Production, study and us© of In pure and applied nuclear research

by

Tor Bjernstad

University of Bergen 1986 THESIS for the degree Doctor Philosophiae (dr.philos.) at the University of Oslo.

Obligatoric lectures: April 25, 1986. Defence of the thesis: April 26, 1986. to tlie thesis

PRODUCTION, STUDI AND USE Of SHORT-LIVED NUCLIUES IH PURE AND APPLIED NUCLEAR RESEARCH

by

lor BJornstad

LOCA!ION IS WRITTEN SHOULD BE

238,, • 235,, REVIEW PAPER, (>.)), line II "...Induced fission of U..." "...Induced fission of U..."

REVIEW PAPER, p.14, Une 4 "Chemical metods..." "Chemical methods.,."

SEVIEW PAPER, p.JO, line 17 "lived mud ides ("or..." "livpd nuclides for..."

REVIEW PAPER, p.29, line 11 '...for sampling of this,.," "...for sampling of thin..."

REVIEW PAPEfl, p.30, line 26 \..T1/2(M) * Tl/JID)..." \..Tj{H) * Tj(D)..."

REVIEW PAPER, p.31, line 11 "...suncrocyclotron..," "... synchrocyclotron.*."

REVIEW PAPER, p.31. line B **... Trie release tfie »uclear..." "...The release of the nuclear. from the bottom

REVIEW PAPER, p.31. line A "...low vapor pressure..." ..low vapour pressure., from the bottom

REVIEW PAPER, p.35, line •'...76Or(tj-1.355)..." \..76V(tj-1.3Ss)..."

REVIEV PAPER, p.47, 1 lues 6, Ptiranthese after the reference on These parflntfieses should be on the 11,1? and 14 ttje level of Hie text line level of the reference number. In addition, there should be Inserted connid between references in line 15 and 12.

REVIEW PAPER, p.48. line 1 "The separted molecular. "The separated molecular,.,"

PfVIEW PAPER, p.48, line 7 "...irridfum..." "......" from the bottom

REVIEW PACER, p.50. line 10 "...IMlA..." "...1NAA..." 1 rom the bottom and p.SI, II»« 10 and 20

REVIEW PACER, p.59, line i "...by elctromaonetic nwiss..." "...Ly electromagnetic mass..." (rom the bottom

r P.I], p.101, fig.4 The 156.0 kcV ,-r,iy should be »Ir.iw. froi.- the 3?).O keV level to the 165.0 keV level.

Paner T.IV, p.M9Z. column 2, "...lhese >-rays, e.o. the transitions "...Those >-rays, e.g. the transitions lines 20,21 and II from the at 183.2 and compocrnt. Ifiese -,-rays, at 103.2 flnd 503.2 keV, are, howpver, bottom e.ci. the transitions at 183.3 and 50.1.2 relatively weak, and thus the counting keV, are, however, relatively wetfc,«nd statistics «re comparitive)y poor." thus the"

Paper P.V, p.l93U, column 1, "...is reduced ans stripped..." "...Is reduced and stripped,.." line 7

Csper CV, p.193?. column ?, "...light lanthanide "... 11 nil c liinthanidp x-ray..." line 7

raner r.VIII, p.377 rorurulA 11?) is fnconrctly derived from the i-rpresstotl I"). FOnmtlft ( 11) K of nr? real Interest and should hp lonared.

Fapv r. VI11, p. V7, I itips 11 If an analytical expression "If .ir- an.iiyticnl expression for p(t; *s 12 from the bottom j)M(i1 latiir, t; «tiny lie calculated f OP- <> prcselpcted value of 2 Ity a nuwrfc? t Inteorét inn mpt'iod. Likewise, the con- stant value of the integral

may then be

rSVpr p.vlil, p.378, line 9 ".,.* and s nonlsotioptc:" "...* «nd j noritsotoplc" from the bottom

Piper P.VIII, p.381, line 21 ".,.practical hindermces..." "...practical hindrances..."

•'»per r.vlll, p.382, line 2 "Cast 1. the measured count(no r«tc. "Case 1. 1hp measured countiixi

i- i1 \\ ! ' . :- i»?. pn-if?nt Paper P.Vill, P.3BJ, eqn(29) i r" .+rp.« T rp. j-1 1J Paper P.VIII. p.383, line 18 "...counting number also have to be. "...counting numbers also have to be..."

Paper P.VIII, p.383, line 2\ "The measuremenys ($..." "Hie measurements are..."

Paper P.VIII, p.385, eqn(34) p(t)-yoexp-[^—:; ln(^ t-jl *D;

Paper P.VI11, p.38?. 1 i ne I "..., I. Rudstad-Haldorsen,..." ,, I. Rustad Maldorsen,..."

Taper p.XI, p. 733, column 1, ,.a stable ion source operation line 15 "...a stable source)."

,the long delay..." Paper P.XI, p.733, column 2, "... the log delay..." line 17 from tlie bottom

Paper P.XI, p.734, column 1, .made of quartz." Une 13 from the bottom ",..made of quarts ."

Paper P.XI, p.734, column 2, "... a water-cooled copperLress..." "...a water-cooled wrapper. line 4

Paper P.XI, p.734, text on fig.3 "Cu-tress for coDlIno" "Cu-wrapper for cooling"

Paper.P.XI, p.735, column 1, "...number of given by..." "...(iuifrf»er of counts oiven by..." line 7 from the bottom

Paper P.XI, p.735, column 2, line 6 from the bottom

Poper P.XI, p.736, column 1, "...the C -fons..." Une 7 from the bottom

Paper P.XI, p.736, column 1, "...different elements)17,..." "...different elements)17',... 1 line 2 from the bottom f'

Paper P.XI, p.736, column 2, "...tarpet container degrade..." ...target container degrades..." line 9 from the bottom

Paper P.XI, p. 737, column I, "...from 12C6*-frr«diatfon..." ...from ICC -Irradiation..." line 6 CO

Paper r.X, p.ZZU. column 1, "...the effect or short time..." ... the effect of short time..." line 20

Paper r.X, p.2Z8, column 2, "...was determied by..." "...was determined by..." line 17

Paper P.KI, p.738, column 1, "...surfaced in..." "... surfaces in..." line 4

Paper P.Xfl, ff.308, coluiwi 2, "., ,a fiobium powder..." "...a powder..." Iine 2 from the bottom

1 ''aper P.XJII, p.37, column 2, ",..missing 3" -» 0" transition. "...missino 3" * flf transition." line 26

Piper P.XIV, several places: Examples: 7,210 keV, 4,351 keV To avoid confusion, this Is enuivairnt first 3 times in the abstract, to 7210 keV and 4351 fceV or /.?IO HeV Ute« 16 places in the text and 4.351 MeV

Taper P.XIV, p.95, column 2, "...and RFA calculations..." "...and RPA caIcut»t(ons..." line )0

Paper P.XIV, p.96, column 2, "The garnna deday..." "The qamma decay..." line 15

Piper P.XV, p.3, lints 4 and "...IHHA..," 7 • text to table 1, p.4, line 22

Paper P.XV, p.8, lLrw 23 "... In a viirm ni trogen..." "...In a warm ..."

Paper P.XV, p.8, last Une "During irratiatlon..." "During lrrJdi

Paner P-XV, p.15, line 11 "... ICH2

I'.i[-r?r P.XVI, p.3 . tflble 1, 14O(O.2f colimi 3

IMfTr P.XVI, p.3, Uble 1, facer P.M. p.-33, column 2, "... the ioj delay..." " ,.,the long delay..." line I " from the bottom

Paper P.Kl, p.?y*, column J, "...made of quarts." ".,.made o( quartz." Une 13 fron tlie bottom

Paper P.XI, p.734, column Zt "... a water-cooled coppertress..." "...a water-cooled copper wrapper.. 1 ine 4

Paper P.xl, p.731, text on Tig.3 "Cu-tress for coo)ino" "Cu-wrapner for cooling"

Paper P.%\, p.735, column 1, "...number of given by,.." "...number of counts n I ven by..." 1 Ice 7 from the bottom

Paper P.Xt, p.735, column 2, line 6 from the bottom

Paper P.XI, p. 73(i, column 1, "...the "C -ions..." line 7 Trom the bottom

Paper P.Xl , p.736, column 1, ",..different elements) ,.,." "...different elements) ,..." line 2 from the bottom

Paper P.Xl, p.736, column 2, ...target container degrade.,," "...target container degrades..." I trie 9 from the bottom

Paper P.Xl» p.737, column 1, ...from I2C6*Mrraditttfon..." line 6

Paper P.X, p.220, column 1, ...the effect or short time..." ... tlie effect of short tine..." line 20

Paper P.X, p.228, column 2, "...was determied by..." was determined by,,." line 17

Paper P.XI, p.738, column 1, "...surfaced in..." "... surfaces in..." line 4

Paper P.xil, p.308, column 2, "...fl nob!urn powder..." "...3 niobium powder..." 1ine Z from the bottom

Paper P.XIIf, p.37, column 2, "...missing 3 •• 0 transition." "...missing 3 * 0 transition." line 26

Paper- P.XIV, several places: Examples: 7.210 JceV, 4,351 lu?Y To «void confusion, this is equivalent first 3 times In the abstract, to 7210 keV and 4351 keV or 7.210 KeV then 16 places in the text and 4.351 MeV

Paper P.XJV, p.95, column 2, "...and RFA calculations. "...and RPA calculations..." line 10

Paper P.XJV, p.96, column 2, "The ganina deday..." "The gatmia decay..." line 15

Paper P.XV, p.3, lines 4 and 7 • text to table 1, p.4, line 22

Paper J>.XV, p.8, line 23 ",..ln a varm nitrogen. "...in a warm nitrogen..

Paper p.XV, p.8, last line "During lrratlatlon,.," "DurLng Irradiation.,."

Paix*r p.XV, p. 15, ll/ie J!

Paper P.XVI, p.3, table 1, coLini 3

P.ipnr P.XVI, p. 3, table 1, cohim 6 f'afor P.XVI, p.6, J ine T, "...lltternturc..," "... 1itutnture..." p.7, line 10 Iran tlw bottom, p.11,tine 10 taper P.XVI, p.11, line 7 "...of 66Ga from..." "...of Go (rom..,"

Pa(x?r P.XVI, p. 11, ll/Mi ?1 "... titan on i "titan an acceptable...

Paper P.XVII, p.3, Hno 5 "...tJius causes Wss..." "...tJius causliitf ID««. l\i|i>r I».XVII, p,2,lliio 18, I».}, iiiw 1

Pajwr P.XVir, p. 14, line 4 ...occotlonally..." .occasionally..." from t/ic totlom PRODUCTION. STUDY AND USE OF SHORT- LIVED NUCLIDES IN PURE AND APPLIED NUCLEAR RESEARCH

BY

TOR BJØRNSTAD

DEPARTMENT OF CHEMISTRY

UNIVERSITY OF BERGEN

1986 CONTENTS Page

Preface 1

Aknowledgements 6

REVIEW PAPER 7

1. INTRODUCTION 7 2. GENERAL METHODS FOR PRODUCTION OF SHORT-LIVED NUCLIDES 11 3. GENERAL METHODS FOR NUCLIDE IDENTIFICATION 14 3.t. Radiochemical separations 14 3.2. Mass separations 16 3.2.1. Traditional mass separators 16 3.2.2. Second generation separators 18 3.2.3. Time-of-flight particle identification 18

4. EXPERIMENTAL METHODS AND TECHNIQUES USED IN THE PRESENT WORK ... 20 4.1. SISAK 20 4.1.1. Gas jet recoil transportation 23 4.1.2. Dissolution of the clusters and the degassing step 24 4.1.3. The liquid extraction step 25 4.1.4. Liquid reservoirs and control systems 27 4.1.5. Conditions for measurements 28 4.1.6. Performance and applications 28 4.2. ISOLDE 31 4.2.1. Targets 31 4.2.2. Ion sources 34 4.2.3. Production yields 36 4.2.4. Delay properties 39 4.2.5. Stable impurity beams 43 4.2.6. Chemical transport techniques 47 4.3. Application of short-lived nuclides 50 4.3.1. Non-destructive 14 MeV neutron activation analysis 50 4.3.2. Medical radionuclitfe production 53 11

Page

5. SOME RESULTS FROM NUCLEAR INVESTIGATIONS 60

•i . 1 . Specti otcopi c measurements nn some neutron r.ich 1 ant.han.i rip

61)

5 . 2 Speet rosciopic: mea suremftnt s on t.he doutily magic nucleus

G. OUTI 00K 7^

REFERENCES 7S

PAPFR P.I. P.XVII Preface

This work is based on fourteen published papers and three research reports. The corresponding experimental investigations, which are of both basic and applied character, were carried out in the from 1973 to 1983. The experiments were performed at various laboratories, partly within the international SISAK and ISOLDE research collaborations (see table below).

Research /topic Laboratory/institution

Basic studies SISAK: S_hort-lived Isotopes Institute of Nuclear Chemistry, Studied by the AK.ufve technique. University of Mainz. On-line chemical separations based on liquid-liquid extractions.

ISOLDE: Separator O_n-L.ine European Organization for Nuclear to the CERN synchrocyclotron ISC). Research (CERN), Geneva.

Applied studies: H MeV neutron activation analysis Department of Chemistry, University of Oslo.

Medical production Institute of Physics, University of Oslo.

The individual papers are:

P.I. N. Trautmann, P.O. Aronsson, T. Bjernstad, N. Kaffrell, E. Kvile, M. Skarestad, 6. Skarnemark and E. Stender, "The combination of the gas jet recoil technique with the fast chemical on-line separation system SISAK", Inorg. nucl. Chem. Letters, H, 729 (1975) .

P.II. 6. Skarnemark, E. Stender, N. Trautmann, P.O. Aronsson, T. Bj«rnstad, N. Kaffrell, E. Kvile and M. Skarestad, "Decay properties of some neutron-rich isotopes", Radiochim. Acta, 22. 98 (1976). p.III. T. Bjørnstad, E. Kvåle, G. Skarneriurk, P.O. Arunsson, N. Kaff r<.0 .1 , N. TrauUnann and f. Stender, "Decay pr opcj It t'i, of la", J. mory. nut. .1 . ftiem., 23. • 1107 ( 1977) .

P.IV. G. Skarnemark, P.O. Aronr.son, T. Bjørti;;tad , E. Kvile, N. Kaffrell, E. Stender and N. Trautmann, "Decay properties of La", ). morg. nucl. C hem. , JJ3, 1487 ( 1977) .

P.V. T. Bjørnstad, F. KvåJe, G. Skarnetnark and P.O. Aronsson, "Oecay properties of some neutron-rich isotopes", J. inorg. nuc.l. Chem. , 19., 1929 (1977).

P.VI. 6. Skarnemark, P.O. Aronsson, K. Broderi, .1. RyiJburg, T. Bjarnstad, N. Kaffrell, E. Stender and N. Trautmann, "An improved system for fast, continuous criemir.a] separation ("SISAK 2") in nuclear spectroscopic studies", Nucl. Instr. Meth. , HI. 323 ( 1980) .

P.VII. S.I. Nøvik, T. Rjornstad, J. Alstad, K. Brodén and G. Skarnemark, "An automatic device for sampling of thin assays of short-lived radionudides in a liquid flow", Nucl. Instr. Meth., 165. 175 ( 1981 ) .

P.VIII. T. Bjurnstad, "A continuous on-line method for fission yield measurements with the combined GJRT-SISAK technique", Nu.:l. [nstr. Meth , J_31, }T> ( 1981) .

P.IX. T. Bjernstad, L.C. Carraz, H.A. Gustafsson, J. H 8. Jonson, O.C. Jonsion., V. Lindfurs, S. Mattsson and H.L. Ravn, "New targets for on-line mass separatjon of nuclei formed in (ifjri MeV proton and 910 Mt?V Me reaction:;", Nuc 1 . Instr. Mclh., Iflf.. 391 ( 1981 i . P.X. T. Bjørnstad, H.Å. Gustafsson, B. Jonson, O.C. Jonsson, V. Lindfors, S. Mattsson, A.M. Poskanzer, H.L. Ravn and 0. Schardt, "Comparative yields of alkali elements and from 3 12 irradiateirradia d with GeV protons, He and C", Z. Phys. A, 303, 227 (1981].

P.XI. 0. Glomset, T. Bjørnstad, E. Hagebø, I. Haldorsen and V. Hjaltadottir. "A -based molten salt target for selective production of neutron-deficient isotopes in high energy proton spallation reactions", Proc. "4th International Conference on Nuclei Far From Stability", Helsingar 7-13 June, 1981, CERN 81-09, p. 732.

P.XII. 8. Vosicki, T. Bjørnstad, L.C. Carraz, J. Heinemeier and H.L. Ravn, "Intense beams of radioactive produced by means of surface ionization", Nucl. Instr. Meth., 186. 307 (1981).

P.XIII. T. Bjørnstad, L.-E. de Geer, G.T. Ewan, P.G. Hansen, 8. Jonson, K. Kawade, A. Kerek, W.-D. Lauppe, H. Lawin, S. Mattsson and K. Sistemich, 13 2 "Structure of the levels in the doubly magic nucleus __Sn '', 5 0 8 2 Phys. Lett. . jy_B, 35 (1980) .

P.XIV. T. Bjørnstad, J. Blomquist, G.T. Ewan, B. Jonson, K. Kawade, A. Kerek, S. Mattsson and K. Sistemich, 13? "Excited states in the doubly closed shell nucleus -ISn. " 50 82 Z. Physik, A306, 95 (1982) .

P.XV. T. Bjørnstad, J. Alstad and T. Johannesen, "The construction of a facility for 14 MeV neutron activation analysis of in ", Report 84-05, Institute of Physics, University of Oslo, 1984. P.XVI. T. Sjørnstad and T. Holtebekk, 6? "ProductioProduction of Ga at tthhe Oslo cyclotron", Report 03-1 Institute of Physics, University of Oslo, 1983.

P.XVII. T. Rjernstad, T. Holtebekk and A. Ruud, "Production of Kr-generators at the Oslo cyclotron", Report 84-02,Institute of Physics, University of Oslo, 1984.

One of the main concerns of this work has been to evaluate methods and techniques for pure and applied nuclear studies involving short-lived nuclides. Hence, the papers P.I, P.VI - P.XII and P.XV P.XVII are all mainly of a technical character. The papers report on the on-line performance of the SISAK system at a nuclear reactor (P.I), on technical development for preparing sources for 0-particlo, neutron and x-ray spectroscopy from the continuously flowing liquid in the SISAK system (P.VI), on important system improvements concerning the liquid hold-up time (P.VII), and point to its possible use in nuclear reaction studies (P.VIII). The papers further report on the development of new production targets at ISOLDE for 600 MeV proton and 910 MeV He particle irradiations (P.IX and P.XII, on tests with a heavy-ion beam of 1 GeV C-particles (P.X), and on the present, availability of mass- separated beams of the elements through new ion-source development (P.XIII).

Some results from nuclear spectroscopic studies of nuclides in selected mass regions when using such new or improved techniques arq given in papers P. II - P.V and P.XIII - P.XIV. The SISAK experiments have been restricted to the transitional region of the light lanthanide elements La, Ce and Pr, while at ISOLDE the doubly magic nucleus Sn has been studied through the decay of 132In.

Examples of techniques for practical application of short lived nuclides in radiochemical analysis and radionuclide production for medical purposes are given in papers P.XV - P.XVI I. The connection between Iheie three papers and those already mentioned lies in the common principle1- and problems related to the production, handling, and measurement of short lived nuclides In the review paper an attempt har. been made to briefly describe the present level of achievement concerning methods and techniques for the production, se- paration, and study of short-lived nuclides. Special attention has been given to the SISAK and ISOLDE techniques. In addition to comments on principles and a few main results, some technical layouts not previosly reported in detail have been described. Acknowledgement s.

I was first introduced tu the field uf niiclcdi studien dm( technology through the inspiring lectures given by Prof. A.C. Pappa; tu whom I owe I Iv decision of graduating in nuclear chemistry.

My closest supervisor as a student and in my subsequent research has been Or. J. Alstad. I would like to express my profound gratitude to him for constructive criticism and helpful advice on numerous occasions.

I also gratefully acknowledge the rewarding collaboration within the SISAK group where my closest collaborators have been the Dr . •; t.R. Haldorsen, E. Kvåle and M. Skare-.tad of the University of Oslu, flr.s P.O. Arnnsson, K. Broder, and G. Skarnemark of Chalmers University of Technology in Gothenburg and the Or.s N. Kaffrell, E. Stender and N. Trautmann of the University of Mainz.

The two years I spent in the ISOLDE-group were fruitful and extremely interesting and I would like to offer my sincerest thanks to my closest collaborators, Mr. O.C. Jonsson and the Dr.s H.Å. Gustafsson, B. lonson, V. Lindfors, S. Mattsson, A.M. Poskanzer and H.L. Ravn.

The last three years have been spent doing applied research at the Oslo cyclotron, and I am deeply grateful for the strong support ami help of Or. T. Holtebekk and for the skillful technical assistance of Mr. A. Ruud and Mr. E.A. Olsen.

Financial support has been obtained from The Norwegian Research Council for Science and the Humanities, The Royal Ministry of Education, The ELKEM A/S Research Foundation and The Astrup Foundation. REVIEW PAPER

1 rNTROOlKTION

Nuclides are conveniently mapped according to their proton number (Z) and (N) in an N-Z plane. Such a mapping, called the nuclear chart, is shown in fig. 1. By letting an axis perpendicular to this plane represent the energy content, the energy (or mass) surface forms a valley of stability with the most stable nuclides lying at the bottom. When moving away from stability, the hill-sides become increasingly steeper (decay energies higher), and the half-lives shorter. Here the nuclear landscape is largely unknown.

Since protons and neutrons can combine together in many different ways, roughly some 8000 nuclides are thought to be stable enough to exist. They are contained within the outer broken lines in fig. 1, the so-called proton and neutron "drip-lines", where the binding energies for the last nucleon, B and 6 respectively, are zero. Out of these nuclides only 280-290 are stable n against 0-emission, and they represent the naturally occuring ones. They appear as the black squares in the nuclear shart of fig. 1.

An additional 1900 nuclides have been detected and studied in some detail. These are confined within the inner broken lines of fig. 1, are all radioactive and decay by emission of j3, a or delayed particles, or by . Our present state of knowledge about the nucleus stems from the - 2200 nuclides composing these two categories.

The importance of extending the knowledge of the nuclear chart was emphasized by the Lysekil Symposium "Why and How should we Investigate Nuclides far off the Stability Line" in 19GG (see in particular the introduction by I. 2) Bergstrom ). New and improved data would, for instance, provide improvements of mass formulas and a closer understanding of the nuclear structure and of nuclear phenomena like fission. This would further have an impact, on the understanding of cosmo-nucleosynthesis, on the development of nuclear technology, etc.

That conference ignited the coals already glowing in this field of research: new techniques were invented and facilities built for the production and study of short-lived nuclides. This resulted in an enormous amount of data during the following years on nuclear masses , half-lives, ground-state spin and 110

100

90

80

60 z z 50 o o £ 40

30

20

10

0 10 20 30 40 50 60 70 80 90 100 110 120 130 U0 150 160 NEUTRON NUMBER (N) Fig.l. Chart of the nuclides. The black squares represent the naturally occuring ones. The limits for detected nuclides are revised to June 198l. moments, nuclear shapes, structure of excited energy levels and decay probabilities. This is apparent in the proceedings from the subsequent conferences in the same series at Leysin in 1970 , at Cargese in 1976 and at Helsingor in 1981

Examples of outstanding results recently obtained are the development of the "Interacting Boson Approximation" model (ISA), which seems capable within limits of predicting the level structure in totally unexplored terrain far from stability, and fascinating and unexpected discoveries that range from the heaviest to the lightest elements, like the shape coexistence in the light isotopes , the very strong deformation in Sr and the 0-delayed two and three neutron decay of Li '

Somewhat in the shadow of the highlights, a number of experiments have produced results that are of general value to the field. An overview (though not complete) of the present status, both with respect to theoretical achievements and experimental development, may be found in some recent conference proceedings and review articles , (see in particular the article by 8. Jonson ).

Although the forces binding nucleons together in atomic nuclei are now broadly understood, a detailed general description of the behaviour of the many protons and neutrons is not yet possible. Different aspects of nuclear behaviour may be approximately described by models designed for each single aspect, especially for nuclides near the stability line. Their validity is, however, limited and they may show serious defects in predicting power when applied to unknown regions of nuclear matter. There is a need to improve and extend the nuclear systematics which is the basis for the model construction and testing: "... the future of the nuclei lies in what we do not know about it and that is a great deal" (D.H. Wilkinson, Cargese 1976 ).

Short-lived nuclides are also found close to the stability line. Such nuclides may be readily produced in simple reactions at nuclear reactors or small particle accelerators (chapter 2), and are extensively used for several practical applications. Depending on their chemical and nuclear characteristics they are of interest to nuclear medicine when measuring short- time variations in physiological functions and for detecting lesions and disorders in the human body (examples are Ga and mKr), in general physio- logy and biology when measuring the metabolism in living organisms of various organic substances (labelled with ^'-emitters like "c, 13N, 150 or 18F), in reserach and education when following short-time changes of physical and 10 chemical processes and in the field of radiochemical analysis. The possibility of making nan-destruct.ive analysis, repeatedly on the same sample if desired, is advantageous. These non'destructive activation methods have had great impact, f specially on the l.iyht elements (7 i. ?n) , where other analytical methods can hardly compete.

The production, study, and use of short lived nuclides require elaborate experiments! techniques. One main aim of the present work has been to develop selective production and rapid radi ochemi ca.l separation techniques, and in some cases employ them to study selected problems of both basic and applied character. ; i

2. GENERAL METHOD1: FOP PRODUCTION OF SHORT-LIVED NUCIIDES

Several method«; may he applied to produce nuc.1 nips far off the stali.il.ity Jinp.

A summary is given in table 1.

The reactions with high-energy protons are the most powerful way of pro-

ducing neutron-deficient nuclides because of the high rate of production

obtained. In the extreme neutron-dcfj c.ient region;;, however, heavy-ion

reactions with compound nucleus formation at bombarding energies normally •; 10

MeV/amu are competitive because of the very high reaction cross sections.

For products on the neutron-rich side, and roughly limited to the mass region

70 < A (, 165, the most important production methods have been spontaneous fis- 252 238 239 sion of Cf and thermal neutron induced fission of U and Pu. As a result of recent experimental development, high-energy charged particle

(proton) induced fission of heavy elements seems currently to be the superior method.

Neutron rich isotopes of heavy elements may be produced by pi»tun- i riducud de«p

spallation of heavy nuclides or by heavy ion damped collisions.

Neutron-rich light nucl.ifies may tie produced iti spallai J on of light elements,

in fragmentation of heavy element targets or relativistii: mediumheavy ions, or

in deep inelastic reactions.

'".bort lived nuclides i.lor.r.' to the stability line are generally produced by

simple neutron reactions like (n,p), (n,u), (n,n') (n,?n) and {n.t] with

neutrons from small isotopi«; neutron sources (C ^ 2 3 MeV) or neutron gene

rators (C ^ K.5 MeV). Some useful charged particle reactions with projec-

tiles f:orn small accelerators include (p.xn) and (a,nnl, where x is an integer depending on the energy of the bombarding projectile Normally, the maximum

available energies at small accelerators are in the range ?0 - 4(1 HpV atid

x i -j 4 far practical .ipplio I ions . max Table 1. Examples of production methods for nuclides far off the line of beta-stability.

Method Reactions Target ele- Projectile energy Products References ment (nuclide) and type

A Heavy ijn reactions A;M(A:M,xp;yn) ^;£v)M Any s 10 HeV/amu Proton-rich, any element (compeund nucleus formation)

Examples: !lY(!§Ar,2p;xn)l27"^Cs 5.9 MeV/aiiw Neutron-deficient cesium isotopes 18) 18!Rh(HNi,xn)1"-ISiTa '5lRh Neutron deficient isotopes 19 ) 2 2S §|Pb(!STi,xn) !;Si04 2S!Pb 4.75-5.15 HeV/amu Isotopes of the "nameless" element 104 20 )

A^Aa^AHx^y^^A^x^y-x^ Heavy ion damped Actinides s 10 MeV/amu Heavy actinides collisions

Example: 2lIU(2liU,22°^Po;xn)f5!Ftn 2 3 a • • s 7.5 MeV/amu Heavy isotopes 21) 92U

A x Spallation reactions >p,xp;yn) -< ^;]H Any 5 200 MeV -.0 Proton-rich, any element (mainly proton-in- several GeV Neutron-rich, heavy elements (deep duced) Spallation).

Examples: 2?Nb(p,5p;yn)71>-!;Rb Natural 600 HeV p Neutron-deficient isotopes 22,23) niobium J J J ?fU(P>6p;yn) "- i;Fr Natural 600 MeV p Neutron-deficient and neutron-rich P.IX. urani urn isotopes 5 3 32 2 !V( He,7p;yn) -i|Ar Natural 910 HeV 'He Neutron-deficient and neutron-rich P.IX. isotopes

Proton-induced Complex Mainly ele- Some hundred MeV Neutron-deficient and neutron-rich fragmentation re- ments heavier to several GeV isotopes of several light elements actions than

Examples: Natural 600 MeV p nLi 10) uran i uni Neutron-rich isotooes out to 24,25) '"Na

Continued on the following page Table t continued.

A A Fission *;H(proj.,f;yn) z;M + '-^-^M Mainly heavy Thermal neutrons» In low energy fission: AA,72 to Au16: , elompnts like fast neutrons, high neutron-rich nuciides. , uran- energy charged par- In high energy fission: a few masses ium and heavier ticles (mainly p) added on eighter side of this region

Examples: T;U(nth,f;yn) Thermal neutrons 18 elements between and bari"n 26)

•"3!u(p.f;yn) 600 MeV p From gallium to the lanthanides P.IX. r J i 3 5(| 2 3 9n.. P.I.-P.VI. Uu, 18Pu(nth,f;yn) 92U, J1.VU Thermal neutrons , cerium and praseodymium

Fragmentation of Complex Light element > 100 MeV/amu Neutron-rich isotopes of light elements relativistic heavy ions

Example: 2Be(2oCa,fragnn.;xp;yn) 2)2 HeV/amu Nuciides from 22N to "5C1 27)

A A, A xiy A,;x;y Deep inelastic re- 1H( 2! H) M Heavy elements < 10 MeV'amu Neutron-rich isotopes of light elements ^1 ^2 i2^X Li+X actions Zs83

] Examples: ^SThCfNCor |0 or f§Ne),prod.) 9.7 HeV/amu N Nuciides in the region "He to 25F in 28) 2 !§Tb 6.8 HeV/amu 0 the transmitted beam 7.9 MeV/amu Ne

MflKJJAr.prod.) 8.5 MeV/amu 21§U Nuciides in the region "Sc to 6"Fe 29,30,31) '§lBi(f|Fe,prod.) 2 0 9n J 8.3 MeV/amu B3»l

Photonuclear reac- flM(Y,yn)A"^M Any Several hundred Neutron-deficient nuciides tions (bremsstrahl- MeV unq)

127, i 2 0 T Example' 300 MeV 32)

Successive neutron Any Neutron spectrum Neutrci-rich nuciides of any element 33) capture in high from nuclear neutron fluxes explosions

Double charge ex- *>M(*rH, , A^M) A'M Light and A few hundred Neutron-rich and neutron-deficient ix ii i ti 7 Zi+Z change reactions 2 medium heavy MeV light and medium heavy nuciides elements

Examples: ;;ca(;;Ti, ;;ca)22Ti 5?Ca 385 MeV JIT,' 34)

'So 164 MeV 35) H

3. GENERAL METHODS FOR NUCLIOE IDENTIFICATION

Nuclear reaction mixtures are often too complex to allow detailed studies of individual nuclides directly. A separation may be required. Separation methods may be roughly grouped as either chemical or physical. Chemical metods provide a separation with respect to the nuclear charge Z. Physical methods often involve electromagnetic deflection, and separate nuclides mainly with respect to their mass number A.

3.1. Radiochemical separations. A large number of ingenious methods have been developed for the chemical isolation of a certain species from complex nuclear reaction mixtures. In the years before the invention of the high resolution detection techniques, a highly refined separation chemistry was absolutely imperative. Since most procedures were designed mainly for manual operation, the separations were often rather time-consuming and hence inapplicable to the isolation of short- lived nuclides. However, in parallel with the improvements of the detectors and the increasing demands for the study of still more short-lived nuclides, a development was continuing towards faster and more automatized chemistry, in some cases at the cost of absolute selectivity. These fast techniques make use of general separation principles like liquid-liquid extraction, precipitation, isotopic exchange, ion exchange, selective chemical evaporation, eletrode- position, electromigration, thermochromatography etc. For a comprehensive overview see the review articles in references 36-38.

Examples of discontinuous approaches may be found in references 39-40. One outstanding, fast discontinuous technique which utilizes several separation principles in a series on a sectioned separation column, has been developed at 38) the research reactor in Mainz . Isolation of niobium from a fission mixture has, for instance, been achieved within 2.0s, which has enabled identification of neutron-rich niobium isotopes as far out as 0.8s Nb

For experiments demanding a large amount of data, for instance coincidence experiments in various forms, the discontinuous separations may be too laborious and time consuming.Continuous separations on-line to the production unit are usually preferred. 15

Thermochromatography has been adapted for continuous on-line separations. The chromatograph consists in principle of a tube (glass, quartz or metal) with a temperature gradient in the axial direction. The tube may be empty or filled with a packing material, or the walls may be coated in several sections with different chemical compounds. The evaporized reaction mixture is led into the high-temperature end of the tube, and the different species are condensed according to their vapour pressure (adsorption chromatography) or adsorbed according to their chemical affinities (chemisorption) ' . If vacuum conditions are not required, the reaction products may be collected in a gas and sw°pt directly into the chromatograph (gas jet recoil transportation) for separation. The gas may also be mixed with reactive chemicals in order to form classes of volatile compounds like halides and oxides which are thereafter •^ ft 1 r\ in I transported to the chromatograph ' . For separation in a vacuum, the reaction products may be collected onto a high-temperature resistant surface (, , tantalum, , and others) as direct recoils from the heavy ion reactions or from a preceding mass separation step Products with sufficient volatility may be continuously reevaporated, either into the chromatograph for further separations, or in order to remove disturbing species and leave a source on the collection foil of simple composition which may be measured directly. In this way and isotopes have been separated from other isobaric fission products and 52) and mercury isobars separated from each other. The same technique has been proposed for sequential enrichment of individual isobars of rare earth elements

Another thermoseparation principle which may be applied is the selective release from the target matrix itself with or without the action of a reactive chemical. Such a chemical may either be a constituent of the target matrix or added as a gas during irradiation. This method is extensively applied in combination with mass separation, and will be further discussed in section ^.2 together with the chemical selectivity of various ion sources.

Continuous separation from liquid media has been accomplished by electrodeposition, isotopic exchange, ion exchange and liquid-liquid extraction. The former principle was used to collect onto a moving copper covered tape and copper onto wool . The other principles have been utilized in the SISAK separations, and will be further discussed in section 4.1. 16

Host of these continuous techniques have been used to study nuclides with half-lives from a fraction of a second and upwards.

3.2. Mass separations. 3.2.1. Traditional mass separators. In traditional mass separators the reaction products are slowed down either in the target material or in a catcher and reevaporated into an ion source. The produced ions are extracted and accelerated to form an ion beam. Provided that the ions are all in the same charge state, they will gain the same kinetic energy in the acceleration potential. By the action of a magnetic field perpendicular to the ion beam, the ions are deflected according to their momentum (mv) which results in a separation of mass.

Normal mass separators (also termed isotope separators) have a mass resolution (m/flm) of a few hundred to a few thousand, thus offering an effective separation of isotopic mass chains (A), while the isobars themselves can not at present be separated by this method.

Normally, ions are extracted from the ion source in the monovalent state. However, higher charge states may occur, and different masses may then mix in the same ion beam after the magnetic analysis. By applying a subsequent electrostatic deflection, which analyses the beam with respect to the kinetic energy, an unambiguous determination of the mass number A may be obtained.

A variety of mass separators dedicated to the separation and study of short- lived nuclides exists (see the proceedings from recent conferences on .Electro Magnetic Isotope Separation IEMIS) " '). A summary of the more successful facilities on-line to thermal reactors, high-energy proton accelerators and heavy-ion accelerators, is given in table 2. The capabilities of the various types of production/separation installations are compared.

As mentioned above, a common characteristic of these on-line isotope sepa- rators is that the reaction products must be stopped, reevaporated and thermally diffused to the ion source. The reevaporation step may be rather time consuming and strongly element dependent. Therefore, the average delay time between the production and the collection of the separated species is generally >> 10 s. Nevertheless, successful experiments have been performed 17

with nudities as short-lived as the 8.5 ms Li . However, several elements can at present not be released at all in usable quantities [ , , ...).

Table 2. On-line mass separator facilities for studies of short-lived nuclides, and their capabilities in general.

Production units Nuclear reactors High-energy proton Heavy-ion accelerator!! accelerators

26 60 22 66 70) Productive on-line 1. OSIRIS ' \ 1. ISOLDE ' ', 1. BEHS-2 , masa (isotope) sepa- Studsvik, Sweden. CERN, Geneva, Dubna, USSR. rator facilities Switzerland. 2. SOLIS61\ Soreq Nuclear Re- 2. ISOCELE67~69\ Oak Ridge National search Center, Orsay, France. Lab., USA. Israel. 3. GSI on-line mass 3. TRISTAM62t63), separator facili- Ames Lab., Iowa, tylBf73), FRG. USA, recently mov- ed to Brookhaven National Lab.6*). 4. OSTIS65), ILL, Grenoble, France.

Products to be Fission products Fission, spallation Fusion products (neu- separated (neutron-rich). and fragmentation tron-deficient) , deep products (neutron- inelastic reaction rich and neutron- products (neutron- deficient) rich and neutron-de- ficient)

Target type and Fissile material, In principle all ele- Sol id targe ts, mos tly amount mainly 235U a* VC ments (hardly gases). metallic elements, 2-3 g at maximum. Examples of targets: 10-50 nig/cm2. 50-60 g/cro.2 U as UCX, > 100 g/cmz for molten metals like Pb, Ln,Sn.

Bombarding particle 5 uA - S.l-lo13 p/a 1012- 2013 part/s intensity

Reaction cross 2 lo"23 en,2 > lO"25 section

Maximum obtainable 10 atoms/s production yields (at saturation)

Two approaches represent variations of the traditional mass separation sequence, and extend the elements available for mass separation to the refractory and non-volatile ones as well. In the first approach, represented by the HELIO$U~77) separator in Mainz and the RAMA78'79' in Berkeley, the recoiling reaction products are stopped in and transported with a gas jet (see section 4.1.1.} into an integrated skimmer/ion source unit. Here, the carrier gas is removed and the products ionized for subsequent mass separation. The transport efficiency is largely element independent, and the delay time is determined by the gas jet transportation time (- Is). In the second 18

a n ft 1 J approach ' the recoiling reaction products are slowed uown in gas Ihigh ionization potential), thus defining the initially multiple charge state for most nuclides finally to +1. These ions are then transported a few mm in a helium flow to a skimmer system leading directly into the extraction and acceleration chamber of the mass separator. The transportation time for this helium-jet ion guide system is of the order of 10 ms and the total efficiency measured to - 5'/.

These techniques are still in the developing stage.

3.2.2. Second generation separators. Limitations in traditional on-line separation such as delay time and elements available for study have partly been overcome with the second generation separators. These are a) the electric and magnetic field recoil separators, b) the gasfilled magnetic recoil separators and c) the kinematic separators. 8 2 83 ) They may be exemplified by the LOHENGRIN ' at the high-flux reactor in Grenoble, JOSEF ' at the reactor in Julich and the velocity filter 86-88 ) SHIP at the heavy-ion accelerator UNILAC in Darmstadt, respectively. The recoiling reaction products are separated in flight, thus obtaining a total separation time of 1-2 us. The former two installations provide a separation in Z as well as in A, but have a rather low efficiency (10 - 10 ). They are especially suitable for the study of e.g. isomers with half-lives down to microseconds, which are formed in relatively high yield. The high separation efficiency of the velocity filter SHIP makes it ideal for investigations of short-lived evaporation residues from heavy-ion reactions with cross sections down to nanobarns.

3.2.3. Time-of-flight particle identification. A recoiling particle emerging from high-energy proton or heavy-ion reactions may be uniquely identified by the combined time-of-flight and energy-loss (dE/dx) technique which involves no chemistry and no mass separator. The time-of-flight information and the measured total particle energy (El uniquely determine the mass number A. The atomic number Z may be obtained from the time-of-flight information and dE/dx-measurements in thin detectors. aq j This technique is applicable mainly for masses A < 50 , and have been used to identify nuclides close to the drip-lines. 19

The remainder of this review paper is devoted to a closer description of the experimental methods and techniques developed and applied in the present work, and to a brief discussion of some of the spectroscopic results obtained in selected mass regions. 20

4. EXPERIMENTAL METHODS AND TECHNIQUES USED IN THE PRESENT WORK

4.1. SISAK

A substantial portion of the present thesis, the papers P.I - P.VIII, is related to the experimental technique SISAK, and this section provides a description of the main features of the system layout and the performance and operation of the technique.

SISAK is an abbreviation for "_Short-lived Isotopes .Studied by the AJLu technique, and designates an experimental technique which includes the production, chemical separation, and detection stages of short lived nuclear species from complex nuclear reaction mixtures.

The idea behind this technique was conceived in Gothenburg, in the wake of the 91 ) development of the H-32 centrifuge that aids in the continuous and rapid separation of two immiscible liquid phases (i.e. aqueous and organic liquids). Thus, the SISAK technique is based on solvent extraction methods by employing such H-centrifuges. The separation technique is continuous and on-line adaptable. Its main aim has been to continuously provide sources of short- lived muclides for nuclear spectroscopy studies, but it has potential for other applications such as studies of nuclear reaction yields (P.VIII) and production of e.g. for medical purposes (section 4.3.2.)

Since the first experiments with the technique in 1970, the separation system itself has passed through several stages of sophistication. At the very beginning manual shaking of the liquids in separation funnels was part of the separation procedure, tn cooperation with nuclear chemists from Oslo, and later from Mainz as well, it developed into a fully automatic four step (four H-32 centrifuges) system. "SISAK 1". A detailed general description of the working principles and technical performance may be found in the references 55, 92 and 93. The configuration of this system (centrifuges located in separate housings) and the rather large liquid volumes involved, rtsstricted the shortest half-life accessible to 3-5 s.

*) AKUFVE is an abtireviation for the Scandinavian "Apparatur for Kontinu- erlig ,Unders«kelse av fordelingskoeffisienten ved Vaeske vaeske .Ekstraksjon", which translated into English reads: "Apparatus for con- tinuous examination of the distribution coefficient in liquid-liquid extraction" 21

A substantial improvement was introduced with the development of "SISAK 2" 94 ) (P.VI), which included the H-10 centrifuge - a scaled-down version of the H-32 model. The centrifuge liquid hold-up volume was reduced by nearly one order of magnitude to only 12 cm without affecting the liquid throughput rate. In order to reduce the liquid hold-up volume in other parts of the system, the centrifuges were placed together in one housing and the liquid transfer tubes between them could be shortened. The present system allows studies of nuclides with half-lives down to 1 s.

In the early stage the SISAK technique was applied to separate rare-earth rt •* (- IS O ft isotopes from fission of ' U at 14 MeV neutron generators in Gothenburg, Oslo and Mainz. The target consisted of an organic anion exchange resin where the uranium was adsorbed as an UO (SO ) -complex. The fission products were continuously eluted and transported with the liquid stream to the first centrifugal separation step. Liquid transport times of several seconds were necessary. The decay of short-lived products became substantial and the final source strength correspondingly reduced. In addition, high-energy ^-background disturbed the measurements. This background emerged from neutron-induced 16 prompt •y-ray emission and from decay of N (t . =7.13 s) induced by (n,p)- reactions on 0 in the elution solution. Still, the first spectroscopic results on lanthanum, cerium and praseodymium isotopes, using continuous on- 92) line chemical separations, were obtained

The development of a technique for rapid transportation of nuclear reaction products in a gas stream, the so-called Gas Jet jjecoil Jransportaion (GJRT) technique ' , made it feasible to apply a thermal reactor in combination with the SISAK technique (see section 4.1.1.). Thus, the transportation time could be shortened and the background disturbances greatly reduced.

The successful combination of these techniques at the Mainz TRIGA reactor is described in paper P.I. This event marked the start of a series of nuclear spectroscopy investigations. Subjects for decay studies in this work have been the neutron-rich isotopes of lanthanum, cerium and praseodymium (P.II - P.V).

In order to convey a more detailed impression of the technique, attention may be drawn e.g. to fig.1 in paper P.I, showing the flow diagram for separation of cerium isotopes, and to fig.2 below which is a layout of the present experimental installation (SISAK 2) at the TRIGA reactor in Mainz. 2 2

V

) v

Fig. 2. Schematic layout of the SISAK equipment at the Mainz TRIGA reactor: a = reactor core, b = reactor shielding, c = target chamber, d = boric acid/paraffin plug, e = jet gas reservoir with flowmeters, f = cluster production and gas jet control, g = transport capillary, h = degasser unit, i = booster pump, j = ventilation system, k = water baths and heat exchangers, 1 = centrifuge cabinet, m = mixer/centri- fuge unit, n = "thin sample" apparatus, o = waste barrels, p = concrete shielding for the detectors, q - Ge(Li)-detec- tors, r = detector cell, s = main control panel for pumps and centrifuges, t = storage barrels for 'tquid chemicals, u » programmable timer, v = counting electronics,

In addition, the following -JOI tmns gji/ij •» (Ji-I.4i.jpij 'lose r i pt j nn uf SP IPC ted

u «I details anil workitiy principlus of ttw pi .••JruH./ upcratiny system. 4.1.1. Gas jet recoil transportation. Nuclear reaction products recoiling out from a thin target may be thermalized 9 5,90) in a gas and subsequently swept: with tin1-1 'jas throuyh .t capillary"' ' . It is experimentally found that the presence of aerosol! in the carrier gas greatly enhances the transport efficiency ' due to formation of clusters between the aerosols and the reaction products. Aerosols may be of different kinds and may be introduced into the carrier gas by different methods ' ' . For the SISAK experiments one has extensively used aerosols formed when ethylene expands from a reservoir above itb critical 45,107, 108] pressure (51.2 bar) , and aerosols of solid salts like NaC.l, KF and 102, 103) KC1'"""' '. A detailed description of the applied gas jet system may be found in ref. 109. The main features will be repeated here.

Aumann et. al. found two conditions for the formation of

To the target Compressed air

Mixing and filtering tube

Mix. vol. To the Class target filter Flowmeters

Fj'j '( . *: Schem.i t-) 1 cliatj 1 ,j 1 1 ri>] fnt t hi' ft hy 1''PH/M i I t ni|tMi h: fi nf 1 hi1 t.ltor <"nr •.,)]! Ai'raini: 24

aerosol generator is given in fig.3b. KCl-powder heated to 650 C (vapour pressure - 1 torr) is swept with a stream of nitrogen gas. The transportable fraction of the formed aerosols is singled out by the filtering tube where the very light and the very heavy aerosols are removed and by a subsequent glass filter. The average size of the transportable aerosols is - 0.1 urn, corresponding to a mass of 6-10 amu.

The target chamber is illustrated in fig.4. The target consists of 0.1 - 1 mg ?35 239 of U or Pu electrolyzed or molecularplated onto an aluminium foil of 10 mm diameter, and has been covered with a thin aluminium layer (<0.1 mg/cm ) by evaporation. The dimensions of the chamber have been minimized ensuring a

Fig. 4. Cross section of the target/thermalization chamber (from ref. 109).

total thermalization and rapid wash-out of all fission products recoiling out of the target into a nitrogen pressure of 2 bar. The chamber volume is 10 cm . With a capillary diameter of 1 mm, a length of 7 m, and a flow rate of about 20 cm Is, the laminar flow conditions necessary for an efficient transport are obtained. A typical transport time is 0.7 s. For the ethylene jet a transport efficiency of about 70 '/ may be routinely achieved and reliably maintained for several days of operation. The maximum transport efficiency with the salt- containing jet is 65-70 t, decreasing slowly with time because the glass filter becomes increasingly blocked with the retained aerosols. Therefore the filter must be changed at least once a day during continuous runs.

4.1.2. Dissolution of the clusters and the degassing step. The end of the capillary is connected to a static mixer where the jet is placed in contact with an aqueous solution (fig.51. However, the organic. 25

— SUCTION

WATER COOLING

BOOSTER PUMP

GAS-LIQUID MIXER GAS JET

TANGENTIAL LIQUID I AOUOUS PHASE J INLET BAFFLES TO BREAK LIQUID TO THE THE LIQUID RO- FIRST CENTRIFUGE TATION UNIT ,

GLASS CROSS TO PRE- VENT THE FORMATION OF A LIQUID VORTEX

Fig. 5. The degasser unit.

cluster in the ethylene jet, which have become partly polymerized in the highly ionizing field close to the reactor core, are not easily dissolved in aqueous media at room temperature. It has been found (P.I) that temperatures of 90 C and more are desirable in order to break down a substantial portion of the clusters. Also the salt clusters are more readily dissolved in hot aqueous media, and the high temperature is routinely kept in the aqueous solution. The carrier gas is then removed in a subsequent degassing step (fig.5). The gas- liquid mixture is injected tangentially to the degasser inner walls, thus creating a large surface for easy diffusion and separation of the gas. The high temperature of the liquid also promotes evaporation of volatile products like , xenon and some .

4.1.3. The liquid extraction step. The hot liquid from the degasser is transferred by a booster pump through a teflon tubing (i.d. 3 mm) into the static mixer for the first centrifuge unit. Here, it is placed in contact with an organic liquid containing an extraction agent for the element)s) of interest,- alternatively for disturbing elements. The mixture is subsequently fed into the centrifuge where the phases are rapidly separated (fig.6 and 7). Phase purity is obtained by adjusting properly the counterpressure (throttle valves) on both liquid branches separately. One phase continues to the next centrifuge step for further purification of the element of interest. The other phase it recircu- 26

HEAVV PHASE OUTLET - LI3H1 PHASE OUTLET

ORGANIC LIQUID INLET S L:cu:u IHLCT

STATIC MIXER INSPECTION TUBES FOR f-HASE I MR 1TY

MANOMETER

THROTTLE VAI..E

CENTRIFUGE HOUSING

OUTLET FOR OVER- FLOW LIQUID •/-—

V; BRAT I Of. ABSORBER

CENTRIFUGE MOTOR COOLING WATER

Fig. 6. The mixer/centrifuge unit. All mechanical components in contact with the liquids are made of glass or -passivated .

CENTRIFUGE CABINET - WITH PLACE FOR FOUR CENTRIFUGE UNITS

FLOWMETERS FOR THE Z

ORGANIC LIQUIDS

CENTRIFUGE UNIT-"'

VESSEL FOR COLLECTION OF LIQUID OVERFLOW — FROM THE CENTRIFUGES

SUCTION r

PUMPS —

STORAGE VESSELS FOR ORGANIC LIQUIDS

Fig. 7. The centrifuge cabinet containing four mixer- centrifuge units. The cabinet has demountable pttrspex walls in order to facilitate service, and the centrifuge support table is made of thick polyethylene material in order to absorb vibration and reduce no ice. 27

lated if it is organic (preferably after a decontamination step) or it is

transferred directly to the waste barrel«; if jt is aqueous. Four centrifuges

are placed together in the compartment which is connected to a ventilation

system in order to remove organic and acidic vapors (fig.7). The purified

element may finally be transferred by the liquid flow to the sampling and

detection stages (section 4.1.5.).

4.1.4. Liquid reservoir and control systems.

The control panel is mounted on a carriage which also contains the aqueous

liquid supply system (fig.8). The various aqueous chemical liquids are stored

— MAIN POWER

PUMP CONTROL UNITS Aouous SOLUTION TO THE GAi-LIOUID MIXER CENTRIFUGE CONTROL UNITS

COS WHEEL PUMP TEMPERATURE CONTROL

LEVEL ALARM CONTROL

FlOWMETERS FOR THE LIQUIDS

STORAGE VESSELS FOR THE LIQUIDS

, FOR LEVEL ALARMS v FOR PUMPS VoR REFILLING OF LIQUID

Fig. 8. Sketch of the main control rack which also carries the storage vessels for the aqueous liquid chemicals. Each vessel is equipped with a cog-wheel pump and a liquid level control. The vessels are made of either titanium or stain- less steel (for the less corrosive liquids).

in containers of 30 1 volume. The storage vessels for the organic liquids are placed close to or on the centrifuge rack carriage lfiy.7). The flow rate may be adjusted individually for each single liquid and monitored by calibrated rut ameter:.. The rotation speed of the electrically cinvt-ri centrifuges may be adjusted by the use of frequency converters up to a maximum of 2";000 rpm. Thermocouples and pH-meters observe continuously any critical temperature 28

and acidity, respectively. The pumps and the centrifuges are manually con- trolled, and require an experienced operator.

t.1.5. Conditions for measurements. The v ray singles and •yf- coincidence measurements may be performed directly on the flowing liquid. However, for nuclides which are not seriously disturbed by longer lived isotopes or their decay products, it may be advantageous to apply beds of solid catchers in the purified liquid flow in order to retain the activity in a small volume in front of the detectors. The solid catchers may consist of ion exchange resins, extraction agents adsorbed on a polymer support (e.g. di-ethyl-hexyl•orthophosphoric acid (HDEHP) on PVC), a preformed precipitate for isotopic exchange reactions (e.g. Ag [ for retension of Ag, I and Br), or metal powder or wool for direct el eet rodeposition. The thinwalled counting cells applied for these purposes are constructed in titanium or polypropylene, and vary in volume from 10 - 1J 0 cm . For half-life determination with the so-called two-detector delay (TDD) method '' , it is probably advantageous to maintain plug flow conditions (P.VIII,p.) past, tlie first detector, through a variable delay volume and finally p.ist a second detector. Tor this purpose special counting cells have been made of teflon tubings which are wound regularly around the detector heads.

Tor efficient detection of vfry .low-energy elet t romagnet.ic radiation l< 20 keV) and for p- arid neutron spectruscopy, a different approach has been devised. The technique is described in detail in paper P.VII. In brief, the purified liquid flow may be directed to an apparatus where the element of interest may be isolated onto ion exchange membranes as thin sources {- 6 rng/cm ) (P.VII, fig.2). The circular membranes are stretched arid firmly fixed between two steel rings IP.VII, fig. I). Thus, they ^re mechanically stable, arid may be conveyed by compressed air into detection position without distortion of the source configuration. Thu process is semi continuous, and the exact timing of all operations in the separation sequence is performed with a programmable electronic controller.

4 . 1 . G . Performance and applications. Studies of short-lived nuclides demand a minimization of the mass delay in the separation system. Removal of any superfluous hold-up volumes is imperative. Therefore, experiments have been perfoimed in order to investigate the mass transport and the delay introduced by the various parts of the separation 29

and transport system. Accumulation of mass delay curves (P.VIII, fig.8) has been accomplished by the method described in detail in paper P.VIII (p.383). Directly after a pulsed (30ms) irradiation of the target, the intensity of the separated activity has been followed as a function of time (multiscaling). Mass delay curves have been recorded at different locations in the apparatus and during different running conditions. With a gas flow rate of 20 cm /s and 3 *) a typical liquid flow rate of 10.5 cm /s (both phases) the measured delays in the SISAK 2 system were: the gas jet 0.7 s, the mixer/degasser 1.2 s and the mixer/centrifuge unit 2.1 s. The liquid transfer time in the connecting tubes adds to the total delay, but is generally < 0.3 s per transfer. Application of the apparatus for sampling of this assays introduces additionally at least 5-6 s delay.

A doubling of the liquid flow rate is feasible, thus reducing the delay correspondingly. A recently developed gas-liquid separation centrifuge has 110) reduced the delay in the degassing step to - 0.2 s . Today, nuclides with half-lives of 0.5-1 s constitute a practical limit of capability for this technique.

The SISAK technique has chiefly been applied to separate and study the decay of short-lived nuclides which were hardly or not at all accessible with ISOL- separation at the time in question. Thus, continuous separation procedures have been developed for the elements copper, , , , niobium, technetium, ruthenium, palladium, iodine, lanthanum, cerium and praseodymium. Detailed descriptions of the individual procedures may be found in references 92,93,111,112 and in papers P.Ill, P.IV and P.V. Total separation times were generally < 10 s and the chemical yields ranged from 25- 100 I.

As was briefly mentioned above, the SISAK technique may be applied in other fields of study as well, e.g. the determination of nuclear reaction yields (cross sections), studies of radical and hot-atom chemistry, and in the production prosesses of short-lived nuclides for technical and medical purposes (section 4.3.2.).

*) Defined as the time between the maxima of the activity (or mass) distribution measured before and after the transport unit. The principles of a procedure to measure on line the reaction yields of short lived product'. ,ue nut.lined in pape; P. VIII. The oasis of the method is the knowledge of the delay probability f urn: t ion . p(t)dt. , whxch ijxpre^es the probab.iJ.ity of finding ;• delay betwi-.'ii t ,uid L • il t at (lie ijle of measurement. This function is directly represented by --

In paper P.VIII .it. has been shown how the experiment 01 dot pl mi na t .1 m; of t he mass delay can be achieved. Separated lanthanum isotopes wen- u-,ed. In addition to pltlclt, the recoil yield and Ihe transport and separation efficiencies must be known as well. Because these parameters are sensitive to vari a t j i>n:. in the runn ing c ondit mni [flow- ratfi, temperatures, acidity etc.), a reliable and stable chemical, procedure is irnper.it ive. The required production and separation .stability may normally be kept for MIVIM al hours of running.

The derived formalism (P. VIII, eqn.(!)H has onl y limited dirfct applicability. In most actual systems, mot her /daci'jli te; uuclide relations mus'. be considered. The- i elation between \Uv recorded number "f count:;, S, and t he- production rate, R , most be established for. each actual casi-. Without yomy fully into detail hero, u few tirief lumiwriti .ifi1 .uldet) he low.

The contribution to % i'roni primary produced and decay • f o rnied '.pecjys depends on the relative production rate uf tin; mother and daughter ruii..l i dc., t tie half life of the. mother nucli.de, T (M), and the i.on f irjur jt LOM of tin? transport path. If T 1M) and the half-lives of the mottier nui. lide (necui sur s are short > dinpaj cd t ti ltu> half .11 f i1 of the 'lauytln-: nut luk-, T ,,([)), the cumulat i ve yield uf the daughter may be measured by ml rodne i. ion uf j delay volume uf sufficient SJ?« to ensure t.nla.1 decay of the motliej nut.li.de before jepa i a\ i on. When T1/?(M)>T1/?(D) the situation is more complicated. If the shape of plt)dt ( an be uiea-..urnd at I hi; :.xte of separation fur two diffi-ient delay vn) uini"; , ! he i ont r ibut i on to S from decay of l.he moth"r nm. lidt may be «Jetermin^d. Ac <:t)i (I uigl y , the urnnarv yield or l.tie dainjht >.-r nut I.ide and t tie i. umulat J ve yield of the mother nucljdt; may be dei i vvd, piovjdeti lhal a total d^iiy of ttie precutsijrs has occurred before separation.

fi'i far, I lie M 7, AK ( o 1 I aboi a ! i un lia '. MO I IJIVI n pre fe i em e I .. i e>n I i nn yield mea s u remen t s. 31

t.2 ISOLOE

The general idea and principles of the ISOLOE project have previously been extensively described ' . A few main features will be briefly repeated here. The ISOLDE on-line separator is coupled to the CERN syncrocyclotron(SC). The SC produces beams of 600 MeV protons, 910 MeV He -particles and 1 GeV 1 C '-particles with intensities of -5uA, -0.5uA and -O.luA, respectively. These beams may be directed onto the ISOLDE target. The nuclear reaction products that are released from the target are transported by molecular diffusion to the ion source. The ions thus created are extracted and accelereated by 60kV and subsequently analyzed in a fringing-field focussing o magnet of 55 with a radius of curvature of 1.5m. The mass range acceptance is ^15Z of the central beam, and the spacing between the neighbouring masses at the focal plane is approximately D=1500/M (mm), M being the mass of the central beam. This offers the possibility, through the use of an elaborated system of electrostatic deflectors in the switchyard, to extract several masses (sequentially) into one beam line without readjusting the magnetic field, and to serve experiments at several beam lines (4) simultaneously provided they work in neighbouring (not crossing) mass regions. Necessary beam handling equipment (i.e. focussing lenses, electrostatic deflectors and beam scanners) are installed in all beam lines wherever needed. One of the beam-lines is extended by vertical sections to experimental areas in the floors above the basement (where the separator is situated), and a similar extension of another beam line is under construction. A sketch of the isotope separator installation is given in fig. 9.

t. 2 .1. Targets. ISOLDE targets are of the thick type (up to 10Z of the range of 600 MeV protons). The release the nuclear reaction products occurs by bulk diffusion and surface desorption processes. The ideal target material should in general have: a. A high formation cross section for the nuclide of interest, b. a high release rate in continuous on-line operation with, preferably, element selective release, c. stability and low vapor pressure (<10" torr) at the operating temperature, d. relatively high specific density with a high mol- fraction of the target element, e. inertness against unwanted high-temperature chemical reactions, and f. resistance to radiation damage. 3 2

Fig. 9. Perspective diagram of the ISOLOE experimental area including some experimental equipment on the floor in 1970. a = the bombarding beam, b = target/ion source unit, c = magnet, d = collection chamber/switch yard, e = charged par tide spectroscopy, f = high resolution mass spectroscopy, g - beta and gamma spectroscopy, h r range measurements of ions in gases, i = atomic beam maynetic resonance, j = op- tical pumping and laser spectroscopy, k - collection of radioactive sources, 1 - isotope separator control desk, m - neutron, beta and gamma spoctroscopy, n - "orange" beta spectrometer, o - beta and gamma -speet ros copy. 33

Possible target materials are as a rule tested according to these criteria in vacuum chamber and off-line mass separator experiments, and later in on-line experiments. Typical test sequences are outlined in paper P.XI and in ref. 1 K .

Target materials may be divided roughly into three classes: the molten metals, the molten mixtures of chemical compounds and the solid refractory metals and compounds. To the first class belong Ge, Sn, La, Au and Pb, to the second class ThF -LiF and TeO -LiCl-KCl and to the third class Nb, Hf, Ta, CaB , 4 2 6 ScC -graphite, VC, YC , LaC -graphite cloth, TaC-graphite, ThC -graphite, UC - graphite cloth and UC -graphite. Results from the testing of these systems and of other materials are given in references ?3,1K,115 and in papers P.IX and P.XI.

The mechanical design of the target-ion source unit is outlined in detail elsewhere ' ' . The target container normally consists of a plugged tantalum tube (diameter 2.5cm, length 20cm) which is mounted horizontally along the beam axis and heated ohmically by the passage of direct current (OC). The transfer line to the ion source for the thermally diffusing reaction products is attached to the middle of the target container, and is usually made of the same material. The line may be heated or cooled independently of the target heating. For particularly corrosive media the container may be lined with a graphite cylinder or it may be made entirely of quartz.

Liquid target material occupies only half of the container volume while a solid target materiaj fills the whole cylinder volume. The solid target, material consists either of a finely grained powder (grain size -10-20um), of discs of graphite cloth impregnated with the target element, or of thin pellets of the target element (or a compound of the target element) with graphite as a diluent. The target thickness ranges from about 10g/cm (ijr -graphite cloth) to about 170 g/cm (Pbl. The product selectivity of a target system is strongly dependent on the vapor pressure difference between the matrix and the various products, and on the difference in surface desorption activation energies of the products. In most cases the target material releases more than one element. If the mass ranges do not seriously 34

* ) overlap, this will generally cause no harm . For overlapping mass ranges a selectivity can still be achieved if a proper combination of the transfer line temperature and type of ion source is used.

t.2.2. Ion sources. Basically, three types of ionization principles are applied at ISOLDE: plasma, and positive and negative surface ionization. The plasma ion source is at present of the versatile FE8IA0 type ' . It has no (or very poor) element selectivity. However, selectivity may be restored for certain target/product combinations, i.e. in the production of Xe -beams from a molten La-target coupled to a cooled transfer line in order to allow only the most volatile spallation products to pass.

A higher degree of element selectivity is obtained in the surface ionization sources. The method is based on the fact that a heated surface is able to "give or to take away" electrons from atoms colliding with the surface, causing negative or positive ionization respectively. According to the Saha- 118) Langmuir formula , the degree of positive ionization increases with increasing surface work function (•), decreasing ionization potential (E ) of the product element, and increasing temperature. The alkali elements (E = 3.9- 5.4eV) are most easily ionized. Almost 1007. ionization is obtained on a tantalum-surface (•=4.2eV) already at 1500 C. Details about the simple 119) tubular source design , inserted as a constriction in the end of the transfer line, may be found in ref. 115. By exchanging the tantalum-surface with a surface which can be operated at highet temperatures (up to 3000 °C) and which has a higher work function like tungsten (•=*.5eV) and rhenium (•=5.1eV), a good yield of elements with ionization potentials up to -6.5eV may be obtained. In addition to the alkali elements, these also include the alkaline earth and the rare earth elements. At these high temperatures the tubular source gives a higher ionization yield than expected from the Saha- Langmuir formula. This has been explained as being partly due to the space

*) Uncontrolled formation of molecular sidebands may, however, lift a lighter element up into mass region» where it might cause problems.

«*) forced Electron fieam Induced Arc fiischarge. 35

1 20 ) charge phenomena created within the tube . Hence this source represents an

pxtenuon of the pure surface ionization principle, and is often referred to

aq the surface/volume ionization source.

It has recently become possible to adapt the negative surface ionisation

principle to the online production of short lived nuclides (see ref. 1?1 and

pappr P.XII). The degree of negative ionization, a, is expressed by ecin, 1 in

paper P.XII. The ionization efficiency is fl=a/(a+1), and is explicitely

expressed by inserting the expression for a to yield:

— exp(<+ + f(E) (1) 9- o

(See explanation of the symbols in paper P.XII.)

This formula shows that the negative iotiization yield increases with a

decreasing work function (

source described in paper P.XII is based on a LaB -surface which has an

effective work function of • ,,3eV). The temperature dependence of (3 is reflected in fig. 10 which shows the negative ion beam yield, measured by (3 -detection nf

Br(t . = 1 . 35s ) , as a function of the emitter temperature. The curve show«; a pronounced maximum at about 1260 C.

TEMPERATURE (°C) 1065 1150 1235 1320 U05 C

120 U0 160 180 LINE CURRENT (A)

Fig. 10. Ralativ« production yivld of th» 1.35 * 76m-8r at a function of th* ionizer (LaBe) tvmparatur«. This behaviour may be understood as the efficiency variation for a porous 1 ?? ) emitter described by Pelletier et al. ' , who showed that the efficiency

increases as a function of decreasing surface couerane followed bv a decrease

when, at high temperature, the diffusion of the halogen in the Dorous system

changes from surface diffusjon to volume flow.

So far, this ion source has been used in combination with DC -graph it e and Ijr-

graphite. cloth targets for the production of neutron rich , bromine

and iodine isotopes, and with a niobium target for the production of ncutron-

deficient bromine isotopes. for the former two taigets the ioni7ation

efficiency has been found to be < 1Z, decreasing with the time. This is

probably due to sur,:ai;e poisoning by carbon, abundantly ou tg

target, a s TO. The initial efficiency may be restored by heating the ioni?er to

1500-1600 C for some minute';. for the "cleaner" Nb target, an i on i/a I i on

efficiency for bromine of {7 3_* 2 *T 1X was found, in agreement with the

expectations when considering the mechanical layout of the ion source.

Additional elements should he within reach if other surfaces with lower work

functions can be put into stable and reproducible operation (P.XFf.p. 312).

4.2.3. Production yields.

The production yield is defined as the measured intensity at saturation of t fie

mass-separated beam in the collet tor l.ank. ami is expiessed, for ins '.HOI. e , by

eqn. ? in P.XI. Normally the yield;; arc measiireff by detection of the p-

ladiation from collected sources, but are occasionally measured with Faraday

riips (charged particle beams), single • hannel electron multiplier«: (charged or

neutral particle bPamsl, or by efciet t i on or .<. f and ilclaycrl neutron and

f rum the sources. The configuration of the most frequently

used exper imenta.l equipment for yield measurement <, if. given in fig. 11 (fig.

9 , pos i lion i] ) .

Production yield measurements «re the basis of the optimi/ation of running

pa r a met err, for various combinations of target, s and ion sources. Studies of decay curves and "y-speitra are necessary to examine the chemiral sci ci t. i vi I y .

Knowledge of yields and beam purity i •> nei «•;-, a i y j •» Hu- pi.win ing of

experiment1.: Ai" pui e beams required, m i;m Hie enpi'r IIUMII s he tail u>rl out w) Hi ,I i c-nlinniNl HIM' I •- I lie obtained yjnlil s"l'fii n-nl hi . nnn>1 ••< •• the

I'hppr i ment s wittiiri a r'sisonahl e I inn-' 37

Ion beam 2n[3- detector Toke-up wheel 4TT/3 - detector shield

2Tfp - detector :• Feed wheel Faraday cup

Single channel electron multiplier

Fig. 11. Configuration of the tape transport system and the detectors attached to the beam line for the measurement of production yields and product delay distribution.

This thesis (P.IX-P.XH) includes new isotopic yield determinations for a number of elements from various target/ion source combinations. A summary is given in table 3.

Table 3. Summary of target/ion source combinations for which the isotope production yields have been measured for selected elements.

Target Ion Ele- Mass region Reference source ment UC2-graphite Aa) Li 7-11 P.IX, fig.l cloth A Na 21-32(3'+) P.IX, fig.2 (P.X, fig.2) •• A K 37-53 P. IX, fig.3 and P.X, fig.3 " A Rb 79-102 P. IX, fig.4 A Cs 123-152 P.IX, fig.5 A Fr 202-230 P.IX, fig.6 and P.X, fig. 5 A Ga 76-78(85) P.IX, fig.8 " A In 114-130(132) P.IX, fig.9 •• A Tl 184-197 P.IX, fig.10 and P.X, fig.4 UC2-graphite Bb) Cl 38-43 P.XII , fig.2 •• B Br 76-94 P.XII . fig.l B I 137,140 _ ThC2-graphite A Tl 184-197 P.IX, fig.12 •• A Fr 207-228 P.IX, fig.13 ScCj-graphite A K 35-38 P.IX, fig.16 VC CO Ar 32-46 P.IX, fig.15 CaB6 C Ar 3<*-35 P.IX, fig-15 Nb-powder B Br 70-85 P.XII , fig.4 TeO2/KCl/LiCl C Sb 111-118 P.XI, fig.5 La (molten) A Cs 114-137 P.IX, fig.17 Pb (molten) C Hg 177-205 P.IX. fig.18 a)A = positive surface ionization, PSI b)B = negative surface ionization, NSI C)C = plasma ionization 38

The refractory carbides of uranium and thorium serve as "multipurpose" targets. Fission, spallation and fragmentation reactions produce appreciable quantities of elements from the lightest to the heaviest. Because the mass ranges for elements with low ionization potentials or high electron affinity are well separated, a high degree of element selectivity may be obtained when these targets are combined with a PSI or NSI source. For production of neutron-rich fragmentation products, a refractory tantalum-powder target is often preferred, thus avoiding the formation of the emanating a-radioactive isotopes.

The other refractory targets in table 3 have mainly been designed for extraction of selected neutron-deficient spallation products. The eutectic mixture of TeO -KCl-liCl (mp. -350 C) is selective for production of antimony isotopes when combined with a plasma ion source via a cold transfer line. The pure molten metallic lanthanum is selective for cesium when combined with a PSI source, for iodine with a NSI source and for xenon when combined with a plasma ion source via a cold line. The molten lead target has been extensively usPd for selective production of mercury beams.

A comparison of the production powers of the three available projectiles 600 MeV protons. 910 (890) MeV 3He2+ particles and 1000 (936)* MeV 12C** particles has been carried out in this work. Results are presented in the figures quoted in table 3, and as yield ratios normalized to the same particle flux and target thickness IP.IX, figures 7,11,14 and P.X, figures 7-H). Particularly for deep spallation and fragmentation products it has generally been found that the yield ratio Y( He)/Y(p)>1. For thallium isotopes with A<187, this ratio is in the range 10-100. This effect may be explained by the higher total energy transferred to the system which favours the evaporation of many particles. The same effect, though not so pronounced, has been established for the C * beam. The explanation may partly be that while the protons and He * particles lose only part of their energy in the target (P.X, table 1), the C-particles are totally stopped in the first few cm of a

*) The initial energy is degraded in the target container wall, and the values in brackets are the energy of the particles bombarding the target material. 39

normal UC -graphite cloth target. Accordingly, the average energy transfer 2 3 per interaction is less than for the He-particles, and the wings of the isotopic yield distribution are correspondingly lower.

For close spallation and fission products the higher energies of He and

C * do not favour the yields.

Production of elements heavier than the target by C irradiation has been demonstrated. and iodine have been extracted from a TeO - KCl-LiCl target (P.XI, fig. 6) and Na from a graphite target (P.X). However, attempts to produce isotopes from an UC -graphite cloth target were not fully successful, although some observed weak a - lines were tentatively assigned to americium. The low yield may be due to formation of refractory americium compounds.

In summary, the C beam does not at present seem «.o offer any new possi- bilities of sufficient interest for ISOLDE production targets. The results obtained with the He -beam, however, are promising. For production purposes the He -beam will become even more attractive with the planned higher beam intensity.

4.2.4 Delay properties A rapid release of the reaction products from the target system is desirable for a successful on-line performance. Knowledge of the delay distribution is also important in order to correct properly for decay when evaluating cross section data from production yield measurements. Therefore, a relatively detailed description is given here of the method used at ISOLDE for delay measurements.

Three fundamental quantities in release and delay evaluation are: a. The

A A wner fractional activity, F(t) = t' • ^ A is the activity of a particular product in a radioactive source after a heating period t and A the initial o activity of the same product, b. The delay probability function, p(t)dt, defined as the probability of finding a delay between t and t+dt, and c. The integrated delay probability, P(t) = / p(t)dt, where Pit) -» 1 when t •• •« and o is expressing the probability of a delay between 0 and t.

The fractional activity is directly measured in properly performed off-line release experiments where the release is governed by the rate of diffusion in 40

solids and/or by surface desorption rates. The diffusion in solids is

described by Fick's second law. The solution of this equation for uniform

spherical particles with an initially homogeneous ar.tivi.ty distribution as its

boundary condition, yield'., fur the single spherical particle .'i fractional ... .. 123) activity

n«, . •', r .', ,. k'*>1 n " k • 1 k '

where u if; the diffusion parameter, u • n D/i , I) the diffusion coefficient, i

the radius of the nucroparticlo and t Mie diffusion t inn-. This function, valid

for diffusion in single spherical particles, tias been successfully fitted to

the off line release curves, and the D values were then determined . This

may be done relatively safely beLHiive the very low mass of I fie soiinei

normally used in such experiments causes no measurable delay due to complex

bulk d.i f f us ion .

In on line delay experiment;: , hiwevcr, one nieae;ures the differential released

activity when collecting and counting on consecutive samp.! <;s . Here we observe a releasing process which .is originated in single spherical partjc.lt"; (as a model.) but which is also influenced by the t-.wyef rompa rtment, it", l:ub" connection to the ion '.our

q I" ' In) - 1 .1 a /A' n)

Hero A^ •• F. a; where A' - I|A| , n, bt.-jng I.ho 1 ran;.poi 1 i'1'fiiinmy. Hie

(luotient a /A' is a good approximation to p(t) at the time t iorrespondinj to sample i , and likewise the (|uotienl J u /T . a is

(.he integrated delay probability, P(t). Toi [lei i e,i •• i mj •. MI lE-t I inn 'unc. the approximation approaches the in f ini t «•; im.i J expii' •> s i on and

r ' (11 1 P11) {>.) wh i c h tjivHs

I'd I dl ' (t )/d( (b! l^^

Here the shape of F'(t) is somewhat different from the off-line Fit) due to more complex mass transport in the target/ion source system. The p(t) versus t-curve is, in principle, similar to those given in paper P.VIII, figs. 7 and 8. The p(t) value increases from zero to a maximum with a subsequent decrease towards the zero level again.

23 ) [n previous on-line studies of the release properties , the function (2) has been taken to represent the actual fractional activity, without any reservation as to its correctness in principle. The value of - dF(t)/dt is monotonically decreasing and incapable of reproducing a maximum. Therefore, it can not represent the true p(t)-function. However, in rapid systems it may still he used as an approximation.

Since it is not feasible to perform a single proton pulse irradiation (duration < a few milliseconds I at the ISOLDE targets, p(t) has to be derived indirectly. The delay measurements are carried out as shown by fig. 12. After a constant ion beam production rate has been obtained, the proton beam is switched off the target and the intensity of the ion beam followed until the background level is reached. The proton beam may then again be switched on and the intensity of the ion beam recorded until equilibrium (constant intensity) is obtained.

[ PHOTON BEAM ION BLAH :•• OFF OH OFF I I ] 1

- TV (count s

V... - NTENS I

in' 450 500 TIME (s)

Fig. 12. On-line delay curve of sodium (24-Na) from an UCs-graphite cloth target at about 2000 °C, accumulated with a single channel electron multiplier (particle counting). The curve must be correc- ted for the general background, of the probe nuclide and the background caused by electron multiplier detection of beta particles from the decaying nuclide. This background can be accumulated after having switched off the ion beam, as shown in the figure. 124 ) An extensive mathematical description has been given by Rudstam for collection of ions on a stationary collector. Assume a constant production rate R in the target for the probe nuclide in the time interval t=0 to t=t P ° and that parent effects may be neglected. Let the collection/counting period At be short compared to the half-life of the probe nuclide. Then the number of collected ions during the period from t to t+At (t>t ) is: o

S(fAt) = R ne J° e'*11*"72"1'' ( / p(t")dt")df (6) P ° Here X is the desintegration constant for the probe nuclide and e the total detection efficiency. If At is short enough so that p(t) may be considered constant and equal to p(t+At/2) over the collection period, eqn. 16) simpli- fies to t

S(t*At) = R neAt J e-Mt*At/2-t >p(t+At/2_t•)dt• (7) P

The number of collected ions in the period At during production ("beam on" at t=0) is

t .A(t.+At-/? f) t + At-f S(fAt) -- R ne J e AlC flt/ ' ( J pit" )df )df (81 t-f

XAt/2 tt4t t*^-t' * R nee J ( J plfldt'ldf P t t'-t The first term corresponds to the contribution from the production before collection start, and the second term to the production during collection. If p(t) may be considered constant over At, eqn.(8) may be approximated by

A(t + At/2 t) S(fAt) = RpneAt J e" "" p(t + At/2-t')df (9) o

2 AAt/2 • R ne At e" p(At/2)

The p(t)-function may be represented by a polynom. The shape of p(t) may then be derived (variables in the polynom determined) by computer fit of one of the eqn.s (B)-(9) to the appropriate part of the experimentally recorded delay curve. Eqn.s (5) and U) then yield F'(t) and P(t). 43

The release performance of a target is often described in terms of the "release half-time". It is defined as the time needed from "beam off" for the ion beam intensity to reach half its initial value. This number only gives information on whether or not there is an appreciably fast component in the release process. In order to compare different target systems for their total release behaviour, we have found it convenient to illustrate the integrated delay probability, Pit), as a function of time on probability paper. From the resulting curve one may directly extract the time necessary for the release of a given percentage of the sample activity.

The fractional yield corresponding to the delay curves in fig. 12 is, however, shown in paper P.X, fig. 6. The time for release of one half of the initial activity is 1-2 s. A similar fast release component has been observed for the other alkali elements under the same conditions, while preliminary measurements indicate a different and slower release for thallium.

4.2.5. Stable impurity beams Although high-purity materials are used in the construction of targets and ion sources, one often experiences a large contribution of impurities from various elements emerging from the target units. The identification of the resulting beams in the mass separator is important in order to understand the performance of the targets so as to make possible an optimization of the output of the desired nuclide and minimize the disturbance from other species. For instance, in measurements of delay-time distributions by means of an ion detector (single channel electron multiplier), the measured mass must be practically free from such stable-beam contribution. Knowledge of the impurities may also assist in explaining why a particular target does not work according to expectations. Two examples are:

a. A production target for chlorine isotopes resulted in very low yields. It was found that a high calsium content led to the formation of molecules like CaCl . A substantial part was not cracked in the negative surface ionization source, and escaped ionization and acceleration. By using a plasma ion source, however, ions like CaCl and CaCl were produced but the six stable isotopes of calsium will distribute a chlorine isotope over several masses, b. In gamma-spectroscopicexperiments on heavy francium isotopes confusing results were obtained until it was realized that a considerable amount of stable was introduced into the system. Molecular ions with radioactive reaction product bromine isotopes of the type BaBr had been formed, thus mixing directly into the heavy francium region. 44

In several cases such problems may be reduced by a proper adjustment of target and ion source parameters. Highly undesirable impurities may in some cases be evaporated away by an off-line heat treatment under introduction of macro- amounts of stable material of the element of interest.

Examples of mass-scans of such stable impurity beams are given in tables 4 and 5. The tables contain the beams emerging from three different UC -graphite cloth targets, one coupled to a positive surface ion source and the two others to a plasma source. The very much cleaner conditions with the former source are striking.

It may also be seen that two similar targets do not necessarily give the same stable beam mass spectrum, probably due mainly to the individual treatment in the mechanical construction and mounting operations.

Table 4. Stable "impurity" beams eme.-ging from

a UC2-graphite cloth target combined with a surface ionization source.

Mass Relative a) Tentative intensity assignment

10 23 7.1-10" "r:a+ 27 0.56 " "A1 + 39 1.9 " "K+ 40 1.5 " »Ca+ 52 0.31 " "Cr+ 59 P. 31 " 5'Co+ 89 1.0 " 8 9 Y+ 136 2.6 " '"Ba+ 137 4.4 " ls7Ba+ 138 30.3 " 138Ba+ 139 14.4 " '3'La + 140 2.8 " 7 + 142 1.3 " '- m 144 0.63 "

146 0.50 " ncNd+ 165 0.50 " "•5Ho+ 219 1.4 " ie.Ta]5F+ 238 580.0 " 248 2.0 " ? 257 280.0 " 276 2.7 "

Target paraneters: 800 A, 1.4 V, corres- ponding to about 2080 °C. Ion source narameters: 380 A, 1.1V» cor* responding to about 24 00 °C. Table 5. Stable "impurity" beams from a UC2-graphite cloth target coupled to a plasma ion source.

Mass Target 13 Target 1 Tarqet 2 Target 2 Tentati ve Relative in tensi ty ReKint. after 3 Rel. int. Rel.int. assignments days of operation Cold line Hot line

1 6.0-10"' 5.6-10-' _ _ 2 1.5 " 4.2 " - 4-10"' 6 5.4 " 1.8 " - - 7 0.54 " 0.53 " - - "•N'\ 7Li + 8 1.1 " 1.8 " - - 10 15.6 " 2.5 " 3-10" 9.5 " "B* 11 53.8 " 9.8 " 19 43 " nB+ 12 161.5 " 60.0 " 68 " >170 " 12C + 13 1.5 " 0.68 " - 2.5 " uc + 14 23.7 " 31.5 " 18 " 38 " >"N+ 16 78.1 " 52.5 " 37 " 89 " 16Q+ 17.5 5.6 " 1.1 " - 4.5 " 18 0.54 " 0.53 " - - :•,(,+ 19 1.9 " 0.30 " 22 21 "F* 20 - - 10 31 " 23 39.8 " 3.8 " 48 " 118 " 23Na + ?4 - 12 69 " 2"Mg+ 25 - - 2.5 " 8.5 " 2!Mg + 26 6.5 " 2.0 " 4.5 " 13 26Mg\ !3BI!O* 27 32.3 " 13.4 " 17.5 " 41 " 27A1+ , ^B'-O* 2b =•200 " >200 " ^•170 " •-170 " 29 - - 34 " 85 " 13B13pt 33 5. 'j 5.6 " 54 " 114 nBisF* 31 5.1 " 1.2 " - 1.5 " !lp+ (?) 3? 3.7 " 0.75 M - 4 "0: 33 3.4 " 0.75 " - -

35 89.4 " 19.8 " 33 70 " i5Cl* 36 95.3 " 20.0 " 5 8.5 " 37 37 38.8 " 6.8 " 11 24 " cr 37.5 - - 16 36 -°Ca'5Cl * 3« 28.0 " 6.9 " - - >Hi7Cl* 38.5 - - 5 18 " "•°Ci!7Cl + 39 12.9 " 6.7 " >I7O " >170 " S'K+ 40 - - 45 108 '•nCa* (?) 41 2.2 " 7.8 " 33 97 B2Kr-\ "Ca* 42 5.4 " 29.3 " - - >°B;'0t , «-Kr:+ 43 6.0 " 12.1 " - - llBl'OJ , a6Kr* 44 48.5 " 8.6 " 97 >170 12C"0J , 1<-Njc0+ 45 3.2 " 0.75 " 35 " 69 1CB'5C1+, 2!Silf0* 46 8.1 " 1.4 " 139 " > 170 llB1bCl , 2 Al' F+, 3 °Si16O' 47 1.1 " 0.15 " 27 " 62 " 1O8"C1*. 1!C5'-C1* 48 2.3 " 0.38 " 21 " 25 " B • C1 » B F - l; 49 - - 82 " 97 " "B'-FJ , C"CT*

51 0.32 " 0.15 " - - i.V* 52 2.4 " 0.38 " - - iiB"0< 53 8.6 " 1.1 " - - '^"B^O* 54 1.8 " 0.45 " - 3 >'B160*

55 16.1 " 12.2 " - 2 •• sMn* 56 7.5 " 0.54 " - - '''Te* 5S 9.9 " 0.30 " 8 10 "'K"F* (?) 5

Continued on the following page. Table 5 continued.

Mass Target 1 Target 1 Target 2 larqet 2 Tentative Relative intensity Rel.int. after 3 Rel.int. Rel.int. assignments days of operation Cold line Hot line

3 9 61 64.6-10' 11.7-10" 7-10"' 10.5-10"' 211Mg"Cl + , 26Mg''cr 62 >200 " 42.0 " 24 34.5 " 27A1'5C1+, 25Mg37Cl+ 63 24.8 " 4.1 " 3.5 " 7 26Mg37Cl+ 64 80.8 " 14.1 " 7.5 " 9.5 " 27A117C1+ 65 - - t 1.5 " 305i!5Cl+, 23Sis7Cl* 66 1.1 " - 26 " 50 2!Si37Cl+

67 - - 1.5 " 4 7cSi!7Cl* 69 2.8 " 0.83 " - - ? 70 3.4 " 1.4 " - - ?

75 - - >170 " >170 " "'Ca^Cl*

77 - - 95 >170 " '"'Ca'7Cl + 79 - - 7 12.5 " ""Ca^cr 80 1.1 " 13.4 " - - aoKr+ 81 - - 2 3.5 "kCa37Cl+ 82 5.4 " 66.0 " - - '-Kr* 83 3.9 " i - 1.5 " 83Kr+ 84 26.9 " >200 " - - 9"Kr+ 86 7.8 " 90.0 " - - 96Kr+ 90 8.6 " 5.9 " - - 5'Mn'5Cl+ 91 4.5 " - - - 56Fe35Cl+ 92 3.2 " 2.0 " - - 55Mn!7Cl* 93 1.8 " - - - 5'FeJ7Cl+ 94 - - 4 7.5 " "•°Ca35Cl19F+ 98 5.5 " 0.15 " - - ? 100 4.1 " - - - ? 102 0.32 " - - - ? 110 - - 19.5 " 28 112 - - 14 19.5 " "°Ca35Cl37Cl+ 114 - - 2.5 " 4 *°Ca37Clt 124 6.5 " 4.8 " - - 8»r"ci + 37 + 126 3.0 " 1.1 " - - B 9y C1 127 2.0 " 0.68 " - - ? lel + 181 0.13 " 0.45 " - - Ta 197 0.33 " 1.4 " - - ieiTal60+

a. The target and ion-source parameters were: Target: 820 A, 2.8 V, corresponding to a temperature of approximately 2130 °C. Ion source: Anode 1 0.18 A, 200 V; Anode 2 0 A, 115 V; Magnet 5 A, 19 V; Elec.bomb. 0.9 A, 250 V. b. The target and ion source parameters were: Target: as for point a. Ion-source: Anode 1 0.1 A, 200 V; Anode 2 0 A, 35 V; Magnet 5 A, 19 V; Elec.bomb. 0.35 A, 235 V. c. The target and ion-source parameters were: Target: 940 A, 1.4 V, corresponding to a temperature or approximately 2200 °C. Ion source: Line cold; Anode 1 100 V; Anode 2 100 V; Filament 28 A, 5 V. d. The target and ion-source parameters were: Target: 920 A, 1.4 V, corresponding to a temperature of approximately 2180 °C. Ion source: Line 300 A, 1.4 V, corresponding to a temperature of approximately 2000 °Cj Anode 1 101 V; Anode 2 100 V; Filament . 8 A, 5 V. These experiments have not been designed to localize unambiguously the origin of each of the species. Yet the main part is expected to originate from the target material itself, and some beams largely from the ion source.

The information from such scans is capable of being used in the chemical evaporation technique treated in the next section.

4,2.6. Chemical transport techniques. About 50 elements are presently available for experiments as mass separated beams at ISOLDE. About 25 of these are obtained relatively free from contami- nation by neighbouring isobars, while some improvements are needed in the target/ion source technique in order to separate the other 25 from such con- tamination . In addition, several elements obtainable as nuclear reaction products have not yet been converted into on-line beams of sufficient strength mainly due to low vapour pressure or complex rhemical behaviour.

By means of chemical reactions it is possible to make elements from this last cats-gory available as on-line beams. Reactive gases or other reactive chemicals may be added to the target. In-situ chemical reactions form voiatili? compounds of otherwise not sufficiently volatile nuclear reaction products. The molecules formed are transported to the ion source. This process is called "the chemical transport technique" ' .

This t t'(. hriiqufi has lung been known as a us.eful too] in chemistry and thenutal physics ), but until recently it was given relatively little attention ,iri connection with mass separation. Even if the possibilities were stated several years ago ), only some scattered investigations have lieen report- 12 9 130 eci ). However, the technique has now started to become i e

There exist several possible reactive -containing gases which in the order of general decreasing reactivity (as judged from their Gibbs energy of formation) are F > SF > CF > HF > BF . In the evaporation of a highvalent 2 6 A 3 element, however, the use of SF may appear more advantageous than the use of 6 F , although this may depend on the vapour pressure of the reactive gas, the actual reaction temperature and the element of interest. The general rule is that the evaporation selectivity will be determined largely by the reaction thermodynamics, i.e. the most thermodynamically stable product species will predominate. For several actual target element/product/reactive gas combinations it is possible to estimate the evaporation behaviour by thermodynamic calculations, but for a number of elements the thermodynamic data are inadequate, incomplete, or not available at all (see tht- review article by Hildenbrand ). In such cases the target development has to proceed via empirical knowledge, and the target chemist may obtain useful information from general studies of the thermodynamics of thermochemistry gasification. The references 134-144 and references quoted therein may serve as an entrance key to this area.

Systematic investigations in this field have recently been started at ISOLDE, where the first goal is to rtfine and enhance the production of fluorinated compounds of the alkaline earths, yttrium and the rare earths. Other elements of interest are the light elements like , sulphur and chlorine, the refractory medium-heavy elements like zirconium, niobium, molybdenum, techneti urn, ruthenium and rhodium and the heavy elements like hafnium, tantalum,tung- sten, rhenium, and irridium. Some of these may be separated as volatile oxides.

A few results from test experiments are shown in table G. These experiments have investigated the effect of fluorine addition on stable impurity beams under "normal" running conditions (more detailed infmma I ion may be found in ref . 1 45 ). The fluorine is added by heat decomposition of a small piece of the plastic material Teflon (KF -CF-I ) placed in a thin tantalum tub« leading Table 6. Stable beams emerging from a UC2-graphite cloth target coupled to a surface ionization source, and the effect of fluorine addition.

Mass Ion type Rel.int. with- a' Rel.int. with a' assignment out F-addition F-addi tion

23 !3Na+ 54-10"' 97.5-10"9 27 27A1 + 41 " 58 39 35K+ 34.5 " 57.5 40 *°Ca + 9 16 41 «iK+ 2.2 " 3 48 *'Ti + 2 10 51 4 23 52 52Cr+ 8 13 56 ">Ca16O+ 2.2 " 4 89 B9y+ 58 " 143 105 B9Y160+ - 3.3 108 8 9y2 9p-f 5 63 127 89YHF + - 7 139 139La+ 180 " 180-10'7 155 180 " 96 158 13'La19F* 180 " 116 165 - 6 48 -10"' 171 l3'La160* 9 113 177 139U19p+ 61 44.5-10"' 238 238U+ 180 " 108 254 238(jl60+ 180 " 32 257 238Ml9F+ 180 " 180 270 23a"l60+ 8.5 " 69 -10"9 273 !3BUl9F16O+ 26.5 " 7 MO"7 276 40.5 " 44

a'Target temperature •*• 2000 °C, and ion source temperature ^ 2400 °C. directly into the target container. The addition is efficiently controlled by adjustment of the heating current.

All the recorded stable beams show an increased intensity after the addition of fluorine with the most substantial enhancement in the fluorine containing ions of yttrium, lanthanum, and uranium. It is interesting to note the large increase in intensity of ions containing oxygen. This increase may be due to the cracking of the transported fluoride compounds while in the ion source with subsequent reformation, now incorporating oxygen (CO is always present as a major constituent of the stable gases diffusing from the target, see table 5).

The resulting mass spectrum is of course a function of the amount and type of the reactive gas, the reaction temperature, and the ion source parameters. This last point is demonstrated in fig.13 with the production of 8a F*. A maximum yield is obtained at an ion source temperature of -2400 C. 50

TEMPERATURE (°C) 1750 2000 2250 2500 2750

600 -

250 300 350 «00 450 LINE CURRENT (A)

Fig. 13. Relative production yield of the fluoride sideband BaF+ as a function of the ionizer (tungsten) temperature, using the 2.55 min 137m-Ba as the probe nuclide.

Off-line release tests have, after the present work was terminated, been 1461 continued at FSOIOE, and have recently given promising results with respect to vaporization of non volatile elements. Addition of CS -gas -1,-3 (pressures of 10 - 10 torr) while keeping the target material at about 1900 C, has strongly enhanced the liberation of, for instance, Zr from Ru, Y from Ru/graphite, Yb, Lu, Tm and Hf from Re/graphite and La and Ce from UC /graphite. There is good hope for a successful online performance of several of these new target systems, but this has yet to be shown in practice.

1,3. Applications of short-lived nuclides.

4.3.1. Non-destructive H MeV neutron activation analysis. Non-destructive, also called instrumental, neutron activation analysis UNNA) with H MeV neutrons produced by the reaction "H(d.n) He, has become a well established analytical method since the first report in 195S . Its success as a universal method for most of the elements in the periodical system is largely due to; a. the development of sealed-tube neutron generators and b. the invention of semiconductor detectors with high energy resolution. Condition a. makes the operation of the generator safer (no leaks of tritium from the target) and considerably simpler than for the older generator types with exchangable tritium targets. The neutron output is of the order of 10 - 11 10 n/s, and as nearly constant in time for several hundred to several 51

thousand hours of operation. Condition b. makes possible a quantitative determination of several elements simultaneously from a relatively complex mixture of reaction products. However, when no ambiguity can arise, a simpler and cheaper system based on Nal(Tl) scintillation detectors may be preferable due to higher detection efficiency.

The present thesis is not the correct forum for discussing in detail the general features of the technique, its area of application, and its limitations. Such aspects are treated in a number of informative review articles, of which a selection is represented by references 149-162. Here, a few brief comments will have to be sufficient: 14 MeV INNA is generally not practical with irradiation times longer than about 30 min per sample. Consequently, the method is best suited for short-lived reaction products with half-lives ranging from a fraction of a second to about 100 min. The shortest half-lives require automatization of the sample handling after irradiation. The detection limits range from about 10 pg to 1 mg for the different ele- ments, and sample sizes up to several cm may be irradiated and analyzed reli- ably. Today, several well equipped 14 MeV INNA laboratories are installed in various countries, many serving the general analytical needs on a commercial basis.

Paper P.XV describes the 14 MeV INNA facility constructed at the Department of Chemistry, University of Oslo.

The construction of this equipment had several motivations, one being the need for fast and reliable oxygen analysis in various matrices including pure metallic samples (i.e. aluminium). The reaction 0(n,p) N {t =7.13 s, E =6.13, 7.12 MeV) offers such possibilities by detection of the high-energy •y-rays from decay of N.

The analysis system roughly consists of the following main parts: A sealed tube 14 MeV neutron generator (Philips PW 5320), a pneumatic sample transfer system, conventional detection equipment based on two 3"x3" Nal(T1)-detectors, and a facility for packing the samples in inert atmosphere (P.XV, figs.1,4 and 5).

The transfer system is constructed for cylindrical sample containers, "rabbits", with outer dimensions r=6.40 mm, 1=55 mm. It consists of a single transfer tube, a biaxial rotation device for the sample in irradiation position, a combined loading/unloading gate and a retardation mechanism (air cushion) for soft landing of the sample in the counting position. For metal rod samples packed in polyethylene containers, the loading gate also operates 52

as a mechanism for separation "in flight" of the sample and the container prior to activity measurement.

The analytical sequence may be executed automatically.

Various factors which may affect the detection limit and the analysis accuracy have been examined:

The standardization method. The oxygen content in normal pure aluminium is often less than 10 ppm which is close to the present detection limit for the method (> 2 ppm). Hence, it has been found sufficient in routine analysis to calibrate the procedure by a secondary standard. It consists of an aluminium sample of normal size (r=6.25 mm, 1=50.8 mm! whose oxygen content has been determined by primary standards of perspex, oxalic acid/graphite, and steel samples with certified oxygen content.

For accurate analysis of samp.les with higher oxygen content, more sophistica- ted primary standards have been devised, taking into account neutron and i-ray attenuation corrections for differences in standard and sample shape and in composition Inormally < 2 '/. correction), or standards which essentially allow such corrections to be neglected.

Procedure for sample treatment and packing. The samples are prepared first by controlled machining of the surface, and then by chemical etching in a strong HF/HNO mixture. After this treatment the initial surface oxide thickness is reduced to approximately 0.2 ug/crn of oxygen. The samples are sealed in polyethylene containers in nitrogen atmosphere.

Optimum fD?rgy__range__for__integration. During irradiation, the matrix activity Na 't./^15 n' i-s formed in the reaction AKn.a) Na. The two coincident t-rays at 1369 keV and 2754 keV in the decay of Na give rise to a sum peak at 4123 keV. Pile-up events due to accidental coincidences between ? 7 the sum peak and the general matrix activity, including Mg(t =9.46 min), form a high-energy tail on the sum peak stretching into the region of interest 1 6 for integration of the N-activity. The contribution from the tail is related 27 ?4 linearly to the ' Mg-activity and exponentially (second order) to the ' Na- activity. With the expression for the relative standard deviation as the optimization criterion (P.XV, table 3), the optimum position of the integration limits has been calculated. The limits (mainly the lower one) vary with the oxygen content and the ntimbpr of rt'plicatp i rrariLttions on the same sample ! P . XV. fig. 11). On the average , the lower limit lies between 4.3 and 4.5 MeV and the higher une at 7.0 MeV.

Optimum combination of .irradiation, clecay_ and counting times. With the expression for the relative standard deviation as the optimisation criterion, (P.XV. table 3) the optimum combination of irradiation and counting times, t. and t , has been calculated for various oxygen contents and numbers of replicate irradiations, the decay time being fixed at 1 <• . Results are given in P.XV, firj.1?. For samples of pure-aluminium one may recommend average "reasonable" values of t. 10 s ane) *.-?s s. aricl '"* replicate irradia- tion/counting sequences.

The described procedure is now well established for analysis of oxygen in metallic samples. With small modifications it has as well been applied to other matrices. The technique is serving research institutions and industry.

For samples with oxygen contents below 1-2 ppm reliable analyses are offered by charged particle activation (CPAA). The detection limit is of the order of a f^w ppb. CPAA also makes possible depth and lateral concentration profiling of the oxygen distribution in irihomogeneous samples. CPAA-techniques are at present under evaluation and establishment by the author at the Cyclotron Laboratory in Oslo. They are thought to be a valuable supplement to the 14 MeV INAA technique.

4.3.2. Medical racjionucj.ide production. The majority of the radionuclides uti.li.«d in nut lear medic ine has tradi- tionally been produced in reactors. Examp] •• , of such nuclidps are P. Ti. , I. ' Xe and ' Au. Among these, " Tc is. by far, the most versatile. It may be chemically incorporated in various organic molecule!, which have different physiological functions, and may thus be used for a number of different examinations. It is tho "working horse" in nuclear medicine departments today, and conventional detection equipment has bpen optimised for its photon energy of 140 keV. In recent years (after 1960) radionuclides produced by charged particle reactions have come into extensive use due to favourable radiation type and energy and suitable half-lives, often combined with specific physiological behaviour.

There are two classes of accelerator-produced radionuclides in nuclear medicine today: 1. Positron-emitting nuclides that are used with the positron emission tomograph and 2. photon-emitting nuclides which are used with the gamma - camera.

The former group has so far mainly been used in the development o-f a promising approach to the in-vivo study of regional physiology - Examples of utilized nuclides are C, N, 0 and F. The latter group has achieved widespread C *7 clinical applications and among the most frequently applied nuclides are Ga, 7?Br, 81mKr, 111In, 123I, 127Xe and 201Tl.

With the installation of the Oslo Cyclotron (Scanditronix MC 35) in 1980, it became possible to start development in Norway of production methods for short-lived medical radionuclides formed in charged particle reactions. The maximum projectile energies available are for protons and a-particles 35 MeV, for deuterons 18 MeV and for He-particles 47 MeV. The external beam currents are generally < 50 pA. These characteristics are sufficient to produce a number of medical nuclides in high yields.

A schematic overview of the Cyclotron Laboratory is given in fig.14. The high- intensity irradiation area is marked on the figure. The irradiation position has been equipped with a special beam tube section with (he necessary vacuum pumps, a collimation chamber with an electrically isolated and watercooled adjustable aperture, and a fluorescent screen for the monitoring of the beam profile. A "snap-on" flange on the beam tube facilitates the exchange of target chambers. 55

y£-^:/^N:^-\'.!:VO:v.W"^

13

; l n «tn^.r.;T.--.^*.» A-.-".v..r>'?.', £23 i-PT'.'.wf.r^Tn i-^^.^-Mj'?i|l

1. CYCLOTRON UNIT 11. ELECTRONIC LABORATORY AND PREPARATION ROOM 2. OUADRUPOLE MAGNETS 12. POWER SUPPLIES 3. HOT CELL 13, VENTILATION AND WATER 4, BENDING MAGNET COOLING 5. ISOTOPE PRODUCTION in. ENTRANCE AND IRRADIATIONS 15. CONTROL DESK 6. SWITCHING MAGNET 16, DETECTION ELECTRONICS 7. SCATTERING CHAMBER 17. COMPUTER WITH ACCESSORIES 8. MINI-ORANGE SPECTROMETER 18. COMPUTER TERMINALS 9, POSITION FOR MULT I- PARAMETER MEASUREMENTS 19, OFFICE 10. CHEMISTRY LABORATORY

Fig. H. A sketch of the Oslo Cyclotron Laboratory. 56

Included in this thesis are methods which have been developed for the production of G7Ga (P.XVI) and "Vr (P.XVII). A brief review and discussion is given below.

The radionuclide Ga has a half-life of 78.?6 h and decays by pure EC followed by -f-ray emission with main energies of 93.3, 184.5 and 300.2 keV which are suitable for detection with conventional gamma-cameras. Complexed with citrate, Ga has gained a reputation as a soft, tissue tumor localizing agent, and its use as a diagnostic tool in nuclear medicine departments had shown a steady increase in the years prior to the start of the present work. Examples from the extensive literature on the medical use are found in references 165-170.

67 Production of Ga may be achieved by proton or deuteron induced reactions in ?inc, or by a- or He-particle induced reactions in copper. Based on results reported in the appropriate litterature and on experimental cost and R ft fi 7 convenience arguments, we decided to apply the reaction Zn(p,?n) Ga using a zinc target of natural composition 118.6 I Zn). The excitation function was measured between 14 MeV and ?9 MeV by the stacked foil technique, and the thick target production yield determined for various bombarding energies (P.XVI, fig 4). A comparison with production yields reported in the literature is shown in table 7. Table 7. Thick target production yield of 67-Ga from the reaction 68-Zn(p,2n)67-Ga at the end of bombardment with 29 MeV protons on zinc of natural composition.a'

Hupf and Beaver, (ref. 171) 1.1-1.4 mCi/yAh Brown et al., (ref. 172) 1.4 Krasnov et al., (ref. 173) 2.2 Little et al., (ref. 174) 1.18 ±0.15 This work 1.6 ±0.4 t*)

a ) ' The original numbers have been recalculated as thick target yields using the shape of the excitation func- tion determined in this work.

b'The quoted error limit i* mainly due to uncertainty in the beam current measurement.

During irradiation of natural zinc, the radionuclide f>6r,a(t =9.3hl is also 66 formed in high yields chiefly in the (p, n)-reaction on ?n (27.8 7.). A target thickness, of 1.3 mm j«, optimal for ?9 MeV protons in order to keep the contamination of Gad 1. after a decay time of 92 93 h. In order to achieve! 57

sufficient cooling during the irradiation with high beam intensity, the zinc disc target is fixed to a copper plate by a thin alloying layer at the contact surfaces. The copper plate served as a window between the vacuum system and the watercooling which is applied directly on the back of the plate (P.XVI, fig. 7).

The chemical separation of gallium from the irradiated zinc target was achieved by cation exchange from 7.5 M HC1, subsequent elution with 3M HC1 followed by evaporation to dryness and dissolution of the evaporation residue in isotonic saline citrate solution at pH 5.5. The procedure also contained a decontamination step for iron which has been reported to disturb the physiological function of Ga-citrate . The total separation time was 4-5 h. After this procedure the content of zinc and iron was less than 2 ppm. Following a general sterile filtration and autoclavation procedure, the Ga- citrate radiopharmaceutical has been tested in animals and subsequently applied in human investigations.

During the development period the use of Ga-citrate in Norwegian hospitals decreased from about 45 mCi/week to less than 10 mCi/week, thus removing the economic basis for a routine production at the Oslo Cyclotron.

Rb. Radioactive noble gases have been used in diagnostic nuclear medicine for nearly 30 years. The soft 0-emitter Xe (t = 5.29 d, Ey= 81 keV) has been, and is still, the most frequently applied radionuclide, mainly due to a simple production procedure and a relatively long half-life which im- plies easy distribution and long shelf-life in the nuclear medicine departments.

However, the diagnostic advantages of the short-lived radionuclide Kr

(t1-=l3 s) has been increasingly emphasized during the last 10 years. It is formed continuously in the decay of the mother nuclide Rb (t =4.57 h). It decays by internal transition with emission of a single -y-ray of energy 190 keV. This energy is nearly optimal for detection with conventional gamma- cameras.

Some of the advantages of Kr as compared to Xe are: a. Better picture quality, b. smaller radiation dose (P.XVII,table 1), c. applicability in combi- nation (simultaneously) with a mTc examination, d. frequent repeatability for pictures in multiple projections, e. possibility of studying dynamic func- tions and f. no waste disposal problems. 58

So far Kr has been applied most extensively in lung ventilation examin- ations where patients inhale the gas mixed with air or oxygen, but also in heart, brain and vein examinations where a Kr-containing 5 1. dextrose or glucose solution is infused in the blood streem. As an introduction to the literature in this field see reference 176 and paper P.XVII, references 8-K.

81 81m Due to the relatively short half-life of Rb, the general use of Kr gene- rators is so far limited by the demand for a suitable production facility, i.e. a particle accelerator, in close vicinity to the medical centres. To Aim serve the needs in Norwegian hospitals a technique for the production of Kr ventilation generators has been developed at the Oslo Cyclotron. This technique is described in detail in paper P.XVI I. 81 With the available particle energies, Rb may be produced by proton and deuteron bombardment of krypton and by a- and He-particle bombardment of bromine (P.XVII, table 2). We *~ave chosen to apply the reaction 8 2 81 8 2 Krtp,2n) Rb, using krypton gas of natural composition (11.6 I Kr), and the proton energy range 30-1* MeV. At 10 bar gas pressure this implies a proton path length of 30 cm. There will also be some contribution from the ft 1 fl 1 fit reaction Kr(p,3n) Rb HI.5 1. Kr). Details of the stainless steel target chamber is given in P.XVII, fig.4. A 25 um thick stainless steel foil (Goodfellow AISI 302) separates the highpressure target gas from the vacuum in the beam tube. The rubidium which is produced during irradiation is deposited on the chamber inner walls from where it is leached by demineralized water. The rubidium is subsequently collected on organic cation exchangers. The absorbtion efficiency of rubidium on a 15 mm long column of Dowex 50x8, 100 - 200 mesh, is close lo 100 I at a flow rate of 10 ml/min. The gas elution efficiency of Kr has been measured to - 84 I for flow rates > 5 mils, in accordance with previously reported results (P.XVII, table 4).

In the proton bombardment of krypton the high-spin (9/2*) isomer mRb ft = 31 min) is also formed. For the energy interval 26 - 17 MeV we have measured the production rate ratio for the ground state to the isomeric state to b« in average R /R = 0.94 *. 0.10. Since th« decay of 61mRb proceeds with 97.6 I to g m the ground state, this feeding must be taken into account whan the generator strength is calibrated. Otherwise one may underestimate the strength at the user time by several tens of percent. 59

This effect is still more pronounced for a- and He-particle bombardment of bromine-containing targets. Preliminary experiments indicate a ratio of R /R g m = 0.4 - 0.5. This strong feeding has not been mentioned previously, and it is unclear whether is has been accounted for in the literature. 81 An alternative, continuous, on-line production method for fib, not previously reported, has been examined: The organic liquid bromoform, CHBr , has been irradiated in a flow target with 40 MeV He-particles, and the rubidium produced has been stripped by water according to the SISAK-technique described in section 4.1. The water was continuously pumped through ion exchanger columns to retain the rubidium while the bromoform after stripping was recirculated to the target chamber. A production yield > 3 mCi/pAh and a chemical yield > 80 I was obtained. However, in long term routine operation the beam intensity must probably be kept below 2-3 uA due to radiolysis and the formation of polymers in the target liquid. Accordingly, this production method is probably no realistic alternative for large scale production of Rb. Further details of the methods will be published elsewhere.

With targets of natural isotopic composition, considerable amounts of fl "} W T ft L Rb are also formed. While these nuclides are harmless in normal 81 elution generators, their presence restricts the use of Rb as an internal generator where the radiorubidium is incorporated in the living tissue by 81 injection. Very pure Rb may be produced by elctromagnetic mass separation, for instance according to the ISOLDE technique. In 600 MeV proton induced spallation of niobium with a beam intensity of 2 uA, the saturation yield of Rb is -10 atoms/s (P.IX, fig. 4). A source strength of 20-25 mCi, corresponding to a normal generator strengt, is obtained by collection of the 8 1 mass-separated Rb beam for 30-40 min. We have started preliminary a 4 feasibility studies for the production of pure and sterile Rb solutions with the ISOLDE technique. 60

5. SOME RESULTS FROM NUCLEAR INVESTIGATIONS 5.1. Spectroscopic measurements on some neutron-rich light lanthanide iso- topes . The neutron-rich light lanthanide isotopes constitute a transitional region between spherical and deformed nuclear shapes. 8y a systematic study of the nuclides in this area, it becomes possible to follow the changing nuclear properties.

This area of the nuclear chart was, until recently, scarcely studied, mainly due to the lack of proper separation and identification methods. The SISAK technique offered new experimental possibilities, and the neutron-rich light lanthanides became a region of special experimental interest.

Chemical separation procedures were developed for selective isolation of lanthanum, cerium and praseodymium from complex fission mixtures. The separations are based on the extraction of the lanthanide ions with the complexing agent HOEHP (bis-2-etnylhexylorthophosphoric acid) in an organic solvent (Shellsol-T) from nitric acid solutions. Detailed separation schemes optimized for La are found in P.Ill, fig. 1, for La in P.IV, fig. 1, for Ce in P.V, fig. 1 and for Pr in ref. 93, fig. 3.15. Thus, the investigated praseodymium isotopes and La are not the primary formed ones, but those formed in decay of their mothers. The rest of the lanthanum isotopes and the isotopes of cerium are mixtures of primary formed and decay - generated species.

The spectroscopic measurements have been performed by f-singles and yy- coincidence detection with high-resolution Ge(Li)-detectors. Half-lives and partial decay schemes were established and a few spins tentatively assigned to low-lying levels in some of the decay schemes. Detailed results are reported for '"La in P.III, for U*-U6La in P.IV, for '"'""L, and U9-15°Ce in

U5 U8 M7 150 P.VI, for - Ce in P.V and for - Pr in p.n.

Recent experimental development, especially of mass separation techniques, has initiated experiments which have added considerably to the spectroscopic: knowledge in this region. Table 8 contains a survey of investigations published after the completion of papers P.It-P.VI. The most productive con- tributors have been the LOHFNGRIN, OSTIS and TRISTAN facilities (see sections 3.2.1. and 3.2.2.1. The lanthanide sources were obtained chiefly as decay products from mass separated cesium (and barium) beams although in some Table 8. Survey of recent investigations on the lanthanide isotopes La, Ce and Pr.

IKiCLIDE REFEREKE PRODUCTION ftTAKATIOS «EAS'JPKMEHTS LKVEL SCHEME HALF-LI FE Qg - VALUE

TU.P1 • 0. THia «ork (P.III) :' ,-«.„«.. ,,-coi„r. .!8 levels, SS -„,V = TlMM'o et »1.(177) 11. '1* • 0. if< mi» 32(10 t 100 keV

1 * L« Thi» wrt i 'W.ri '"'!-, i.'\S": 10 levels, H» Y-rays Ski IB levels, 2b -Y-rayu Monnsnd «t al.l>?9) 5100 t 100 heV Stippler et •!.('79) Walters et al.(iflo) levtla, ^r. Y-r&ys btlow C MeV, 1 Mi ctirl «kakis *t «l.n levels, 1^6 -r-rays, T™

Thi, vort (P.Wt 10 levels, 16 Stippler «t KI.CiT9) piass (A) "y-ainK-, lil 10 t 100 keV Pfti ff«r et *1. ( tS?> " !:,A) B.Y-3U1R li I IC ! 100 keV Ennler « «j.iiBD

tll6 U Thx« vork (F.IV) 9 levels, ' SkarnciMrk (93) U levels, ffcnnand «t »1.( 6.? a (l.sp.), 10.0 s Cl.sp.J Kryier «t *1.() 6850 t 100 keV Il.sp.) 6C60 t !?n keV (n.sp.) main levels, I 6.U • O-t s (Lap. J 100 keV (l.sp.) 30 keV (l.sp.)

Thim wort (P.VI) On-line che». (Z) •y-sinp. 13 Y-rays ?-2 * 0-3 s Stippler et »1-C79) On-line masa {Al Ø-ainR.,

En«l*r et •1.C33) 6-sin€. U.I* t 0.5 s Schussler rt »1.f «99! 22 levels, ^7 Y-rays below H^O keV li s STai J « ai. I i9C' 15 levels, ?fl Y-rays belov 050 KeV U.I18 i 0.06 s

On-line them, !~; •j-ain«. SMirne««!-« <9Ji j levels, it Y-riys ? 1 Emtl«r «t «1.1iPJ) A.,_ »i ««. IA) B-sin«. waiters « »l.('3of Y-sinp.., n main levels, I* Bnmirr et »l.i *6S1 6020 + 100 keV Sill et *i.i'<^i ?8 levels, b2 Y-ra; 1.19 * O.05 s Gill tt «1.1 »«) i.Oi 1 0.01 s

C« This wort .' 3.0 • 0.2 min 2530 t 70 keV Taylor et »i.it91} 9 levels, 16 >-rays \78 t 0.08 min t *1. ( 19*" J '0 levtls, ?'? Y-rays ?.O! • O.06 min 260D t 100 KeV l.?0 * 0.21* rain 21*12 i 100 keV

Continued on the following page. Tabl« 8 continued.

1 1 .*> >.,» .->rk f. .' ' 'n.r.r ••••• :^ * ii-..i.- T.-r. "• Y-smc, ^Y-roinc. 7 levels, 'f Y-rays il-.? * 0.1 min 3«mrnr^r« . J • " " " M 1-vels, 19 Y-r&ys

Tiaaeo* • *'. ni. '• *- " . ff-i;ne -f-em.' I" > £.1 -s-.np. , BY ,YY~ "line. fl Uvels, y1 Y-rsys M. V * O.i)min 9',c + ^0 keV

Eton« rt a. . ' i* ' .'fi|,f >"*•• '-n * N<-. reporit i.in B.X-c lirn". - lU * ' min

r *•> This work .-.«"' '"'"•"•h' "* •'•*''"''" ^ Un-lir.e ^he-m. 17'• i|-«sinfi. , Y>-coinc. ' level;:, U Y-rays ^,,.T • .'.i -, Stipplrr *'. »l.i'"1* " On-line mass !A' P-ninfi., BY-"oinc. ~ - Vi'^ » •*[, k,.-,'

Totsuka et al. ,."3i. " .Tf-line cheir-'T.) Y-aifiR., y>-('oinr. '' levels, H Y"roys ^7 ' '> s-

H|*e This work .F.Vi ' ^'J( n, /'"''•'.V*' ' U - CJn-lin*» cheB. I") Y^airiK. , YY-coinr. H IPVCIS, ?U y-rny;, S0.'> f 1.(, s

Stippler rt al.I'791 " Cr-liiif ifiass .(A> 6-sing., BY-ooinc. " " ^Oi0 * 10C k«*v

Kboiv « H.iml "^CflBp.fi'^C^^^U S" St. separation flX-coine - W.o • i.C r ?^t i U)i ]ii-V

9LV Thi« work IP.VI! ?^U(n .f)1 9Ce*' U * On-line chem. i'.'' i-sin^. - 'i.! * ^-b s

SharntaBrfc [91) " " " 8 y-rays "

an-line mnos (A) 6-sinf-.- - "--T * O.S s r -t al.I '9f

1 IP-VII "-V..I uln-th* rrr: «,-.*'^e* U . Or-lin- rhe«. ! = h Y-s:r.f.. - h.fi • D.!, i,

^•. al.:-«. " m-line rrj.SE (A' " 1 Y-ray "i . fi • O.f; s

« B...WI ' '.'(n,,,!" 'U. •< e-sin«., BY-romc. 1080 * if Ifi keV

iHJHUKl ;-rr-lin.. ch»..r.l 6,Y-,.r,C. . 8Y.TY-c=in- JO 1••vels. 8,T -rw,

' Fr This verk iP.II» ''^ju^.f114"-« ^ n-nnp "heir. [" Y-sinr., »-.-oinc. f. ievei^, ?O Y-raya .'.."4 U, 1 »in

Skam-aark I9V " " " IL 1"VP1S, ?!• Y-ray=

Suppler »•• »1.:ITO; 3^L'ln ,f''^^Fr* "'^'<• S" ';II-1H:P ranss '.«« ri-«in,;., 6Y-'-oinr. - - ;>(1'O * iff". KPV

l ?i 1 Ih«-(]a '• \1. li V-nt ,f • '-^Pr*' "^;'? $ ;"fT-linr ohfm..^ P,^-siti?.. BY ,Y>e-<:uinr • ''• Ifv^ls, IC Y-i*ay5, l" ,'..': ' *J.IJ'. ir i., 'l.-j,.) tHf.o • ,'0c keV T levels, 3 Y-rays, i" ;'.T * O.i min 'tt.fp.l

SktncMrk i^>, " " " il levels, V> Y-rnys

1 Pr-iffcr « al.i "9^1 ' \in.fc,ri Frt > " Dn-line mass A) Y-sing. li 1 Y-my3

Pirston vi a.. "*?i " " Y-SIOR., vY-coinc. .'U Levels, VJ Y-r«ys ;'.^'J ' n.^? min

?fei*T#r #t al.' Wt ; * ^"."in ^.f/ "fr* '' V * ^n-lm» RSSS !Al Y-sin«. u Y-rays *i.i ' d.8 i.

II'.M *? «I. ''••C1 '^JMfn,,,^.,.,?.1 ' ' Pr !li sfpnra'iun Y-sinR. .' l»veli;, "1 Y-fiys ', • i i,

K*ys«r «*. ai. ,-0» • *"""!n ,f'r'"i' * Oti-Iinv r.ass A fl-sini?., PY-'-oinc. - - "i'!"'i ' ^ k".-V 63

cases also as primary formed products (LOHENGRIN). Conversion electron and >n~ angular correlation measurements have allowed the determination of spins of a number of excited levels, and 0 - energies have been determined by f3*y- coincidence measurements. Some aspects of the more conclusive part of the pre- sent experimental knowledge are briefly discussed below.

Nuclear structure. So far, the most complete sets of data have been obtained for the even-even isotopes of cerium and . Based on a quasi 206 ) band description the level systematics are derived for the cerium isotopes with N=84 to N = 92 and for the neodymium isotopes with N=86 to N=92, and they are shown in fig. s. 15 and 16, respectively. Dashed lines are drawn to guide the eye and are used to suggest related structures in successive isotopes. For cerium a smooth trend towards lower energies for each neutronpair added is apparent for all the lowest-lying members of the ground state band (0 ,2 ,4 ), octupole band ( 1" , 3" ) , P band (0+ ,2* ) , and -y-band (2*,3*). The first 2*-state (2 1 drops by a factor of -0.63 with the addition of each successive pair of neutrons while the first 4 -state (4,) drops more slowly by a factor of »0.70. Thus, the ratio E */E -i, plotted in fig. 17, appears to shift from near the

2.2 2.0 \ 4* (4+) 1.8 1.6 i

0.6 0.4 O.2 0 QL-

Fig. 15. SyttaiMtics of a selection of low-lying l*v«lt in th« «van-avtn cerium isotopes with 8* < N < 92. Thu data hav« bttn taken from raf.c 101, 106, 19? and 203. f. 4

1.8

1.6

1.4

pl.2

s:

>- 1.0 • \j UJ m 0.8 _j

LkJ ;> UJ -• 0.6 -

0.4 "

0.2 - ?

0 . Si- — 146., 148 150, 152,, "rifMd 88

Fig. 16. Systematics of a selection of low-lying levels in the even-even neodymium isotopes with 86 _i N ^ 92. The data have been taken from ref.s 198. 20.1, 204 and paper P.I!

i

Ce © Nd . ».- -•-• Sm 3.0 -«- Gd F : J . LU m V. 7!' 2.5 Oi LU il /f I

• •

2.0 - -

1 1 8-1 86 88 90 92 NEUTRON NUMBER N

fig. 17. Systematic tromis of \.Ui< \ u^l\ ;^ ia(ici<. for even even nuclides with 58 < ?, t 6* and il : N ; 97. The data for Sm and Gd wsrc takt-n (rom rcf. ?fl 1. 65 vibrational limit of 2 to near the rotational limit of 3.33, suggesting a smooth transition towards deformation. The same trend is apparent for the neodymium isotopes although the change is not characterized by the same degree of smoothness. The sharp increase between N=88 and N=90 is similar to the trend found for the and isotopes where the onset of deformation is known to occur between N=88 and N=90.

The ratio E +/E + is considered to give a good indication of the onset of deformalion . This ratio is plotted for cerium and neodymium nuclides in fig.18 and compared to data for neighbouring barium, samarium and gadolinium isotopes. For neodymium, samarium and gadolinium isotopes the ratio gradually increases with the neutron number to a value of about 3 in the

1 1 1 1 i i i

16 -o-- --o Ba

14 —o Nd —• Sm Q —o Gd 12 /•' '•''

+cTl0 u

«*8 u

6 jf

4 r

2 -

0 i i i • i i 812 84 86 88 90 92 94 NEUTRON NUMBER N Fig. 18. Systematic trends of th« E02/E2T ratios for •ven-tvan nudidts with 56 < Z < 6* and 82 < N < gt. Th« data for Sm, Gd and 142-Nd war« taken from r«f. 203 and th« data for 6a from r«f.s 203 and 205.

A level at 723 k«V in Nd has been interpreted by tkeda et al. ' as beeing the band head of the quasi 0-vibrational band. This would lower the E *lt * ratio from 3.04 (from the 916 keV 0*-level presently °2 Z1 used) to 2.40 and represent a slight deviation from the systematic pattern constituted by the neighbouring elements. The general conclusion about the shape transition would, however, remain unchanged. 66

vibrational area, drops slightly, and then rises sharply after N = 90 to values above 10. This may indicate a rather sudden onset of deformation at N=90. For barium isotopes the situation is different. Here the ratio rises continuously to N=90, which could indicate a more gradual change in nuclear shape. The existing data for the cerium isotopes suggest an intermediate situation.

To sum up, systematic evidence suggests that the onset of deformation in the cerium isotopes should not be considered to have occurred prior to N=90 1 U B ( Ce). The transition to the well-deformed region follows a rather smooth pattern in contrast to what is found for the neodymium isotopes which show a more sudden onset of deformation at N=90( Nd).

154 204> Since the E4*/E • -ratio for Nd is 3.23 , the value of 3.26 for the Nd indicates that deformation reaches a maximum for even-even neodymium isotopes at N = 92. A similar effect is observed for samarium isotopes where maximum deformation is found at N=94. No such effect has, so far, been observed for cerium isotopes.

The proposed level structure of the cerium isotopes has been compared with 192) calculations using the Interacting Boson Approximation with neutrons and protons treated separately IIBA-2) . By applying a set of parameters which reflected the subshell closure at Z=64, a reasonabely good agreement was obtained for the level energies of the even-even isotopes with N=86,B8,90,92. However, the calculations failed to reproduce the experimental transition 1 4 R probabilities for Ce. A more rotational structure would be necessary in order to reproduce the observed transition probabilities.

On the basis of the calculations performed and published so far, the general impression is, however, that I6A-2 seems to work well in this crucial transitional area of low-Z rare earths, and shows promises as a predictive tool in this region.

For the even-odd, odd-even and odd-odd isotopes of low-Z rare earths, general conclusions are difficult to draw on the basis of the existing data. In spit« of the large amount of •xperimental information which has appeared during the last few years, knowledge about the level systematics is still inadequate, tn order to begin to understand these nuclides, there is an obviour. need for still more refined measurements, such as angular correlations. 67

Masses. There are at present several reasons for determining accurate atomic masses (see for instance the summary article by 3. C. Hardy, ref. 209), one being the study of systematic trends, i.e. for further refinement of mass formulas which are used to predict properties of unknown nuclides far from stability. Such refinements have fundamental significance (i.e. in astrophysics ) as well as practical applications (for instance in the calculation of decay heat and the estimation of delayed neutron probabilities in thermal reactors).

For the mapping of the mass surface of short-lived neutron-rich nuclides, two methods are mainly applied. One is to determine mass differences between neighbouring nuclides by direct mass spectroscopy. The other is to determine the mass difference between two adjacent isobaric nuclides by measuring the total p-decay energy in p-singles or Pf-coincidence experiments. The first method gives very precise values with error bars of a few tens of keV, but has until now mainly been applied for the alkali elements . Future experimental improvements will certainly extend the applicability of this method. In the second method plastic scintillators have traditionally been used for detection of the p-particles with precision of typically a few hundred keV. Recently, however, hyperpure detectors, Ge(HpGe), have been introduced, offering the potential for determining p-particle end-point energies with higher precision than is possible with scintillation detectors, due to superior energy resolution, ease and accuracy of calibration, linearity of response, and stability.

The Q„-values given in table 8 have all been obtained by measuring the endpoint of P-spectra using either scintillation or Ge(HpGe) detectors. The accuracy of such measurements depends upon a good knowledge of the decay characteristics, and necessary data have been taken partly from the present work.

The weighted means of the measured Q.-values for the lanthanum, cerium and praseodymium isotopes are compared to predictions from a selection of mass formulas and shown in figures 19, 20 and 21, respectively. The chosen formulas are those of Myers and Takahashi et al. based on the droplet 68 model, the liquid drop model formula with shell corrections by Seeger and Howard"217' ) ', th. . e semi-empirica. . . l, shellmode.__.,, ,_l, formul* ,a_ of„ iLiraj n and. -,_,_•_Zelde_s 2 1 8 ) and 219 ) the empirical mass relation by Comay and Kelson

For the praseodymium isotopes the Liran-Zeldes and Comay-Kelson formulas fit the data resonably well while the others seem to underestimate the Q -values p somewhat. For the lanthanum and cerium isotopes the agreement between prediction and measurement is good, and the deviations largely stay within *_ 1 te. 500 keV except for La. For this nuclide all the considered formulas predict notably higher Q -values than those measured. The existing value was obtained p from the p-singles spectra because the collected activity proved to be too 1 88 sparce to permit reliable p^-coincidence measurements ). Accordingly. the error bars may be considerably underestimated because weak contribution from high-energy transitions may have escaped detection. The measurements should be repeated at a production facility which is capable of producing higher source strength, e.g. at ISOLDE, where a beam of Cs with intensity of 10' -10 atoms/s is routinely obtainable from a normal UC -graphite cloth target (P.IX, fig. 5).

1 • i i .""LZ .' GHT La- sotopes ; y; 1000 A 143 32802 100 keV 144 5545 2 150 •• / 145 4110 2 100 •• / ' 1 146 6400 2 100 •• > • : /M 147 4750 2 120 ••

148 60202 100 •• / . ••/ soo •' • ••'/

/ ;

0 .,f] i

%\ / 500

i i i U3 R5 K6 146 MASS NUMBER A

Fig. 19. Qg-valu«» for tht ntutron-rich lanthanum i»o- topas: Comparison bttwatn maaturamtntt lztro-lin«t and prediction* from the nats formulas by Mytr» (HI, Taka- haihi-v.Groot«-Hilf (TGH), S»»g«r-Howard (SHI, Liran- Z*ldas ILZI and Comay-Kalson (CK). r* < -n r* < TI » •> H O S> H- ••. c • (Q Ul ro O CALCULATED 0B - MEASURED Of) (keV) -«> • Ul Ul -ta O — O ••• •! O ,-. O «3 CALCULATED Qfl - MEASURED 0,, (keV) o w o o o S SI -i o

3 o o o o I "1 (S »-•

The precision of the various mass predictions may conveniently be compared using the root mean-square deviation. These numbers are displayed in table 9 1 t ft for all the nuclides in question except La (due to the above arguments).

Table 9. Root-mean-square deviation (in keV from Qg(calc.J-QgCmeas.)) of some mass pre- dictions for the elements La, Ce and Pr in the mass region investigated here.

Nuclides Myers Takahashi Seeger Li ran Camay Janecke Janecke v.Groote Howard Zeldes Kelson (Garvey Eynon Hilf Kelson)

143-147, La 155 110 298 289 124 144 176

145-150Ce 253 202 177 227 404 458 475 147-15Opr 441 437 745 166 109 170 234

Total 292 267 436 236 271 314 339

An additional two empirical mass relations are included, i.e. those "arrived 220) 221) at" by Janecke and by Janecke and Eynon . The best overa.1.1 fit is obtained with the Liran-Zeldes mass formula (containing 178 free parameters). 222 ) This agrees with observations by E. Lund et al. for <,2 measured values 211) with Z<50, by G. Audi et al. for elements in the francium region, and by M. Epherre et al. for cesium isotopes. It should, however, be noted that the 16-parameter droplet model formula by Myers also shows a good fit, in contrast to the results by E. Lund et al." who found that the Myers' formula gave the largest discrepancy. In the present c«.impari;«n the Jargest discrepancy is observed for the mass relation of Seeger and Howard (9 free 202 I parameters), in agreement with observations by Keyser et al. for a few measured values with 50 < Z < 57. 71

5.2. Spectroscopic measurements on the doubly magic nucleus Sn . 132 The nucleus Sn with 50 protons and 82 neutrons is of special interest since it is the only medium-weight nuclide with closed major shells both of protons and neutrons. Hence one would expect a high stiffness against deformation.

This nuclide has been studied at ISOLDE, and the experiments and results are reviewed here in somewhat more detail than the previous discussion for the lanthanides since, in this case, the interest is concentrated on the single nuclide rather than on systematics of isotope and isotone chains.

132 At present, Sn is only accessible through fission. Previous investigations 223) - 132 at the on-line separators OSIRIS (|3 -decay of separated In) and JOSEF (direct separation of the 1.7 (is isomer) led to the construction of a level scheme including a few of the main transitions (fig.22).

Fig. 22. Level scheme of 132-Sn deduced from experiments at the OSIRIS (ref. 223) and JOSEF (ref. 224) isotope separators.

These initial experiments left a number of important questions unanswered, as for example: Was the assignment of the 4 0^1 keV level as the first excited state unambiguous? [f so, was the spin and parity 2* or 3" ? Both alternatives 223) 9?i ) are suggested by systematics , by the deexcitation pattern and by 225) random-phase-approximation (RPA) calculations In the heavier double-magic 208 nuclide Pb the spin of the first excited state (26K keV) is 3"

132, At tSOLOE '" In may be produced at an intensity of 10*-105 atoms/s by 600 MeV proton-induced fission in a "normal" UC -graphite cloth target. Two experi- ments were performed. The first was designed to determine the multipolarity of 72

the 299 keV and 374 keV transitions and the second to determine ^-y-coincidence relationships and half-lives of the individual -y ray peaks (fig.23 a and b). The main results are presented in papers P.XIII and p.XIV.

Ion beam

Ion beam Collector tope

Ge(Li)-o>1eclor

fig. 23. ». Schematic diagram of the experimental set-up in measure- ment of conversion coefficients for multipolarity determinations, b. Detection set-up for beta-gated gamma singles and gamma-gamma coincidence measurements. In both cases the sources were collected in the measuring position on a tape which was stepped regularly in order to reduce disturbances from grown-in products and 132-Cs which was present in the primary beam.

The measured conversion coefficients of ot„(299keV)=0.028+0.005 and a (374keV)=0.020^0.007 are only consistent with an C2 or M1 transition (P.XIII, fig.2). Simultaneously performed delayed tt-coincidence experiments at the JOSEF isotope separator gave a half-life of the 4 415 keV transition of 2.1±0.3 ns (ref.226 and P.XIII), later remeasured to 4. 0+.0.3 ris (P.XIV), which 225 1 agree with an E4 assignment. Together with RPA-calculations , these data strongly suggest a spin sequence 8+ (4647 keV) -* 6*(4714 keV) •• 4*(4415 keV) •» 2* (4041 keV) -» 0* (g.s. ) .

The ^-measurements showed a number of new transitions (fig.24) with the same half-life (fig.25) as the previously known transitions. From the coincidence relations the decay scheme given in fig.26 was established. The scheme is based on 215000 "good" coincidences only, due to increasing background disturbance fram the Cs activity which made a continued data-taking non- valuable. Therefore, only transitions with I > 10/! showed clear coinci- dences. A few weaker transitions are placed in the scheme on th« basis of half-life and level energy differences. COUNTS. RELATIVE INTENSITY CQUNTS. RELATIVE INTENSITY

f- MM«

Ul m t-- 1— TI (cDr 1£— 3T 3 0 (S H rt- ri- Q 3 v> •1 rf rt- O 3T « O •t H- H- H1 C a> a (O o —1 rfr 3 •i n> 3" £ 3- 3" rf 3 (S It 3 -) n •o 3 c H- 3- (D ^- £ T3 rf O •1 W IC w n- Ul 1 I a IN3 74

Fig. 26. Decay scheme for 132-In, and levels in 132-Sn.

From shell model considerations one would expect the J=7 ground state of 132 - 1 V In to decay mainly by an allowed Gamow-Teller transition (vf ' 9q ,?) 7 * lvfnT™9nn'i:' • The new -T-evel at 7210 keV is most likely the expected 6 - lie y / c b state, supported by the low value of log ft = 4.5. The coincident cascade depopulating this level suggests for the lower levels a spin sequence 5' (4942 keV), 4" (4830 keV) and 3"(4351 keV). tt should, however, be pointed out that a direct experimental proof of the 3 -assignment to the 4J51 keV level is still lacking. All the same, the present experiments have provided further evidence for the assignment of the 4041 keV level as being the first excited state in 132 Sn.

The very large gap between the ground state and the first excited state is in fact the largest known for any nuclide above 0. Even allowing for an A scaling reflecting the variation with the size of the nucleus, the gap is ? 0 fl considerably higher than in Pb. Thus, the emerging picture is that of a nucleus with unusually strong shell closure, the strongest of any nuclide known. 75

OUTLOOK

The number of printed contributions to the four conferences arranged so far on "Nuclei Far From Stability" ' is shown in fig.27. The graph may illustrate

Fig. 27. Number of printed contri- butions to the four international conferences on "Nuclei Far From Stability".

the trend of activity in this field of research, though not necessarily in a true quantitative measure. The accelerating growth is partly due to the continuous improvement of traditional methods and occationally the invention of new experimental techniques at already well established facilities, and partly to new facilities joining in. Examples of such facilities which have just star- ted taking data are the Oaresbury Isotope Separator and the recoil separator on-line to a 30 MV Tandem Van de Graff accelerator (England)226), the He-jet ISOL at the Kyoto University Reactor (Japan) ), the EMIS facility at the 2 2 B Tohoku University Cyclotron (40 MeV p and 50 MeV a. Japan) ), the Chalk River Isotope Separator on-line to a Tandem Van de Graff accelarator with prospects to reinstal the facility at the new Chalk River Superconducting Cyclotron O10 7 0 Q ) MeV/amu for all elements up to uranium. Canada) the recoil separator CARP at the Osaka University AVF Cyclotron (-10 MeV/amu for elements up to A = 22 i ,230) Japan) . and the isotope separator TRIS on-line to a 1 GeV proton 231 ) synchrocyclotron at Gatchina (USSR) . A few more are in the planning staqe.

At the ISOLDE facility the experimental program is expected to continue well into the nineties. From 19B« the CERN SC operates exslusively for the rSOlDF. experiments, and »-he available number of shifts (1 shift = 6 hours) will increase from about ?S0 to possibly 500 per year. Operation at this scale is, however, not technically possible with the present installation alone. A new mass separator is planned, ISOt.OF III. Tt has been designed to handle bpam currents in the mA-reqion. A maximum mass resolution of -100O0 may be obtained 76

under ideal running conditions, thus making possible a reasonable separation

of those isobars which have a difference .in Q -values of approximately 3 MeV

and more. Ihis high resolution will also allnw discrimination of contaminating

molecular ions, thus producing verv pure ion beams.

the prospective upgrading of the ISOLDE facility necessitates. and makes

possible. increased allocation of time for on-line target tests. An effective

target/inn source rteve1opment, repeatedly termed the "bottleneck" at on-line

isotope separators, is imperative for an efficient operation of a multipurpose

facility like ISOinF. Of special interest are the possibilities inherent in

the chemical evaporation technique. Several new possible target systems

are awaitinq on-line tests (see for instance ref.146).

The combined G1RT-SISAK technique has been useful as a first advanced experimental tool in regions which were not attainable with mass separation techniques. The production efficiencies of the various elements are not, as in most traditional mass separation technique-, dependent upon vapori7ation of

the nuclides to be studied. Thus, the SISAK technique is suitable for studies of refractory elements like Nb, Mo. Ru, Rh, Hf, Ta, W and Re.

So far, the technique has mainly been applied to separation of fission products, but it is probably readily adaptable to separation of other reaction products resulting from medium to high energy charged particle reactions.

Thus, neutron-deficient spallation products and heavy ion fusion products may be studied as well. At present, the SISAK system is about to be tested at the heavy ion arcelaratnr UNII.AC in Darmstadt where separation of actinide elements from heavy ion fusion will be attempted

Application of short-1 i vod radionuclides to other (not basic nuclear) fields of study will probably continue the strong growth we have experienced during the last 15 years.

In nuclear medicine special interest is devoted to the positron emitters C

13 1S (t 12 18 (t1/?=?0.3 min.), N (t1/? = 9.97 min.). 0 1/?" ? *' and F (*• ., = 109.R min.) These nuclides can be incorporated into biologically active molecules and administered to a biological system. By mean*; of a positron emission tomograph which detects the two coincident 511 kf>V photons from fl*- annihilation. Irhree-dimensLona 1. picture? of thp activity distribution may be recorded. Thus, physiological processen may to»* followed as u function of time or external stimulation Extensive research will nintinue ;n order to improve product inn methods and '.peed no labelling pror cdur »•<, making t tiftn '.«mi 77

continuous or fully continuous and remotely controlled.

Methods for the production of other radionuclides with documented or potential medical application are also subjects to continuous improvements. As an 1 23 example, isotopically pure and sterile I may be produced rapidly and in large quantities by applying the reaction Te(p,2n) [ using a target 1 ?4 material moderately enriched (»90'/) in Te, a subsequent non-destructive separation of iodine from the target by dry-distillation and a final mass separation of the collected iodine by means of a negative surface ionization source. Experiments are at present in progress at the ISOLDE laboratory to determine the optimum conditions for each step in the process.

In elemental analysis the use of nuclear techniques is in several cases the best (or only) realistic alternative for solving uniqup analytical problems. Irradiation with thermal or fast neutrons for element, bulk determination or irradiation with charged particles (E « 50 MeV) also for depth and lateral concentration profiling. are powerful analytical tools when combined with suitable detection techniques. The fast (14 MeV) neutrons are especially useful for rapid and reliable determination of light elements from nitrogen to phosphorus. Charged particles at present constitute the most sensitive probe for a number of light elements from to fluorine.

While thermal (reactor) neutron activation analysis has achieved a general application, the routine use of K MeV neutron activation analysis is limited, and charged particle activaton analysis is still rather exclusive. However, with the increasing availability of small particle accelerators, these methods will in the future be more commonly applied as standardization procedures and even be used for a limited number of routine analyses. Opvelopment of improved and new techniques for this area of application is an important and interesting task.

In conclusion, with the constant experimental reinforcement indicated, the increase in scientific output of both pure and applied character in this field of research is likely to continue for quite some time. Quantity is not necessarily quality, however, and critical voices are warning against signs of superficial data collection. My belief is. none the less, that the maturity of the numerous experiments already reported, and of those which will be carried out in the future, will provide data useful for applied purposes and/or for the process towards a more complete understanding of the various aspects of nuclear behaviour. I do not believe that the only valid reason for experimental attack is a firm theoretical prediction of new phenomena. T8

Sometimes experiments come first. The past has repeatedly shown that unexpected discoveries may result even from a systematic study of some nuclear parameter. Detailed establishment of reliable systematics is, in addition, crucial for model testing. This declaration does not imply an a priori approval of all possible experiments. It will remain an important challenge to the scientist always to search for that particular action which is likely to provide most valuable information.

The present author looks forward to a continued engagement in this field of research, and hopes to be able to contribute to its development for many interesting and rewarding years to come. 79

REFERENCES

1. W. Forslmq, C.3. Herrlander and H. Ryde (eds.). Pror. Tnternat. Symp. on "Why and How should We Investigate Nuclldes far off the Stability Line", Lysekil. Sweden, Aug. 21-27, 1966, Arkiv for Fysik. 36., (19671.

2. I. Bergstrbm, in. ref. 1, p. 3.

3. Proc. Internat. Conf. on the "Properties of Nuclei far from the Region of Beta-Stability", teysin. Switzerland, Auq. 31-Sept. 4, 1970. CERN 70-30, Vol. 1 and 2 . ( 1970) .

4. Proc. of the "3rd International Conference on Nuclei far from Stability", Cargese. Corsica (France), May 19-26, 1976, CERN 76-13, (1976).

5. Proc. "4th International Conference on Nuclei far from Stability". L.O. Skolen, Helsingor (Oenmarkl, June 7-13, 1981, CFRN R1-09, (1981).

6. F. [achello, in ref. 11, p. 281.

7. J.H. Hamilton, Nukleonica, 2JL. 561(1979).

8. R.E. Azuma, G.I. Borchert. L.C. Carraz, P.G. Hansen, B. Jonson, S. Mattsson, 0.8. Nielsen, G. Nyman, [. Ragnarsson and H.L. Ravn, Phys. Lett. , 8_6B, 5( 1979) . 9. R.F. Azuma, L.C. Carraz. P.G. Hansen, B. .lonson, K.-l. Kratz, S. Mattsson, G. Nyman, H. Ohm, H.L. Ravn, A. Schroder and W. Ziegert, Phys. Rev. Lett. , 4_3 , 1 652 ( 1979) .

10. R.E. A?uma, T. Bjørnstad, H.Å. Gustafsson, P.G. Hansen. B. Jonson, S. Mattsson, G. Nyman, A.M. Poskanzer and H.L. Ravn, Phys. Lett., 96B. 31 ( 1980) .

11. J.H. Hamilton, E.H. Spejewski, C.R. Binqham and F.F. 7ganjar (eds.), Pvoc. Internat. Symp. on "Future Directions in Studies of Nuclei far from Stability", Nashville. Tennessee, Sept. 10-13. 1979.

12. T. Von Egidy (ed.l, Proc. Conf. on "Speetrnscopy of Fission Products". ILL. Grenoble, France, Hay 1979, published by the Institute of Physics. Bristol, England.

13. P.G. Hansen, Ann. Rev. Nucl. Part. Sci., 23.. 69(1979).

H. 5.A. Nolen, Jr. and W. Benenson (eds.), Proc. "Sixth International Cnn- fprence on Atomic Masses", Fast Lansing, Mirhiqan, Sept. 1H-?1, 1979. (Plenum Press, New York, 1980).

15. R.M. Diamond and J.O. Rasmusspn (pris.l, "Proceedings nf the International Conference on Nuclear Physics", Berkeley. California. Aij'i. ?4~1fl, 1980. Muc-J . Phys., A3J.4 , (1981).

16. B. Jonson. in ruf. 1"), p. 77c.

17. O.H. Wilkinson, in ref. I, p. 71. 18 K.H. Burkard, W. Dumanski, R. Kirchner. 0. Klepper <»iid f. Ro«?i:kl.. Nucl. Inqtr. Meth.. 139.. 275(19761 80

19. S. Hofman, W. Faust, G. Munzenberg, W. Reisdorf, P. Armbruster, K. Giittner and H. Ewald, Z. Phys., A291 . 53(1979).

20. G. Munzenberg, S. Hofman, F.P. Hessberger, W. Reisdorf, K.H. Schmidt, W. Faust and P. Armbruster, in ref. 22, p. 755.

21. H. Gåggeler, M. Schådel, W. Briichle, J.V. Kratz, K. Siimmerer, G. Wirth, N. Trautmann, P. Peuser, G. Tittel, R. Stakemann, G. Herrmann, E.K. Huiet. R.W. Lougheed. J.M. Nitschke, R.L. Hahn and R.L. Ferguson, in ref. 22, p. 763.

22. H.L. Ravn, L.C. Carraz, J. Denimal, E. Kugler, M. Skarestad, S. Sundell and L. Hestgaard, Nucl. Tnstr. Meth., 139, 267(1976).

23. L.C. Carraz, S. Sundell, H.L. Ravn, M. Skarestad and L. Westgaard, Nucl. Instr. Meth., 1M, 69(1979).

24. C. Détraz, D. Guillemaud, G. Huber, R. Klapisch, M. l.angevin, F. Naulin, C. Thibault, L.C. Carraz and F. Touchard, Phys. Rev., £19, 164(1979).

25. C. Thibault, M. Epherre, G. Audi, R. Klapisch, G. Huber, T. Touchard. D. Guillemaud and F. Naulin, in ref. 20, p. 291.

26. G. Rudstam. Nucl. Instr. Met., J39, 239(1976).

27. G.D. Westfall, T.J.M. Symons, O.E. Greiner, H.H. Heckman, P.J. Lindstrom. J. Mahony, A.C. Shotter, D.K. Scott, H.J. Crawford, C. McParland, T.C. Awes, C.K. Gelbke and J.M. Kidd. Phys. Rev. Lett., 4j, 1859(1979).

28. A.G. Artukh, V.V. Avdeihcikov, J. Ero, G.F. Gridnev, V.L. Nikheev, V.V. Volkov, and J. Wilczynski, in ref. 8, p. 47.

29. 0. Guerrau, J. Galin, B. Gathy, X. Tarrago, J. Girard, R. Lukas and C. Ngo, Z. Phys., A295, 105(1980).

30. H. Breuer, K.L. Wolf, B.G. Glagola, K.K. Kwiatkowski, A.C. Mignerey, V.E. Viola, W.W. Wilecke, W.U. Schroder, J.R. Huizenga, 0. Hi]scher and J. Birkelund, 0R0-5172-0017 , (1980).

31. K.K. Kwiatkowski, H. Breuer, A.C. Mignerey, V.E. Viola, L. Wolf, B.G. Glagola, J.R. Birkelund. 0. Hilscher, J.R. Huizenga, W.U. Schroder and W.W. Wilecke, abstract 23.90*W for the conference mentioned in ref. 15.

32. B. Forkman and G.G. Jonsson, in ref. 1, p. 57.

33. O.W. Dorn, in ref. 1, p. 41.

34. C. Brendel, F. Beck, J. Carter, G. Delic, M. Marinescu, A. Richter and G. Schrieder, in ref. 5, p. 664.

35. K.K. Seth, in ref. 5, p. 655.

36. G. Herrmann and H.O. Denschlag, Ann. Rev. Nucl. Sei., 19, 1(1969).

37. N. Trautmann and G. Herrmann, J. Radional. Chem., 32, 533(1976).

38. N. Trautmann, in ref. 4, p. 30.

39. H.V. Weiss and W.L. Reicherl, Analyt. Chim. Acta, 14. 119M9fifi). 81

40. E. Ohyoshi, A. Ohyoshi, T. Takemi and M. Shinagawa, Nurl. Sei. Techn., iO , 101 ( 1973 ) .

41. H. Ahrens, N. Kaffrell, N. Trautmann and G. Herrmann. Phys. Rev.. f_1_3 , {19761 .

42. K. Bachmann, V. Matschoss and S. Tsalas, Proc. "4+h International Trans- Element Symposium", Baden-Baden, Sept. "S. p. 37(1976).

43. K. Bachmann, V. Matschoss, .1. Rudolph, A. Steffen and S. Tsalas, in ref. •5 8 . p. 343.

44. N. Trautmann, N. Greulich, II. Hickmann. N. Kaffrell. E. Stender. M. Zendel, H. Gågqeler, K. Prodén, G. Skarnemark and D. Eriksen, in ref. 5, p. 727.

45. R.J. Silva. N. Trautmann, M. Zendel, P.F. Dittner. F. Stender and H. Ahrens, Nucl. Instr. Meth., J.4.7 , 371(1977).

46. Jahresbericht 1977, Institut fur Kernchemie der Universjtåt Main7. ( 1978) .

47. Jahresbericht 1978, Institut fur Kernchemie der Univers]tat Mainz. ( 1979) .

48. Jahresbericht 1979, Institut flir Kernchemie der llnivers.it at Mainz. I 1980).

49. 8. Grapengiesser and G. Rudstam, Radiochimica Acta. 2J1. 85(1973).

50. B. Grapengiesser, Research Report LF-59, Studsvik, (1974).

51. G. Rudstam, P. Aagaard, P. Hoff, B. Johansson and H.U. 7wicky, in ref. 59, p. 365.

52. P. Hornsho.i. B. Jonson, H.L. Ravn and L. Westgaard, in ref. 57, p. 379.

53. G. 1. Beyer, A.F. Novgorodov, A. Kovalev. F. Pra?ak. V.A. Khalkin and Yu.V. Yushkevich, in ref. 59, p. 401.

54. T. Bjørnstad, Thesis, University of Oslo. 1973.

55. P.O. Aronsson, B.E. Johansson, J. Rydberg, G. Skarnemark, J. Alstad, 8. Bergersen, E. Kvåle and M. Skarestad, J. innrg. nucl. Chetn. , 3j>. 2397(19741.

5fi. H. Wagner and W. Walcher (eds.t, Proc, Int. Conf. on "F.lectromagnetir Isotope Separators and the Techniques of their Applications". Marburg, Sept. 7-10, 1970, Research Report BMBW-FB K70 ?8. (1970).

57. C. Andersson and G, Holmen (eds.) Pioc. 8th Int. Conf. on "Low-F.ner

58. S. Amiel and G. Fngler (eds.), Proc. IXth Int. Conf. on "Fleetrrimag- netii: Isotcpe Separators and Related Ion Accelerators" , Kirvat Anavim, Israel, May 10 13. 1976, Nucl. Instr. Meth., 12? . (1976). 59. H.L. Ravn, E. Kugler anet S. Sundeli (eds.), Proc. ; 1th Int. Conf. on "Electromaqnetic Isotope Separators and Techniques Related to their Application-;", Zinal, Switzerland. Sept. 1-6, 1980. Nud. Instr. Meth., 18!, ( 19B1 ) .

60. S. Borg. I. Bergstrom, G.B. Holm. B. Rydberg, I .-F. De 6eer, G. Rudstam, B. Grapengiesser, E. Lund and L. Westgaard, Nuci.. Instr. Meth., 9J , 1(1971).

61. S. Amiel, Arkiv fur Fysik, 3£, 9(19fi7).

62. W.L. Talbert, Jr. and [>. Thomas, Nucl. Instr. Meth., 3,8. 306(1965).

63. W.L. Talbert, Jr. and 3.R. McConnell, Arkiv fur Fysik, 16. 99(19671.

64. O.S. Brenner, R.E. Chrien, R. Gill, M. Stelts, J.C. Hill and F.K. Wohn, in ref. 11, p. 389.

65. J. Miinzel, H. Wollnik, B. Pfeiffer and G. Jung, in ref. 59, p. 343.

66. A. Kjelberg and G. Rudstam (eds.), CERN 70-3, (19701.

67. R. Foucher, P. Paris and J.L. Sarrouy, in ref. 57, p. 341.

68. P. Paris. V. 8erg, A. Caruette, J. Obert, .1. C. Putaux and J.L. Sarrouy, in ref. 58, p. 251.

69. P. Paris, C.F. Liang. D. I. ecouturier, M. Arianer, J. Obert, A. Caruette, A. Ferro, J. Fournet, J.C. Putaux and J.L. Sarrouy, in ref. 59, p. 91.

70. V.A. Karnaukov, D.D. Bogdanov, A.V. Demyanov, G.I. Koval and I.A. Petrov, Nucl. Instr. Meth., 1_20, 69(1974).

71. E.H. Spejewski et al., in ref. 57, p. 318.

72. E.H. Speiewski, in ref. 59, p. 71.

73. C. Bruske. K.H. Burkard, W. Hiiller, R. Kirchner. 0. Klepper and E. Roeckl, in ref. 59, p. 61.

74. G. Kromer, T. Lund, A.K. Mazumdar, E. Stender, N. Trautmann, H. Wagner and W. Walcher, in ref. 46, p. 24.

75. M. Brugger, T. Lund, A.K. Mazumdar, E. Stender, N. Trautmann. H. Wagner and W. Walcher, in ref. 47, p. 15.

76. M. Brugger, A.K. Mazumdar. N. Trautmann, H. Wagner and W. Walcher, in ref. 48, p. 9.

77. A.K. Mazumdar, H. Wagner, W. Walcher, M. Brugger and N. Trautmann. in ref. 59, p. 131 .

78. D.M. Moltz, R.A. Gough, M.S. Zisman, D.J. Viejra, H.C. Evans and ). Cerny, Nucl Instr. Meth., ±12, 507(1980).

79. n.M. MoUz, J Aysto, M.D. Cable, R.F. Parry, l.M. Wouters and J. ferny, in r

8P. .1. Arje, J. Åysto, 3. Honkanen, K. Valli and A. Hautojårvi, in ref. 59, p. U9 .

81. J. Årje and K. Valli, Nucl . Iristr. Meth., 171. 533(1981).

82. E. Moll, H. Schrader, G. Siegert, M. Asghar, J.P. Bocquet. G. Bailleul, J.P. Gautheron, J. Greif, G.I. Crawford, C. Chauvin. H. Ewald, H. Wollnik, P. Armbruster, G. Fiebig, H. Lawin and K. Sistemich, Nucl. Instr. Meth., 121, 615(1975).

83. P. Armbruster, M. Asghar, .1. P . Bocquet, R. Decker, H. Ewald, J. Greif. E. Moll, B. Pfeiffer. H. Schrader, F. Schussler, G. Siegert and H. Wollnik, in ref. 58, p. Z13.

84 . H. Lawin, 3. Eirlens. .1. W. Borgs, R. Fabbri, .1. W. Griiter, G. Joswig, T.A. Kahn, W.D. Lauppe, G. Sadler. H.A. Selic, M. Shaanan, K. Sistemich and P. Armbruster, Nucl. Instr. Meth., 13 7. 103(1976).

85. H. Lawin. J.W. Borqs. R. Fabbri, J.W. Griiter, T.A. Kahn, W.D. Lauppe, G. Sadler, H.A. Selic. M. Shaanan, K. Sistemich and P. Armbruster, in ref. 58, p. 227.

86. H. Ewald, K. Guttner. G. Miinzenberg, P. Armbruster, W. Faust, S. Hofmarin, K.H. Schmidt, W. Schneider and K. Valli, in ref. 58, p. 223.

87. P. Armbruster, in ref. 4, p. 3.

88. G. Miinzenberg, W. Faust, F.P. Hessberger, S. Hofmarin, W. Reisdorf, K.H. Schmidt, W. Schneider, H.J. Schott, P. Armbruster, K. Guttner, B. Thuma, H. Ewald and D. Vermeulen, in ref. 59, p. 423.

89. G.W. Butler, in ref. 4, p. 15.

90. J. Rydberg, H. Reinhardt and J.O. Liljenzin, Ion Exchange and Solv. Extr. , 3 . 1 1 1 ( 1973) .

91. H. Reinhardt and J. Rydberg, Acta Chem. Scand., 21. 2773(1969).

92. P.O. Aronsson, Thesis, Chalmers University of Technology, Gothenburg, 1974 .

93. G. Skarnemark, Thesis, Chalmers University of Technology, Gothennurq. 1977 .

94. J. Rydberg, H. Persson, P.O. Aronsson. A. Selme and G. Skarnemark. Hydrometallurgy, 5, 273(1980).

95. R.D. Macfarlane and W.C. Me.Harris, in "Nuclear Spettroscopy and Reactions A", by ). Cerny (ed.), Academic Press, N.Y. 1974, p. 243.

96. H. Wollnik, in ref. 58, p. 311.

9?. W.J. Wiesehahn. G. Bisrhoff and ).M. O'Auria, Nucl. Instr. Met., XZi . 2 21(19 7 5).

9fl. W.J. Wiespfiahn, G. Rtsrhoff and 3.M. n'Kuria. Nut 1 . Initr. Met., 1?9. 187(1975). 84

99. H. Wollnik, H.G. Wilhelm, G. Robig and H. Jungclas, Nucl. Instr. Meth. ±H, 539( 1975) .

100. H. Wollnik. J. Postenddrfer, G. Robig and H.G. Wilhelm, Nucl. Instr. Meth. , lii, 247 ( 1977) .

101. D.E. Alburger and T.G. Robinson, Nucl. Instr. Meth., 1£4, 507(1979).

102. A. Ghiorso, J.M. Nitschke, J.R. Alonso, C.T. Alonso, M. Nurmia, G.T. Seaborg, E.K. Huiet and R.W. Lougheed, Phys. Rev. Lett., i?, 1490(1974 ) .

103. E. Stender, N. Trautmann and G. Herrmann, Radiochem. Radioanal. Lett., 4.2, 291(1980).

104. H. Feldstein and S. Amiel, Nucl. Instr. Meth., 113. 181(1973).

105. W.-D. Schmidt-Ott and K.S. Toth, Nucl. Instr. Meth., 1_2J, 97(1974).

106. H. Schmeing, V. Koslowsky, M. Wightman, J.C. Hardy, J.A. MacOonald, T. Faestertnann, H.R. Andrews, J.S. Geiger and R.L. Graham, in ref. 5 8, p. 335.

107. O.C. Aumann, R. Presuhn and D. Weismann, Nucl. Instr. Meth., 165. 35( 1979 I.

108. W.J. Wiesehahn, J.M. 0'Auria, H. Oautet and 8.0. Pate, Can. .1. Phys., 51, 2347(1973).

109. E. Stender, Thesis, Johannes Gutenberg-Universitåt, Mainz, 1979.

110. G. Skarnemark, private communication, June 1981.

111. K. Broderi, G. Skarnemark, T. Bjørnstad, D. Friksen, I. Haldorsen, N. Kaffrell, E. Stender and N. Trautmann, J. inorg. nucl. Chem. , 43, 765(1981).

112. G. Skarnemark, K. Brodén, Mao Yuri, N. Kaffrell and N. Trautmann, Radio- chim. Acta., 3J. 97(1983).

113. H.L. Ravn, Phys. Rep., li, 201(1979).

114. L.C. Carraz, I.R. Haldorsen, H.L. Ravn, M. Skarestad and L. Westgaard, Nucl. Instr. Meth., 141, 217(1976).

115. H.l. Ravn, S. Sundeli and L. Westgaard, Nucl. Instr. Meth.. 123. 131(1975).

116. R. Kirchner and E. Roeckl. Nucl. Instr. Meth., 122, 307(1975).

117. R. Kirchner and E. Roeckl, Nucl. Instr. Meth., J_3_1 , 171(1975).

118. I. Langmuir and K.H. Kingdom, Science. $2, 58(1923),

119. M.F. Harrison, AERF Report GP/R 7505, (1958).

120. A. latuszynski and V.I. Raiko, Nucl. Instr. Meth., 120, 69(1974).

121. H. Shrrud, G. Engler, I. Yoresh and E. Skurnik, in ref. 58, p. 349. 85

122. J. Pelletier, C. Pomot and .1. Cocagne, J. Appl. Phys.. 50. 4517(1979).

123. J. Crank, "The Mathematics of Diffusion", Clarendon Press, Oxford,

124. 6. Rudstam, Research report IF B5, 1975.

125. H. Schafer, "Chemische Transport reaktionen", Verlag Chernie, Weinheim, 1962.

1?6. P. Kofstad, Tidsskr. Kjemi, Bergv. Metallurgi, 2_5, 33(1965).

127. G. Sidenius and 0. Skilbreid, Proc. Int. Symp. on "Electromagnetic Separation of Radioactive Isotopes", Vienna, 1960 (Springer Verlag, Vienna 19611, p. 243.

128. J.H. Freeman and G. Sidenius, Nucl. Instr. Meth., Ifjj • 477(1973).

129. H.L. Ravn, S. Sundeli, L. Westgaard and E. ftoeckl, ). inorg. nucl. Chem. , 3_7, 383(1975).

130. P.L. Reeder, I..J. Alquist, J. Lin and J.F. Wright, Mucl. Instr. Meth., 133 , 501(1976).

131. P. Hoff, L. Jaccibsson, B. Johansson, P. Aagaard, G. Rudslam, and H.-IJ. Zwicky, Nucl. [nstr. Meth.. J_72, '13(1980), and in rof. B?, p. 381.

132. J.C. Putaux. J. Obert, L. Kotfila, B. Roussiere, J. Sauvage-Letessier , C.F. Liang, A. Peghaire, P. Paris and J. Giroux. in ref. 62. p. 321.

133. C.r. Liang, P. Paris, A. Peghaire, B. Weiss and A. Gizon, in ref. 22, p. 487.

134. D.L. Hildpnbrand. in ref. 59, p. 357.

135. J.W. Green, G.D. Blue, T.C. F.hlert and J.L. Margrave, J. Chem. Phys., 4J, 2245(1964).

136. T.C. Ehlert, G.D. Blue, J.W. Green and J.I. Margrave, .1 . Chem. Phys., 4J., 2250(1964 I .

137. K.F. Zmbov and .1. L. Margrave, Advanced Chem. Ser., 12, ?r.7(196fl).

138. B.K. Anrus and S. Oatz. J. Chem. Phys.. 6j>. 4468(1977).

139. D.L Hijdpnbrand, J. Chem. Phys., G_2, 3074(1d75)

140. D.L. Hildenbrand. J. Chem. Phys., 65, 614(1976).

141. O.L. Hildenbrand, J. Chem. Phys.. 66, «788(1977).

142. P.O. Kleinsi.htnidt , 0. Cnbicciott i and O.L. Hildenbranri, .). riectrochem. Sot., 125. 1543(19781.

143. K.H. Lau and D.L. HUdf nbrand, J. Chem. Phys., 2J.. 1572(1979).

144. P.C. Nnrdine and O.E. Rosner, Semi Annular Progress Report "Aerodynamic Ipvita'.ion Reactor ftudier. of Fluorine Reactions with Refreartorv ChPramics" , 0c':. 1, 1979 - March 31, 1980. Yalp Uni vi?rs 11. v. New Havt'ii Connec tIL ut . 86

145. T. Bjarnstad, K.Å. Gustafsson, O.C. Jonsson, V. Lindfnrs, A.M. Poskanzei and H.L. Ravn, I SOIDF internal report, 9.5.1980.

US. P. Hoff, O.C. Jonsson, E. Kugler and H.L. Ravn, Nucl. Instr. Meth.. _?_2_1_. 1131 1984) .

14 7. S.E. Turner, Anal. Chem., ZÅ. 1457(1956).

148. 0. Reifenschweiler, in Proc. Int. Conf. on "Modern Trends in Activation Analysis", National Bureau of Standards, Gaithersburg, Maryland, Oct. 7-11, 1968, NBS Special Publication 312 (J.R. OeVoe and P.O. La Fleur, eds. ) , p. 905 ( 1969 ) .

K9. W.W. Meinke and R.W. Shideler, Nucleonics, 2_0 , 60(196?).

150. W.E. Mott and J.M. Orange, Anal. Chem., 37, 1340(1965).

151. R.F. Coleman and T.B. Pierce, Analyst, 9_2 , 1(1967).

152. W.O. Ehmann, Fortschr. chem. Forsch., J_4, 49(1969).

153. S.J. Bone, R.A. Greig and R.E.J. Porritt, Atomic Energy in Australia, 11, 24(1970).

154. M.T. Kelley, Pure Appl. Chem., 23.. 65(1971).

155. J. Hoste, J. Op de Beeck, R. Gijbels, F. Adams, P. Van den Winke.1 and O.De Soete, "Instrumental and Radiochemical Activation Analysis", CRC - monoscience series, Butterworths, London, 1971.

156. O.E. Wood, Nucl. Instr. Meth.. 92, 511(1971).

157. K. Nagorny and K. Zietlow, Chemie-Ing.-Techn., U., 1201(1972).

158. 0. Brune, Kemisk Tidsskrift, 1^8. 36(1972).

159. J. Op de Beeck and J. Hoste, Analyst, 9 9, 973(1974).

160. S.S. Nargolwalla and E.P. Przybylowicz, "Activation Analysis with Neutron Generators", John Wiley £ Sons, 1973.

161. T. Bjørnstad and. J. Alstad, Tidsskri. Kjemi, Bergv., Metallurgi, 36, 15(1976).

162. K. Bernstein, Int. lab., Jan/Feb.. 85(1981).

163. M.E. Pelps, Semin. Nucl. Med., 7, 337(1977).

164. M.M. Ter-Pogossian, M.E. Raichle and B.E. Sobel, Scient. Amer., fl£t., 141(1980).

1H5. G.A. Andrews and CL. Fdwards, JAMA, Z21. 1100(1975).

1G6. G.S. Johnston, A.F. Jones, M.S. Milder and R. I. . Frankel, J. Surg. Oncol., 5, 529(1973).

167. R.E. Hpnkin, .1.1. Ouinn and P.E. Weinberg, Radiology, iOJt. 595(1973). 87

168. E.B. Silberstein. A. Kornblut, O.A. Shumrick and E.L. Saenger, Radiology, 110, 605(1974).

169. J.B. van rier School, A,S. Groen and .1 . de Jong, Thorax, 2.1, 51T ( 197 2 ) .

170. S.O. Richman. S.M. Levenson, A.E. Jones and G. S . Johnston, Sem. Nucl. Med. , 5, 103(1975).

171. H.B. Hupf and J.F. Beaver, Int. 3. Appl. Rad. Isotopes, 2±. 75(1970).

172. l.C. Brown, A.P. Callahan, M.R. Skidmore and T.B. Wilson, Int. J. Appl. Rad. Isotopes, 24.. 651(1971).

171. N.N. Krasnov, P.P. Pmitriyev, I.O. Konstant inov, N.A. Konjakin. A.A. Ponomarev, A.A. Ognev, A.E. Romanchenko and V.M. Tiiyev, in Proc. Conf. on "The Uses of Cyclotrons in Chemistry, Metallurgy and Biology", St. Catherine's College, Oxford, ?2-2"3 Sept. 1969, C.B. Amphlett (ed.), Butterworths, london, 1970, p. 159.

174. F.E, Little and M.C. lagunas-Solar, Int. J. Appl. Rad. Isotopes, 34. 61 ( 1 9 8 1) .

175. I .J. Anghileri, P. Thouvenot , F. Brunotte. C. Marcha.1 and .1. Robert, Eur. J. Nucl. Med., _7, 266(1982).

176. L.P. Lavender (ed.), "Clinical and Experimental Applications of Krypton 81m", Brit. 1. Radiol., Special Report No. 1_5 , 1978.

177. H. Yamamoto, Y. Ikeda, K. Kawade, T. Katoh and T. Nagahara, J. inorg. nur.l . Chern.. 4_2, 855(1981).

178. F. Monnand and B. Fogelberg, in ref. 4, p 503.

179. R. Stippler. F. Munnich, H. Schrader, K. Hawerkamp. R. Decker, B. Pfpiffer. H. Wollnik. E. Monnand and F. Schussler, Z. Physik, A285. 267( 1970) .

100. W.8. Walters, C. Chung. O.S. Brenner, R. Gil], M. Shmid, R. Chrien, H.-r. linu, G. Gowdy, M. Stelts, Y.Y. Chu, F.K. Wohn, K. Sistemich, H. Yaniamnto and R. Pelry, in ref. 5. n. 557.

181. E. Michelakakis, W.O. Hamilton. P. Hungerford, G. Jung, B. Pfeiffer and S.M. Scott, .1. Phys . G: Nurl. Phys. . 8, 111(1982).

IB?. B. Pfeiffer. F. Schussler, 1. BJairhot, S.I. Feenstra, '. van Klinken, H. Lawin, F. Monnand, G. Sadler. H. Wnl'nik and K.D. Wiinsch, 7. . Physik, A_2 8J . 19 1 11 9781 .

181. G. Engler, Y. Nir - F.I, M. Shmid and S. Ami el . Phys. R»>v. C. ig, H48( 1979) .

184. F. Mcuinand, 1. Blachot, l.P. Bncquft , F. Schussler, H. Lawjn, G.Sadler, K. Sistern.ich, G. lung and K.O. Wiinsch, Vcrh, nt . Phyiiknl. Ges., 4, 845(1978)

185. ll. Kpvser, H. Berg, F. Mimiiich, K. Hawerkamp H. Schirtrtt.r. P. P'eiffer and F. Monnjnd, 7. . Phy.'jk. A?B9, 407(1978). 88

186. G.M. Gowdy, R.E. Chrien, Y.Y. Chu. R.L. Gill, H.I. Liou, M. Shmid, M.L. Stelts, K. Sistemich, F.K. Wohn, H. Yamamoto, O.S. Brenner., T.R. Yeh, R.A. Meyer, C. Chung, W.B. Walters and R.F. Petry, in ref. 5, p. 562.

187. R. Decker, K.D. Wunsch, H. Wollnik, G. .Tung. J. Muri7el, G. S.ieqert and E. Koglin, Z. Phyzik, A 3 0.1, 165(1981).

188. D.S. Brenner. M.K. Martel, A. Aprahamian, R.E. Chrien, R.L. GiJl. H.I. Liou, M. Shmid. M.L. Stelts, A. Wolf, F.K. Wohn, D.M. Rehfield, H. DeTbakhsh and C. Chung, Phys . Rev., C2.6, 2166(1982).

189. F. Schussler, B. Pfeiffer, H. Lawin, E. Monnand, .1. Munzel, .1. A. Pinston and K. Sistemich, jn ref. 5, p. 589.

19 0. M. Shmid, Y.Y. Chu, G.M. Gowdy, R.L. Gill. H.I. liou. M.L. Stelts, R.E. Chrien, R.F. Petry, H. Dejbakhsh, C. Chung and D.S. Brenner, in ref. 5, p. 576.

191. R.L. Gill, R.E. Chrien, M. Shmid, G.M. Gowdy, H.I. Liou, n.S. Brenner, F.K. Wohn, K. Sistemich, H. Yamamoto, C. Chung and W.B. Walters, in ref. 5, p. 569.

192. R.L. Gill, M. Shmid, R.F. Chrien, Y.Y. fhu, A. Wolf. D.S. Brenner, K. Sistemich, F.K. Wohn, H. Yamamoto. C. Chung and W.B. Walters, Phys. Rev., C21. 1732(1983).

193. H.W. Taylor and B. Singh. .1. Phys. G: Nucl. Phys., 5_. 1445(1979).

194. H. Yamamoto, Y. [keda, K. Kawade and T. Katoh. ). inorg. nucl. Them., 4_2. 1539(1980).

195. I.D.ll. Fbong and R.R. Roy, Z. Phys.. A2j)9. 209(1981).

196. B. Pfeiffer, J.P. Bocquet, A. Pinston, R. Rousille, M. Asghar, G. Bailleul, R. Decker, .1. Greif, H. Schrader, G. Sieqert, H. Wollnik, J. Blachot, E. Monnand and F. Schussler. J. Physique, .38. 9(1977).

197. H. Yamamoto, C. Ichihara, K. Kawade and T. Katoh, J. Phys. Soc. Japan. 41. 729(1976).

198. Y. Ikeda, H. Yamamoto, K. Kawade. T. Takeuchi, T. Katoh and T. Nagahara. J. Phys. Soc. Japan. 4J, 1039(1979).

199. J.A. Pinston R. Roussille, G. Sadler. W. Tenten, l.P. Bocquet. B. Pfeiffer and D.D. Warner. 7. Phvsik. A28_2. 303(1977).

200. S.W. Yates and A.J. Filo, J. inorg. nucl. Chem.. 40. 963(1978).

201. M. Totsuka, S. Fujita, K. Mio, K. Kawade, H. Yamamoto. T. Katoh and T. Nagahara, J. Nucl. Set. Technol.. 19, 765(1982).

202. U. Keyser, H. Berg, F. Munnich, B. Pah.lmann, K. HawerMmp, B. Pfpiffpr, H. Schrader and E. Monnand. Verh. Otsch. Phys. Ges., 9. 1161(1980).

203. CM. Lederer and V.S. Shirley (eds.). "TabJes of Isotopes", 7th edition, John Wiley & Sons, INC., New York, 197(1.

204. J.I". Hill.. H. Yamamoto and A. Wnlf, Phys. Rov.. C27, 2857(1901). 89

205. S.M. Scott, W.O. Hamilton, P. Hungerford, O.D. Warner, G. Jung, K.n. Wunsch and B. Pfeiffer, .1. Phys. G, £, 1291(1980).

20G. M. Sakai, Nucl. Phys., A_KK. 301(1967).

207. fi.F. Casten, O.D. Warner, D.S. Brenner and R.L. Gill, Phys. Rev. Lett., 41, 1433( 1981 ) .

208. F. Iachello, "Interacting Bosons in Nuclear Physics", Plenum Press, New York, ( 1973) , p. 1.

209. .l.C. Hardy, in ref. 14, p. 515.

?10. M. Arnould, in ref. 14, p. 375.

211. G. Audi, M. Fpherre, C. Thibault, R. Klapisch, G. Huber, F. Touchard and H. Wollni k, in ref. 14 , p. 281.

212. C. Thibault, M. Epherre, G. Audi, R. Klapisch, G. Huber, F. Touchard, 0. Guillemaud and F. Naulin, in ref. 14, p. 29 1.

213. M. Epherre, G. Audi, C. Thibault and R. Klapisch, in. ref. 14, p. ?99.

2K. "The 1975 Predictions", Atomic and Nuclear Oata Tables, J_7 , No. "5-6, (1976).

?15. W.D. Myers, in ref. 2U, p. 411.

216. K. Takahashi, H. v. Groote and E.R. Hilf, "Gross Theory of Nuclear masses and Radii. I. Masses". IKDA 76/26, (1976).

217. P.A. Seeger and W.M. Howard, in ref. 214, p. 428.

?18. S. Uran and N. 7eldes, in ref. 214, p. 431.

219. E. Comay and I. Kelson, in ref. 214, p. 463.

220. J. Janecke, in ref. ?14, p. 455.

??1. J. lanecke and F.P. Fvnori, in ref. ?14, p. 467.

??.2. E. Lund, K. A lek lett and G. Rud s lam, i.n ref. 14, p. 451.

??i. A. Kerek, G.B. Holm, l.-F. De Geer and S. Porq, Phys. lett., B44. 252(19 73) .

224. W.-D. Lauppe. K. Sistemich, T.A. Khan, H. Lawin, G. Sadler, H.A. Selic and O.8.W. Schult, Proc. Int. Conf. Nuclear Physics, Tokyo, 1977, ). Phys. Soc. .lapari, 44(SUPD.1. ) . 335(1978).

7?*i. l.S. Oehesa, W.O. lauppe, K. Sistemich and J. Speth. Phvs. Lett., R74 . 10") ( 1178) .

22P.. P.I. Nolan, jn ref. 11. p. 103.

221. K. Okano. V. KavtMse, K. Kawarle. H. Vamamoiro, M. Han«<1a , T. Krftnh anrl I. Fujiwara. in ref. 59. p. 115. 90

226. M. Fujioka, T. Shinozuka, E. Tanaka, Y Arai, S. Hayashibe and T. Ishimatsu, in ref. 59, p. 121.

229. H. Schmeing, J.C. Hardy, E. Hagberg, W.L. Perry, J.S. Wills, J. Camplan and B. Rosenbaum, in ref. 59, p. 47.

230. S. Morinobu, I Katayama and H. Nakabushi, in ref. 5, p. 717.

231. E. Ye Berlovich et al., izv. Akad. Nauk SSSR. Ser Fiz, 4j0, 2036(1979). PAPER P.I. INORG. NUCL. CHEM. LETTERS Vol. 11, pp. 729-735, 197S. Pergamon Press. Printed in Great Britain

THE COMBINATION OF THE GAS JET RECOIL TECHNIQUE WITH THE

FAST CHEMICAL ON-LINE SEPARATION SYSTEM SISAK

N. Trautmann+, P.O. Aronsson*, T. Bj6rnstad++, N. Kaffrell+,

E. Kvåle++, M. Skarestad++, G. Skarnemark* and E. Stander*

The SISAK Collaboration

(Received 10 June 1975)

ABSTRACT

The fast chemical on-line separation system SISAK has been connected to a gas jet recoil transport arrangement. Pure fractions of short-lived La, Ce and Pr nuclides have been isolated from complex mixtures of reaction products obtained by thermal-neutron-induced fission of 2 3 U5.

INTRODUCTION

The Gas Jet Recoil Transport (GJRT) sysxem has been used successfully to transport short-lived reaction products from a highly radioactive irradiation site to a low background area

Institut f (Ir Kernchemie, Universitåt Mainz, D-65OO Mainz, Germany Department of Nuclear Chemistry, Chalmers University of Technology, Fack, S-UO22O Goteborg 5, Sweden Department of Nuclear Chemistry, University of Oslo, Oslo 3, Norway ?30 Fast Chemical On-line Separalinn System SISAK Vol. 11, No. 11

(1-5). This technique has not only been applied for charged- particle reaction products (1,4) but also for radioactive products resulting from neutron-induced fission (3), spontaneous fission (2,5) and radioactive decay (6). The nuclides recoiling out of the irradiated target are thermalized in a gas, usually He mixed with some organic compounds to form the necessary clusters. The products are then transported through a capillary, the length of which may be several meters. In most of these experiments, the nuclides transported by the jet have been collected by impinging on a collector plate (1-5). Only in a few cases, attempts have been made to separate individual elements from product mixtures by combining the jet system with chemical procedures (4,6-9). So far, no severe efforts have been undertaken to utilize the GJRT technique in combination with a continuous chemical separation system for the selective on-line isolation of elements from complex mixtures of reaction products, e.g., fission products. In this paper, we will report on the connection of the fast, continuous radiochemical separation system SISAK (10,11) to a GJRT arrangement and its application to the separation of short- lived fission products. The chemical system previously used for studies on neutron-rich isotopes of La, Ce and Pr (12,13,14) as well as the gas jet have been modified to enable the successful achievement. In the y-ray spectra of the La, Ce and Pr fractions nuclides with half-lives down to B s have been observed.

EXPERIMENTAL

This paper will mainly concentrate on the Ce system although chemical systems for La and Pr were tested and found to work properly. The whole system shown in Fig. 1 consists of three main parts: the GJRT arrangement, the mixing-degassing unit and the chemical separation system.

The GJRT system. The target, consisting of 450 vg 235U covered with a 400 pg/cm Al layer, is situated in a recoil chamber of about 12 cm volume. The recoil chamber is placed Vol. 11, No. i 1 I-ast Chemical On-line Separation System SISAK

C;«. N,

l.Z5HBN0j 1M HND; IHHNO, 1H H,SO, 0 IMHjSO, D.ObH SAA D.2M K,Cr,Q, 0 05MK,Cr,0-

FIG. 1 Flow diagram showing the system for the isolation of Ce nuclides. M:ga3-liquid mixer, Dg:degassing unit, FP:fission products, C1, C2 , C3:mixer-centrifugal separator units, E: HDEHP/PVC column, D0-D7:Ge(Li) detector positions, Y0-Y4:Ce recovery yields. YO > 80 % (calculated), Yl > 90 % (measured), Y2 -v. 80 % (measured), Y3 > 30 % (measured), Y4 ^ 55 % (measured) in one of the beam holes close to the core of the Mainz TRIGA reactor, having a thermal neutron flux of about 10 11 n cm — ?s —1 The fission products recoiling out of the target are thermalized in a 1:1.4 mixture of C^Hj^ and N„, the former substance serving as a cluster producer (15,16). The gas pressure in the recoil chamber is kept at 1500 torr, and the gas flow rate at some 20 cm s . When thermalized, the fission products are transported to the mixing-degassing unit via a 7 m poly- ethylene capillary (inner diameter 1 mm). The transport time from the target to the end of the capillary has been measured to amount to about 1 s.

The mixing-degassing unit. After the transport through the capillary, the gas is mixed in a static mixer with the first aqueous phased. 25 M HNOj for Ce and Pr or, if studying La, a UNO, solution of pF 1.4). The temperature of this solution is kept at about 90 'C, since a strong influence of the temperature 732 past Chemical On-line Separation System SISAK Vol. 11, No. I ]

on the fission product dissolution was noticed. The gas-liquid mixture is then fed into a degassing unit, simply consisting of a coned funnel with a tangential inlet.

In this funnel, the C2H. -N2 mixture is swept off together with more than 95 % of the noble gas activity.

The chemical separation system. The degassed liquid is pumped to a second static mixer, where it is contacted with the first organic phase (2 M HDEHP in Shellsol T). In this step, most of the Y, Zr, Nb and Mo as well as part of the Br and I are extracted into the organic phase. The phases are then separated in the first H-centrifuge C 1 (17). The Ce, remaining in the aqueous phase from C 1 together with the majority of the fission products, is oxidized to the

tetravalent state by adding HNO3, H?S04 and K2Cr207. It is then extracted almost quantitatively (in C 2) into the second organic phase (0.3 M HDEHP in Shellsol T). In C 3, Ce is stripped from the organic phase by using H„S0. , sulfamic acid

(SAA) and H2<32. After re-oxidation achieved by adding the proper amount of K-C^Oy , the Ce is adsorbed on an HDEHP/PVC column, as described in ref. (13).

Measuring equipment and data evaluation. The measuring system consisted of a Ge(Li)-detector (efficiency 6.4 %) connected to a <+ K multichannel analyzer. The data obtained were stored on magnetic tapes and evaluated with computer programs.

RESULTS AND DISCUSSION

As mentioned above the dissolution of the fission products is influenced by the temperature of the first aqueous solution. Generally, a higher temperature to an increased yield of the desired element and a better decontamination from interfering elements. This is probably due to a partial extraction of the undestroyed clusters into the organic phases at low temperatures. This assumption is supported by the fact that in the cerium fraction the amount of the trivalent lanthanides decreased rapidly with increasing temperature of the first aqueous solution, whereas the yield of Ce is increasing. At temperatures of about 90°C, the strong 397 keV peak of 1U1*La, which served as a monitor Vol. ] 1, No. ] 1 Fasl Chemical On-line Sepaiation System SISAK 733

for trivalent ianthanides, disappeared in the Ce fraction. On-line measurements were performed in the different positions indicated in Fig. 1. The column E as well as the aqueous phase leaving the unit C 3, showed almost no other activities than those from Ce nuclides and, to some extent, the grown-in Pr daughters. A y-ray spectrum of the Ce fraction on the column E 14 7 is shown in Fig. 2. In this spectrum, the peaks of 56-s Ce 1 4 9> and48-s Ce can be identified, as well as those of the longer- lived isotopes 3-min 145Ce and 13-min 145Ce (12,14). The efficiency of the gas-liquid transfer of Ce was determined in the following way: the activity of the strong 317 keV peak of 146Ce was measured in a Ce sample collected on the HDEHP/PVC column and compared to a "direct catch" measurement, obtained by passing the gas through a fiber glass filter. The ratio column/ direct catch was calculated to be 0.1. From this value and the measured chemical yields indicated in Fig. 1., the Ce dissolution yield in the gas-liquid mixer can be calculated to be >_80 %. By using the chemical separation procedures for La and Pr, strong sources of 11-s La and of 6-s Pr, respectively, were produced. The results obtained show that the combination of the GJRT technique with a fast, continuous chemical separation system like SISAK enables on-line investigations of specific short-lived fission products. Detailed spectroscopic studies by y-y- coincidence and even y-y-angular correlation measurements should be possible. In addition to studies of fission products,the described system can be used for the investigation of nuclei formed in spallation and heavy-ion-induced reactions.

ACKNOWLEDGEMENTS

The authors express their gratitude to Professors G. Herrmann, A.C. Pappas and J. Rydberg for their interest in this work. We wish to thank Mr. R. Heimann for his help and the staff of the Mainz TRIGA reactor for numerous irradiations. Financial support from the Swedish Atomic Research Council, the Norwegian Research Council for Science and the Humanities and the Bundesministerium filr Forschung und Technologie is gratefully acknowledged. Fasl Chemical On-line Separation System SISAK Vol. 11, No. 11

UOO 1600 Channel number

FIG. 2

y-ray spectrum of neutron-rich cerium isotopes measured in position D 6 of Fig. 1

REFERENCES

1 . R.D. MACFARLANE, R.A. GOUGH, U.S. 0A> EY and Il.F. TORGERSON, Nucl. Instr. Methods 7J_>28 5 (1 3 f. 9 ) 2. K. WIEN, Y. FARES and R.D. MACFARLAfJE, ?lucl. Instr. Methods 103, 181 (1972) 3. H. DAUTET, S. GUJRATHI, W.J. WIESEHAHN, .7.M. D'AliPTA and B.D. PATE, Nucl. Instr. Methods 107, U9 (1373) Vol. ! 1. No. 11 Fast Chemical On-line Separation System SISAK 7,

4. K.L. KOSANKE, W. MCHARRIS and R.A. WARNER, Nucl. Instr. Methods llb_, 151 (1974) K.L. KOSANKE, Thesis, Michigan State University, East Lansing 1973 (COO-1779-76) 5. H.G. WILHELM, H. JUNGCLAS, H. WOLLNIK, D.F. SNIDER, R. BRANDT and K.H. LUST, Nucl. Instr. Methods 1J_5, 419 (1974) 6. P. PUUMALAINEN, J. AYSTb' and K. VALLI, llucl. Instr. Methods 112, 485 (1973) 7. R.J. SILVA, J. HARRIS, M. NURMIA, K. ESKOLA and A. GHIORSO, Inorg. Nucl. Chem. Letters _6, 871 (1970) 8. D.C. AUMANN and D. WEISMANN, Nucl. Instr. Methods 1\1_, 459 (1974) 9. K.-H. HELLMUTH and K. VALLI, Nucl. Instr. Methods 1_2_5, 99 (1975)

.10. P.O. ARONSSON, B.E. JOHANSSON, J. RYDBERG, G. SKARNEMARK, J. ALSTAD, B. BERGERSEN, E. KVÅLE and M. SKARESTAD, J. Inorg. Nucl. Chem. 35, 2397 (1974) 11. P.O. ARONSSON, Thesis, Chalmers University of Technology, Goteborg (1974) 12. P.O. ARONSSON, G. SKARNEMARK and M. SKARESTAD, J. Inorg. Nucl. Chem. ^6, 1689 (1974) 13. P.O. ARONSSON, G. SKARNEMARK and M. SKARESTAD, Inorg. Nucl. Chem. Letters 1C), 499 (1974) 14. P.O. ARONSSON, G. SKARNEMARK, E. KVALE and M. SKARESTAD, Inorg. Nucl. Chem. Letters 10, 753 (1974) 15. W.J. WIESEHAHN, J.M. D'AURIA, H. DAUTET and B.D. PATE, Can. J. Phys. 5J., 2347 (1973) 16. R.J. SILVA, N. TRAUTMANN, M. ZENDEL and P. DITTNER, to be published 17. H. REINHARDT and J. RYDBERG, Acta Chem. Scand. _2_3, 2773 (1969) PAPER P.II. Raflioehimioa Acta 23, 38-103 (197(i) tf~f) by Akadomisehe Verlagsgesollschaft, Wiesbaden 1976

Decay Properties of Some Neutron-Rich Praseodymium Isotopes

JJy (i. SKAK.NE.MAKK*, E. STE.NDKR**, N. TRAUTMANN**, P. 0. ARONSSON*, T. UJORNSTAD***, N. KAFFREI.L**, E. KvÅi.E*** and M. SKARESTA»***, The SISAK Collaboration

With o figures. (Keceived June 11. 1976; in final version August 20, 1976)

• Department of Niirk-nr Chemistry, Chnlmor« University of Technology, Fack, S-402 20 d'otcborg fl, Sweden. •• Institut fiir Kernchemie, Johannes Gutenberg UniveraitSt, Postfaeh »980, D-6500 Mainz. (Jermany. •*• Depart ment of Nuclear Chemistry, University of Oslo, Oslo 3, Norway. Summary Half-lives and some y-rny energies of 118]'r (2.0 min) Xeutron-rieh Pr isotopes producer] in the thermal neutron- and 1491V (2M mill) were reported by OHYOSHT et al. [4|. induccd fission of 2a5U liiive been investigated by means of These nuclidcs were isolated from fission products by y - y coincidence experiments. The nuelides have been sepa- an electromigration teehniciue. Shortly after that. rated from the fission product mixture, using the fast chemical separation system S1SAK in connection with a gas jet recoil AKONSSO.N et al. [5] published energies and relative transport system. Tho results include assignments of several intensities for several new y-Iines attributed to the new -/-ray energies and partial decav schemes for 147Pr, 148Pr, ll8 150 decay of Pr (2.2 4- 0.1 min) on the basis of y—y ""}'<• and Pr. coincidence measurements. Tlio ]>raseodymium sources were produced by the milking technique on cerium fractions chemically separated from fission product Neutronenreiche Prnscodym-Isotope, horgestellt durcli Npal- mixtures. The presence of l4SPr in the milked praseo- tung von -35U mit thcrmischen Neutronen, wurden mittels •/--/ Koitizideiiznu-ssungcn untersucht. Die Ahtrcnnung der dymium fraction could be excluded due to a delay Aktivitiitpn aus dem Spaltprnduktgemisch gelang (lurch time between production and separation of the cerium Ankupplung eines (las-Jet-Transportsystems an das schnelle isotopes, in which "9C'e had decayed. chemische Trennverfahren S1SAK. Mit den erhalh-nen Er- Some evidence for an ~2.3 min activity attributable gebn issen korinen einige neue "/-Linien zugeordnet und partielle I49 117 11B I49 JM to Pr was found by HOFFMAN and DANTF.LS [1] in Zcrfallssc-hemata fur Pr. Pr, Pr und Pr amVestellt 150 werden. irradiations of Xd with bremsstrahlung. The exist- ence of a 2..'5 min I49Pr was confirmed by VAN KLINKEN 150 Késiiiné and TAPF [,. f|.oln the growth-and-deeay curve of the only known y-ray transition assigned so far to the decay of IB0Pr, which occurs at 131 kcV. 1. Introduction From the neutron-rich Pr isotopes, the nuclides '«Pr 1J8 and Pr with half-lives of 12 min and 2.2 min, 1. I). ('. HOFFMAN and W. H. DANIKLS. .). inorg. nml. CIH-III. respectively, were first identified by HOFFMAN and 2«. 17(i!) (1B04). DANIKT.S [1 ] who separated praseodymium radio- 2. .F. A. Pl.VSTOX, l{. Roi'KSILLK, ('•. BAIM.EII,, .1. B1.ACIIOT. .1. P. ROCQI'ET, K. MONNAND, B. PKKIFFKH. H. SCIIHAIIKH and chemieally from fission products. For the assignment V. Scin-ssi.F.u. \ucl. Pliys. A 2-1«. Sflii (1»7.1). of the 12 min Pr-aetivity to mass number 147 they :i. M. DOHIKKXS and 1,. DOUIKKXS-VAM-HAIT. Z. Physik also performed (y,p)-reactions on enriched 148Nd. In A 27», :nr»( i»7.r>). addition to the half-lives they obtained a partial decay 4. K. Ouvosin. A. OIIVOSIII, T. TAKKMI and M. SIIINAIIAWA, scheme for "'IV «ml ^„-values of 2.7 -± 0.2 MeV and Nucl. Sei. Techn. 10. Ull (Ill7:t)- ii. P. O. AROXSSON, (i. NKAHNKIMIIK, K. KVAI.B anil 4.5 ± 0.4 MeV for "'Pr and "»Pr, respectively. The M. SKAKKSTAII. Inorg. nucl. ('hem. I.CIIITH 10, 7">:i (1074). decay of "'Pr has also been thoroughly studied II. J. VAN KLISKKN nnd L. M. TAIT, Xml. I'liys. AIM). 47M recently by PINSTON et al. [2j ; -ing samples which were mass-separated from f . products and by 7. T. E. WAKD. N. A. MOHCOS and V. K. KritoiM, Pliy». Hev. DORIKKNS and DORIKENS-VANIT.AET [3] who obtained C 2, 24(0 (I!)70). I4 8. P. O. AIIOSSSON, (j. SKAIINKMAUK mid i! SKAHKSTA». the activity via a *Xd (y,p)-reaction. Inorg. niu-1. Chein. Ix-tters JO, 41HJ (1974). Decay Properties of Some Ni-iitron-Kich Praseodymium Isotopes 99

1"Pr r,,2=12itiin

139S 2 1350 3 1310 8 1254 3

792 6 7«9 3 i8*2"888

517 0 463 6

314 7

214 5

,1i 49 9 i O D i?00 1400 1600 1800 "Nd Channel number Kii». 1. -/-ray singles spectrum of the neutron-rich Pr isotopes, Fig.2. Partial decay scheme of "7Pr. The scheme is baaed on mainly 147Pr, 118Pr and I48Pr the coincidences shown in Table 1

The present investigation has been carried out using A typicaly-ray singles spectrum of the three longer-lived the continuous radiochemical separation technique nuclides 147Pr, J4BPr and "9Pr is shown in Fig.l. The SISAK [9j in combination with a gas jet recoil trans- decay of the various nuclides is briefly discussed below. port system [10]. y—y coincidence measurements have been performed on 147Pr, "8Pr, J48Pr and 160Pr with 3.1. The decay of 147Pr chemically separated fractions. The decay scheme of 147Pr is shown in Fig.2. The scheme is based on the coincidence data given in Table 1. In order to discuss the present scheme, we 2. Experimental procedures have to review some of the earlier reported data on The radionuclides studied in the present work were the decay of 147Pr and excited levels in I47Nd. produced by thermal neutron-induced fission of 236U With a 14eNd (d,p)-reaction, WIEDNER et a/.[12] have in the Mainz TR1GA reactor. The fission products observed nine states in 147Nd below 600 keV. The were transported with a gas jet to the SISAK system, first excited state located at 78 keV which is in agree- and the Pr fraction was separated from the fission ment with the earlier reported decay scheme of product mixture as described in ref. [11]. HOFFMAN et al.[l]. This disagrees with the results of The y-ray singles and coincidence measurements were ROUSSILLE et al. [13] who have found an energy of carried out by using two Ge(Li)-deteetors with relative 50keV for the first excited state in the "«Nd (n,y)- efficiencies of 23.6°/0 and 6.4°/0 and an energy resolu- reaction. However, if we consider the assignment from tion at 1332 keV of 2.3 and 1.75 keV, respectively. the (d,p)-rcaction as an erroneous interpretation of the In addition, standard electronic equipment with the 50 keV level as the ground state, while the real ground possibility of spectrum stabilization was used. The state has not been excited in the one-nucleon transfer resolving time of the coincidence set-up was 2T = 25 ns

The coincidence events were stored in a 4K v. 4K 9. P.O.ABONSSON, U.K..JOHANSSON, .T.KYDHKHCI, G.SKAIIKK- matrix and subsequently transferred, event by event, MAHK, J. AI.HTAI), 13. BKIIOERSEN, E. KVAI.K nnri M. SKAHK- onto magnetic tape. All data were evaluated by mean« STAD, J. inorg. nucl. Chem. 86, 21)97 (1974). of appropriate computer programs. 10. N. TRA). U. P. 0. ABONSSON. (!, SKARMKMARK arid M. SKARESTAD, 3. ltcsults and discussion J. inorg. nucl. Clu-ni. «I. 1689 (1974). From the y y coincidence data obtained, decay 12. C. A. WlEONKR, A. HEtlSI.ER. 3. SOI.K «I1<1 .]. P. Wl'HM. 147 Nucl. Phy». A 108, 43S (19(17). schemes have been deduced for the nuclidc» Pr, Vi. K. KOII.SSH.MV, J. A. PINMTON. H. BURNER, H. K. KOCH l60 '"I'r, "«Pr and Pr. and D. HECK, Nucl. PhyK. A 24«, 380 (1975). Un) (•. SKAKVKMAKK. K. STKNDKK, N. TBAITTMANN. 1*. O. AKOVSSDN, T. K.KWXSTAI», \*. KAKFHKLT,, E. KVÅLK and M. SKAKESTAD

Table 1. The coincidence ihtta resulting from the. (heat/ of urPr. s (-"•.strong) indicates that the peak is higher than '.in (standard deviations) in the background, m (-• - medium) indicates ( - 'A a and \t (- weak) less than 1 a. The intensities listed are relative intensities /.,(«/„) - -• n

49.9 21.5 S H s s 77.9 (13.(1 s Hii.n Hlj.7 -s ^ (x) 127.9 42.9 s !H«.7 5.4 S s *• 214..r, 7.(1 249.0 (>.» 314.7 1110.0 :!2K.s 20.0 K R !• :«S.7 25.1 S « fc 38». 1 10.3 S R 413.7 T..3 4«7.() H.9 m 477.1) 2(1.5 517.(1 23.0 5r>4.7 33.8 a s s ;->77.i) 78.3 s 8 S S (»27.5 < 1 641.4 84.4 s s s »42.2 5.8 990.(1 8.5 lim.r, .- H3(i.r> 7.9 11S3.II 5.1 12(14..'i 7.» 1300.4 nu

The uncertainties in the energies are estimated to -{: 0.3 keV, and in the intensities to -}~ !0°/0 for the stronger ^-transitions (/,, •• 30»/,,) and ± 20% for the weaker ones. The coincidence matrixes should be regarded only as a presentation of the raw. unprocessed data obtained. Because the symbols s, in and w refer only to the statistical significance of a certain peak (which is not always the same if gating on the weak or the strong component of a cascade composed of two y-rays of different intensity), the matrix is not completely symmetric.

Table 2. Thr coincidence data resulting from the decay of l48Pr. For explanat ion of the letters in the matrix, see Table I

A1 /1/«V^ 7/0/ ^ p »o p » æ TJ* oc « x 10 o *+ ir: OT M x os o

148.0 < 1 I71./5 < 1 247.0 0.7 s s 301.8 100.0 ni m s s s s s s s in s s s s s s s 450.(1 5.3 s s «15.4 5.4 B 030.8 2.5 (1(10.3 3.6 m s «97.8 10.« 721.5 8.3 K25.5 3.1 809.4 7.5 OO3.fl 3.4 947.3 3.4 1023.2 9.0 1248.8 0.6 1357.9 10.4 1381.« 4.7 2133 2635 Decay Properties of Some Xeutron-RUrh Praseodymium Isotopes 101

Table It. The. coincidence data resulting from the decay of I49Pr. For explanation of the letters in the matrix, see Table 1

A\,(kcV) x aj M « a « x oi x -c fi "i w h ci -t h x x w ifi 6 o w cc so o o « « w w i«

332.8 53.2 305.9 35.7 300.3 35.7 400.3 33.1 432.0 41.3 reaction, a systematic shift of 50 keV in the level 0.127-0.078 MeV cascade which is not seen by us. sequence from the (d,p)-e.\periments has to be taken Our scheme is in good agreement with the independ- into account. Doing this, the (d,p)-data are in excellent ently determined schemes of PINSTON el al. |2| and agreement with the (n,y)-data. This is further sup- DOBIKENS el al. |3|. There is one disagreement con- ported by the results of the present work. cerning the reported 12(il-4(l keV cascade. Although The present decay scheme disagrees with the scheme we have seen the 1261 keV y-line in our singles spec- of HOFFMAN el al. [1| with respect to the energy of trum, we have found no evidence for a coincidence the first excited Jevel and the 0.61 0.335-0.105- relationship with the 4!) keV peak.

T„2=2.9min\ 0„ =3.0MeV

403 •

1113 4 3B5 9 1»5» f

270 > 1249 0 (2*) mi 1171 2 2*

0'

7K2 4 4* 1314 10* t

301 • 2'

0 0 0'

Fi(i.3. I'nrtinl «Ipiiiy wlipnie of ""Pr. The si'hpmp is baapcl on Fig.4. Pnrtinl dcrny Hchcnie of "»Pr. The schcinc is husctl on the iiiincidi-iif'PH HIIOUII in Tiiblr 2 the roiticidcni-e shnun in Tnhh1 .'i 102 G. SKARNKMAUK. E. STKNDER, N. TRAUTMANX, P. O. AROMSSON, T. BJORNSTAU, X. KAFFKELL, E. KVÅLE and 11. SKARESTAD

Table 4. The. coincidence data resulting from the decay of J50Pr. For explanation of the letters in the matrix, see Table 1

AV(kcY) ! i sl § S si I s•i» I ccS os = C s (rv o^ rain tro- r-m 54 5 9 72 0 6

46 9 r* to w o m 130.2 loo.o s m s w s s s »n 251.2 s s \v = 852.7 1 469 = 850.B (2*) m m 545.9 15.9 553.3 72O.(i 722.4

804.4 B2.7 25 1 2 (1 3 4 ) "O 852.7 o o 931.5 18.1 i I Illil.li 12.4

3.4. The decay of 160IV Fig.5. Partial decay scheme of I50Pr. The scheme is based on 150 the coincidences shown in Table 4 In this experiment, 6.2 s Pr was milked from a Ci: fraction containing 4.0 s 150Ce [8]. This half-life is close to the lower limit accessible with the prosent- M8 system, and in combination with the low cumulative •A.2. The decay of Pr 160 148 fission yield of O, only low-in tensity samples can Table 2 shows the coincidence data obtained for Pr. be obtained. Therefore only a preliminary decay The corresponding decay scheme is given in Fig, 3. 146 scheme can be deduced from the obtained coincidence Spins and parities are taken from the Ncl (t,p)-reac- data, which are summarized in Table: 4. The scheme tion data reported by CHAPMAN et al. [14] and the is shown in Fig. 5. I48Nd (d.d')-reaction data of VAN DER BAAN et al.[lo]. The levels at 130 and 381 keV correspond to the The levels at 301.8, 752.4, 909.6, 1171.2 and 1683.4keV 2J' and 41 levels of the K •-— 0 ground state rotational agree fairly well with those of the reaction experiment band [15]. The 676 level is assumed to be identical mentioned above, whereas levels at 917.2 and 2074.6 with the 672 keV level observed by VAN I>EK BAAN keV may correspond to the 911 and 2086 keV states et a!..[15\ in the IMXd (d,d')-reaction and the 67.5 keV found by (!HAI*MAN r.t al. [14]. level of CHAPMAN et a?.[14J in the148 Xd (t,p)-rcaction, The 1020 keV level found in the (d.d')-experiment [15J and may be interpreted as the 0' member of the should be identical with the 1023.2 keV state observed K = OjS-vibrational band. The 848 keV level (2'. 1") by us. The differences in the level energies between of VAN UER BAAN C( al. is presumably identical with the (d,d')-reaction work and the present decay study the 851 keV level (2') of CHAVMAN rial. In the (d,d')- arc typically 0-3 keV. Thus our level at 1249.0 keV reaction work|15| they found that the cross section and the {d,d')-reaction level of 1240 keV may not be for the 848 keV level (which is interpreted as the the same. Therefore this state at 1249.0 keV and the 2'" member of the K =-0 /J-vibrational band) was levels at 1659.9, 2434.9 and 2936.9 keV seem to be higher compared with the similar level in IS2Sm. They fed only in the /? "-decay. concluded that the 848 keV peak may be an unresolved doublet which also contain» some contribution from a 1" state. In our decay work, this level has split up 3.3. The decay of 14>Pr into two levels with energies of 851 and 853 keV. The present results confirm the earlier reported half- .An exact spin assignment of these two levels is not life of 2.3 min 11] of »19Pr. possible. However, from the decay pattern of the two The coincidence» obtained for "»Pr are shown in levels the 853 keV state may be interpreted as the Table 3. Krom these data a partial decay scheme, I member of the K --• 0 octupole band while the presented in Fig.4 has been deduced. This Hchemu includes the y-lincs at 108.5, 138.4 and 165.0 keV Mf already assigned to Pr by VAN KLINKEN and TAFP[6] 14. K. CIIAI'MAS. K. MCLATCIIIK «nil .1. K. KITCIIIN», XHCI, and several new, weak transitions. The levels are in Phys. A 18«, (KB (11172). good agreement with the results of HECK et al. [16], /5. J. (i. V.IN I'Ktt B.IAN. I*. K. CjlJUSTB.NSKN. .1. ItASMI'NNKN from the14B Nd (n.yj-reaction and with a study of the and 1', (). T.loM, Priviite conimiiiiiciition (11)75). l49 3 If!. 1). HKCK, I'riviilc ciininiiiiiiiiitinii (III7">). levels in Nd using the (d,t) and ( He,e*)-reactions 17. D.

:{ • member fit ilSfi ki'V should lic identical with the Ackmnvlcdtfcmcnt

:j levelal !i:5l keVin rei. |ir,|. .,.,„, .l)Jth,„., • ,.s)>1.t.ss ,|«.i,- ..vatitu.le to IVotcssovs

Then. the Sol ke\ stale can be tcntatncly a11r,buted (; ,, ,,1(1;MA N N A.'M'AITAS and .1. KVMKKK.: tor

to the 2 „.emher..: the/i-vibralmiml hand. tl|rj|. .|Mi|.|vst .|n |l)j;J wijr.,. W(. w|s)| ,o tl|.lnk ^

The |I«W k«-\- level „I.oul,! l;e i>lr>,li<-.il with th<- (»mil R_ ,,,,,,,.ANX ,„,. ,li(i ;i,.,Mill,ee .\urina Hir exferiments

l,<,,,l ot th,. -vibr.itK.n which has b.-e,, 1m,n,l at aM(1 t,|t. s(;i|,. (>|. (|i(, M;iinz T|{|(;A r<,at;|i. )()|. t||r

I0-.7 keV by VAN I.KK BAAS rf«/.|l5|. im,,li«1i,,,,s. We also «ish t.> thank Mrs. K. .I.,MAK

In siiiMih,,ry. ll.c ,.iK-«.-iit rcs,,ll« sh»»- s»in<- i»terestmK (.()l. f|u. ,,„„„„ „,-,,„. :it,nvmf,s Kiimirri«! s.i,.,N>rl eharaetenstM-s ol nuele, ,n the transitional n-.on ,i.om t(u. iSwi.(1|sll Atomi(. R^,,,,^ ,.(Uin,.i|. tll(. between sliliei'iial and . both nuclei for the second excited 0 states, which 44'!'M',': .. ,. lmvc bceti iound |IH. l!l| in the Sin isoto,,es as sha|>e ,;. H KKK.MNN ;,„., K. K. SI:VI:. 7.. Pliysik A 17>. L».H coexisting states. (1SI7.TI. PAPER P.III. J inorg. aud, Chem.. 1977, Vol. 39. pp. IIO7-]|||. Pergamon Press. Printed in Great Britain

DECAY PROPERTIES OF ""La

T. BJORNSTAD and E. KVÅLE Department of Nuclear Chemistry. University of Oslo. Oslo 3, Norway

G. SKARNEMARK and P. O. ARONSSON Department of Nuclear Chemistry, Chalmers University of Technology. Fack, S-'O2 2'J Goteborg 5. Sweden

and

N. KAFFRELL, N. TRAUTMANN and E. STENDER Institut fur Kernchemie, Johannes Gutenberg Universitat. Postfach 3980, D-65 Mainz. Germany

(Received 9 September 1976; accepted 17 November 1976)

Abstract—^singles and y-y coincidence measurements have been performed on samples of '"La produced in thermal neutron induced fission of "'U. The nuclide was isolated by means of the on-line chemical separation syslem SISAK in combination with a gas jet recoil transportation (GJRT) system. From the decay of the strongest T-lines, the half-life of '"La has been determined to be 14.23 ± 0.14 min. A number of new -y-rays are reported, and a decay scheme involving many new levels has been derived.

INTRODUCTION EXPERIMENTAL Irradiations and chemical separations. The ""La was produced The neutron-rich muclide '"'La has been surprisingly in thermal neutron induced fission of Z"U (450 fig) at the Mainz little studied despite its relatively long half-life of 14 min. TRIG A reactor (flux ~10" n/cmzxs). The target arrangement The half-life is sufficiently long to allow the isolation and was identical to the one described in Ref. [8J. the purification of the nuclide from a fission product In thermal neutron induced fission of "'U, the proton-number mixture by traditional techniques such as precipitation, nearest to the most probably charge {Zp) of the mass number 143 ion exchange or off-line solvent extraction. is 56, corresponding to the element Ba. The cumulative chain Gest and Edwards [1] showed already in 1951 that the yield of the 13.6s ""Ba is 5.50%. while ihe independent yield of lanthanum precursor of l43Ce had a half-life of about ""La is 0.44% and the total chain yield 5.95% [9]. Accordingly, it 19 min. However, th? first serious attempt to investigate is advantageous to remove the primarily formed lanthanides and milk the ""La from "'Ba The chemical separation system for the decay of ""La was carried out about ten years later isolation of '"La is shown in Fig. I. The jet gas carrying the by Fritze et al.[2]. They published half-life, 95%) are removed, the published. aqueous solution is fed into the first mixer-centrifugal unit Cl. Therefore we decided to perform experiments in- Here it is contacted with an organic phase, org 1, consisting of volving y-ray singles and •y-y coincidence measurements 0.3 M HDEHP (bis-2-ethylhexylorthophosphoric acid) in on the decay of ""La, utilizing the fast on-line chemical kerosene, and the lanthanides are extracted into org I. leaving Ba in the aqueous phase, aq 1. A delay, Dll, of about 15s in aq 1 separation system SISAK [6,7] in combination with a gas allows 50-60% of the present ""Ba to decay into '"La while the jet recoil transportation (GJRT) system [8]. These ex- growth of "!La from the 10 min M2Ba is kept at a reasonable periments are part of a larger decay study project on the level. The grown-in lanthanides (mainly '""'"•""La) are extruded neutron-rich isotopes of La, Ce and Pr in the transition into org 2 (in C2) which has the same chemical composition as region between spherical and deformed nuclear shapes, org 1. In a delay. DI2, of 200s in org 2 more than 96% of the but this paper will exclusively deal with M3La. 42.1s 1<4La decay to the longlived '"Ce. In C3. the '"La is

Org.l Otg 2

O ObM K2Cr Fig. I. Chemical separation system for the isolation of '"La. Cl, C2, C3 = mixer-ccntrifugal separation units; Dll, D12 = delay lines; GJ = gasjeUN; + C:H«):M = static mixerfor gas and liquid; DG = degassinguniuNG = noble gases; FP = fis5ion products; C = counting cell (DOWEX 50Wx4.50-100 meshl; D = detectors.

/INC VOL » NO. 7-A 1107 nos T. BJORNSTAD el al.

» I o

a 2io

il." LO

e r 0 ri o L-J HK CHANNEL NUMBER Fig. 2. (al y-Ray spectrum of ""La covering the energy range 15-30011 keV. The speclrum is accumulated an the ion exchange resin in the counting cell, and energies are given in keV.

.,14

I o -,200 400 600 800 cc a. 21 in z D O o

,10* 1200 1400 1600 1800

2200 2400 2600 2800 CHANNEL NUMBER Fig. 2. (b)T-Ray spectrum of 14!La covering the energy range 15-3000keV. The speclrum is accumulated on the ion exchange resin in the counting cell, and energies are given in keV. Decay properties of ld'La 1109

backextracted to an oxidizing aqueous solution of 1 M HNO,, Table I. Half-life of '"La as determined from

0.1 M HiSO., and 0.05 M K:Cr,O,. leaving CeUVI in org 2. the double peak of the two most intense y-ray Finally the "'La is adsorbed on a cation exchange column energies 620.6 and 621.7 keV (I)owex 50W x 4. 50-10(1 mesh) and the y-ray counting is carried out directly on this column. Run Half-life Half-life (min) The specifications ofthe chemicals used are the same as given no. (min) mean value in Ref. |IO|. •/•Singles and y-y coincidence measurements. The polypro- 1 14.30 ±0.25 pylene counting cell containing the ion exchanger resin is -» 14.14 ±0.33 identical with the flow cell described in Ref. [10]. The cell allows 3 13.86 ±0.34 14.2310.14 detection of low energy y-transitions down to 18-20 keV. 4 14.31 ±0.32 Since the production-separation system used in the present 5 14.43 ±0.32 work is a real on-line system, the build-up of ";La on the cation exchanger will, to an increasing extent, affect the measurements of ""La. To reduce this contamination, the resin was frequently Criterion 1 or 2 + 3 have each been accepted as renewed during the y-y coincidence measurements. sufficient. If only one of the criteria 2 or 3 is fulfilled, the The half-life determination was carried out hy the traditional technique of sampling (on the ion exchanger) and subsequent assignment is still in question. decay measurements. The measuring periods were designed so as The energy calibration curve was composed of three to facilitate correction for possible long-lived components in the parts: 0-300 keV with the standard nuclide '"Tb, 300- 7 y-ray peaks. 2000 keV with standards "Co. "Y and " Cs and above The Ge(Li) -^ray detectors used had an energy resolution at 2000 keV where "'Co was used as a standard. The final 1332 keV of 2.3 keV and 1.75 keV and relative efficiency of 23.6% curve is constructed by a linear fit within each of the and 6.4%. respectively. In addition, to improve the energy energy regions. resolution in the low energy region an X-ray detector was used In some cases the energies given are less accurate. The (resolution at 122 keV - 600 eV). peak at 475.5 keV is the weak component in a doublet The electronic equipment was identical with that described in with the stronger 476.7 keV and both the energy and Ref. 1101. intensity are estimated from the coincidence spectra. For l4 RESULTS AND DISCUSSION the 2056 keV peak the probable 'La-component is y-Ray spectra completely covered by the stronger 14!La-component and A typical y-ray singles spectrum is shown in Fig. 2b. is found just as a shorter lived part of the decay curve of Figure 2(a) shows the low energy part of the spectrum 2056 keV. The 1299keV peak is not clearly shown in the (recorded with the X-ray detector). The r-ray peaks singles spectra and is found only through the coin- belonging to l41La and the most intense peaks from the cidences. Generally, the energy values given in the contaminants ":La, "'"'"Te and l4*Pr are labeled with regions 0-300 keV, 300-2000 keV and above 2048 keV are energies (in keV). the mean of 3, 5 and 2 single values respectively.

Half-life Decay scheme Two of (he must intense •y-ray peaks at 620.6 and The decay scheme derived is shown in Fig. 3 and for 621.7 keV were used for the half-life determination. comparison the level scheme from Refs. |ll] and [12] is These peaks form a doublet, but a computer fit shows parallelly drawn. The Qi-value of 3.3±0.1 MeV is the identical half-lives for the two peaks. In order to obtain one measured by Fritze el al.{2]. It is in agreement with better counling statistics and to avoid the extra uncer- the value calculated by Viola et a/.[l3] of 3.383 MeV. tainty introduced in the fitting procedure, the total net Some general comments on the decay scheme are area under the doublets was used. Five runs were necessary. The scheme is drawn with the symbols used performed. The half-life was calculated by a least in Nuclear Data Sheets. Many of the coincidence peaks squares fit to the decay data, and the results are are rather weak and in many cases it has been difficult to summarized in Table 1. decide whether the coincidences are random or true The final value assigned, 14.23±0.14min. is the cascade coincidences. In the cases where cross checking weighted mean of the single values given. It agrees well have revealed the same result, the coincidences have with the literature values of 14.0 ±0.1 min [2) determined been accepted as true. by energy gated ^-measurements on chemically purified The lowest lying level in the ground state multiplet of samples and 14.32 ±0.73 min [4] obtained from computer- ""Ce is measured by means of the atomic beam method fit of decay curves from ^-measurements on mass to be 3/2114] in consistence with angular correlation data separated samples. and decay schemes of this nuclide 115], ""Ce has 58 protons and 85 neutrons in the nucleus. According to the y-Ray energies shell-model theory each of the three neutrons outside the The y-ray energies attributed to l4'La are shown in closed shell 82 is in the /7/2-state, and the coupling Table 2 together with relative intensities and coin- between three identical particles in the same /-state cidences. usually gives / = /-l|l6]. This should yield I = $12 When assigning the y-ray energies to '"La, we have which is not observed. The measured anomalous spin of used three criteria. The y-ray should: 3/2 might be explained if long-range Majorana forces (1) have coincidence(s) with •y-ray(s) unquestionably were in operation (14). belonging to '"La (mainly those with /, > 20%), In a polarization experiment, Graw el a/.|l7) have (2) show the right half-life (14 ±2 min). found a low-lying excited level of spin and parity 7/2-. (3> fit into the level scheme constructed from the According to I.essard et al. 111 ], who studied the levels in id. r>)-reiiciion diila by Lessard et «/.|ll|. the thermal l4!Ce by means of the (d. p)-reaction. this level has neutron caplure-y data of Groshev el tj/.[12] and the y-y probably an excitation energy of 20keV and it is coincidences from the present work. assumed to be identical with the 18.9 keV level derived in lill) T. BJORNSTAD et al.

Table 2. y-Ray energies assigned to N'l.a with relative intensities and observed coincidences

hnergi Rel. ml " Literature Values [2] Ey (keVl h \%) Observed coincidences*: Ey (MeV) /y (%)

23.4 ±0.3 620.6(s).774.9(w|. 1053.2(m). !122.9iw). 1556.6(ml 0.200 + 0.01 0.08 1938.0IW) 0.440 + 0.02 0.13 42.3+0.3 620.6M. 774.9(ml. 1053.2tw|. 1556.61m) 0.625+0.01 1.0 19.38.IKw) 0.800 ±0.01 0.44 300." ±0.5 1.7 Not puled 9.9|5±0.02 0.08 433.1 ': 0.5 4.1 Not gated 1.07 ±0.04 0.26 545.2 ±0.5 560.f(s).6:i l878.4(m).20IM.l(w) 643.9± (1.3 71.6 433.l(w). 454.2(s).560.1(ml. I592.5(wl. 174«.9(u/ 774.9 + 0.3 15.1 23.4(w).42.3(w).300.2(w). 860.31 w). I24l).2(wi 798.3 ±0.3 47.8 300.21ml. 860.3(m). 1240.21 m) 860.3 + 0.3 S.I 798.31 w) 919.3±0.5 9.6 I556.6(m) 942.7 ±0.5 3.0 Not gated IO53.2±O.3 25.6 23.41s). 42.3(nu.475.5(m).581.9{.si. 1211.81«) 1064.2 + 0.3 2.8 I423.6IW). 1475.5lwl 1076.5 ±0.3 15.2 581.9(5). 1475.5(w) 1086.8 ±0.3 2.6 Not gated 1122.9 ±0.3 13.6 23.4(w). 42.3(w) II 39.6 ±0.3 7.9 1148.61m) II46.O±O.3 31.0 None 1148.6 ±0.3 39.9 H39.6(m). 1475.5(w) IIM.9 + 0.? 14.4 None 1167.7 ±0.5 4.2 None 1201.1+0.3 9.7 I299(wl 12ll.8±0.3 4.1 Not gated 1240.2 + 0.3 5.3 Nol gated 1299.0+1.0 Nol gated 1402.6 ± 0.5 2.3 454.2(s).476.7(w), 620.6(ml. 643.91 w 1 1423.6 + 0.5 2.3 Nol gated 147s\s ± 0.3 7.9 1053.2(w). 1076.5(w). 1148.6(w) 1556.6 ±0.3 42.3 23.4(m).42.3(m).9l9.3(w) 1592.5 ±0.3 4.5 Nol gated 1611.7 ±0.3 3.5 454.2(m).47fi.2(w).ft2a6(ra) 1707.9 + 0.3 16.4 None 1740.9 ±0.3 4.9 620.6(w).643.9(wi. 1838.1+0.3 8.4 None 1876.1+0.3 9.6 621.7(m) 1878.4 + 0.3 8.7 621.7|m) 1938.0 ±0.3 16.0 Not gated 1961.5 ±0.3 40.1 None 1980.4 ±0.3 14.2 None 2004.1 ±0.3 10.2 Not gated 2056+1 2.7 None 2066.3 + 0.5 2.6 Not gated 2385.2 + 0.5 7.2 None 2500.0 + 0.5 29.3 None 2625.0 ±0.5 12.8 None 2710.2 ±0.5 3.5 None 2825.6 + 0.5 5.1 None 808.1 ±0.38 1346.8 ±0.3« 1664.2 ±0.3§

•The error limits are estimated to be +I(W for the stronger lines I/, > 20%) and +20^ for the weaker ones. is. strong: m. middle strong; w. weak. §y-Energies showing the right half-life, bul can not be lilted into the decay scheme Decay properties of 1J'La

•• ::::*is :t

"JCe. , Fig. 1. Decay scheme of '"l.a the present work. The transition from this 18.9 keV level 2. K. Fritze. T. J. Kennett and W. V. Prestwich. Can. J. Phys. to the ground state should thus be an E2 transition, 39. 662(1% II. which is almost completely converted. This explains the 3. A. Ohyoshi. T. Tamai and M. Shinagawa. Bull. Chem. Sue. absence of an 18.9keV line in the measured y-ray (Japanl44. 34X9 119711. spectra. 4. B. F.hrenberg and S. Amiel. Phys. Rer. C6. 618 (1972). 5. K Buchtela. Aiomkernenergie IATKEI. 22. 268 11974). The next level found at 40keV in the id.p) 6. P. 0. Aronsson. B. R. Johansson. J. Rydberg. G. Skar- reaction 111 ] may correspond to the state observed by us nemark. J. Alstad. B. Bergersen. F.. KvSIc and M. Skarestad. at 42.3 keV; it is depopulated by rather strong y-ray ;. Inorg. Nucl. Chem. 36. 2397 (1974). transitions of 42.3 and 23.4 keV to the ground state and 7. P.O. Aronsson. Thesis. Chalmers University of Technology. the 7/2 level at !8.9keV, respectively. As the (j- 1) Goteborg 1974. state is also expected at a low energy we would propose 8. N. Trautmaiin. P. O. Aronsson. T. Bjornslad. N. Kaffrell. E. J" =5/2 for the 42.3 keV level. Concerning the other Kvåle. M. Skareslad, Cj. Skarnemark and E. Stender, lnorg. levels, although some of them are close in energy to Nucl. Chem. Lett. 11.729(1975). those found in the (d.p) reaction, no definite spin and 9. K. Wolfsheri!. Los Alamos Scientific laboratory. Report parity assignments can be deduced. I.A-5553-MS. IV74. Ml. E. Stender. N. Irautmann. V. (). Aronsson. [. Bjornstad. N. Acknowledgements—The authors express their gratitude to Kaffrell. E. Kvåle and G. Skarnemark, tn be published. Profs. G. Herrmann. A. C. Pappas and J. Rydberg for their 11. I.. I.essard. S. Gales and J. I.. Foster. Jr.. P/i.r.i. Rev. C6. 517 interest in this work. We wish to thank Mr. R. Heimann for his (19721. assistance during the experiments and the staff of the Mainz 12. I.. V. Gmshcv. V. N. Ovorctskii. A. Vt. Dcmidov and M. S. TR1GA reactor for the irridations. Financial support from the Al'vash. Sen: J. Slid. Pliys. 10. .192 1197(11. Swedish Atomic Research Council, the Bundesministerium fur 13. V. E. Viola. Jr.. J. A. Swam and J. Graber. \tomic Dtila and Forschung und Technologie and the Norwegian Research Nuclear Data Tables 13. 35 (19741. Council for Science and the Humanities is gratefully acknow- 14. I. Maleh. Phys. Rn. 138. B7<* (1965). ledged. 15. P. R.Gregory. L.Schellenberg.Z, Sujkowskiand M. W.Johns. Can. J. Pins. 46. 2797 il%8). REFERENCES 16. M. G. Mayer and H. Hans I). Jensen, l-lcincnlary Thet>>> of 1. H. Gest and R. R. Edwards. National Nuclear Hnergy Series. Nuclear Shell Structure. Wiles. Ne» Yolk 119551. Plutonium Project Record, IV. 9. p. 1144. McGraw-Hill. 17. G. Grau. G, (ilaiisnii/cr. R Fleisi/hmaim and K. Wienhard. New York 119511. Phys. Lett. 2«B. 5K3 (l%9i. PAPER P.IV. J inarg. nucl. Chem . 1977. Vol. IV. pp \W-Wi Pcrgamon Press. Primed in Greai Britain

DECAY PROPERTIES OF l44I46La

G. SKARNEMARK and P. O. ARONSSON Deparlmenl of Nuclear Chemistry, Chalmers University of Technology. Fack. S-402 20 Göteborg 5. Sweden

T. BJÖRNSTAD and E. KVÅLE Department of Nuclear Chemistry. University of Oslo. Oslo 3. Norway

and

N. KAFFRELL. E. STENDER and N. TRAUTMANN Institut fur Kernchemie, Johannes Gutenberg-Universität. Postfach 3980. D-(o00 Mainz. Germany

[Rti-eived 1? November 1976)

Abstract— y-ray singles and coincidence measurements have been performed on the neutron-rich nuclides '44La. '"'La and ""La, produced in thermal neutron-induced fission of 5"U. The studies were facilitated by the combination of a gas jet recoil transportation (GJRT) system with the rapid on-line chemical separation system SISAK. Half-lives of 42.1 i 0.7 s, 25.2 ±2.6 s and 8.5 ± 1.0 s were obtained for '"La, "'La and "*La. respectively. Several new y-ray energies have been assigned to each of the nuclides, and partial decay schemes are proposed for l44La and l4hLa together with a preliminary partial decay scheme for l4)La.

INTRODUCTION In two papers (11.121, Aronsson et al. reported on the Until recently, the neutron-rich transitional nuclei in the half-life, y-iay energies and relative y-ray intensities of light lanthanide element region, such as La. Ce and Pr. 144La and proposed a preliminary decay scheme. The same have been only sparingly studied, mainly due to the lack papers also presented the first direct half-life measure- of proper fast separation procedures. However, during ments and y-ray energy data for '"'La and 148La. as well as the last years a number of new separation techniques have a simple decay scheme for ""'La. The La isotopes were been invented and existing methods have been further isolated from fission products by means of the rapid refined. This experimental progress has made possible a on-line chemical separation technique SISAK[13.14]. In a new approach for nuclear spectroscopy investigations, recent publication [15], we have described the successful e.g. in a classically difficult element separation area like combination of a gas jet recoil transportation (GJRT) the rare earths. technique with the SISAK separation system. In the This paper will deal exclusively with neutron-rich present paper, we will report on y-y coincidence nv isotopes of the lightest of the rare earth elements. La. measurements on '"La. La and '"La performed with although we have also performed studies on neutron-rich this new arrangement. isotopes of Ce and Pr[l,2]. The nuclide ""La was first EXPERIMENTAL identified by Amarel el a/. (3). who isolated the nuclide by on-line mass-separation and determined the half-life to Irradiations and chemical separations 41 ± 3 s by /3-ray counting. In two papers, Ohyoshi el The La isotopes were produced as fission products by irradiation of :"U (450 jig) with thermal neutrons in the Mainz al. [4,5] reported on more comprehensive investigations TRJGA reactor. The thermal neutron flux amounted to about of ""La. La was separated from fission products by means 10" n cm ! s '. We have used the GJRT system described in Ref. of an electromigration technique. The -y-ray measure- 115|. Figure I schematically shows the chemical separation system ments were started about 4 min after the end of the used for the isolation of La isotopes. The jet gas carrying the irradiation. From the decay measurements of the two fission products is extensively mixed with a nitric acid solution ot most prominent y-peaks at 397 and 541 keV they pH 1.4 I-90T) in a static mixer. The gas-liquid mixture is then determined the half-Mfe to be 42.4 ± 0.6 s. On the basis of fed into a degassing unit where the jet gas is swept off together this half-life they also assigned seven other •y-ray energies with more than 95% of the noble gas activity. In the first to 144La. mixer-centrifugal separation unit. Cl. all trivalent lanthanide* (mainly La. Ce and Pr) are extracted into an organic phase Seyblft] utilized a fast off-line separation technique for consisting of (UMHDEHP (bis-2-elhylhcxylorthophosphoric light rare earths and reported on the half-life and some acid) in kerosene (Shcllsol-T) together with some other fission y-ray energies for "*La. while Wiinsch el at. [7] briefly products like Zr. Nb. Mo. etc. After the phase separation. La (and reported on a few y-ray energies for '"La obtained from Pr) is back-extracfcd (in C2) inlo an oxidizing aqueous phase mass-separated fission products. For "sLa and l4*La, less consisting of I M HNO,. 0.1 M H.SO, and 0.05 M K.Cr.O,. In this conclusive data are found in the literature. Grapengiesser step, cerium as Cc(IV) and the other elements remain in the el a/, [8] detected a component in the mass-chain 145. organic phase. The y-ray measurements arc then carried out an decaying with a half-life of about 36 s. The clement the aqueous phase leaving C2. The organic phase runs in a closed circuit. This requires a cleaning step (C3) in order to prevent suggested was La. in accordance with the results of accumulation of longer-lived sp>\ k « Wilhelmy(9). Seyb[6]. has indirectly measured the half- Chemicals. Most of the chcinuals were of pa. grade. The lives of '"La and '*La to be 28±3s and 8.3s, HDEHP wa< supplied by Farhenfahriken Bayer AG. Loverkuscn. respectively, while Fasching[IO] found 24±5s and Germany, and used withoul purification. Shcllsol-T (kerosene) 15 ± 10 s. also from indirect measurements. was used as organic diluent.

1487 G. .SKARNEMARK el al.

C2H4 . N2 Noble gases

Gas jet from target

H2O 0 2511/1 HNO3 1 M HNO3 1 M HN0-, 0.1 M H2SO4 0 05 M NH2SO3H 0.05M K2Cr207 0.05 M H2O2 Fig. I- Fl"w '•heel showing the chemical system used for Ihe isolation of [.a isotopes. M = mixer. Dg = degassing unit. CI-C3 = mixer-centrifugal separator units. 1 = detector cell. D = detectors.

Singles y-rax and y-y tnincidence measurements of Ihe i.oincidence setup was 2r = 20 ns. The coincidence events The counting flow cell was a cylinder of polypropylene were stored in a 4K x 4K matrix and subsequently transferred, Idiamcter 3.5 cm. thickness 1.0 cm) with diametrically opposite event by event, onto magnetic tape. The spectrum analysis was inlet and outlet. The walls facing the detectors were 0.1 cm thick in carried out by appropriate computer programs. order to allow the detection of low energy transitions (a20 keV). Half-life measurements. The half-life measurements of the The y-ray measurements (singles and coincidence! were carried three nuclides have been carried out by the traditional technique out by using two Ge(Li)-detcctors with relative efficiencies of 23.6 of sampling and subsequent decay measurements. For l44La. the and d.47r. and energy resolutions al 1332 keV of 2.3 keV and measurement time sequence was repeated 40 times, while the 1.75 ke\". respectively. In addition, slandard electronics wilh the sequence fur *'l.a and ''"'La was repeated 150 times. possibility of spectrum stabilization was used. The resolving time The half-lives were calculated from the decav of the mosi

z u

V) z O u

CHANNEL NUMBER Fig. 2. y-ray spectrum of neutron-rich l.a isotopes. The spectrum «as record«! on-line during 30mm Decay properties of l44l4ALa 1489 prominent y-peaks with a least squares fit to the decay data and the Cheifetz el a/.[16] determined the energy of the ground value finally assigned is the weighted mean of the values obtained state 2'-»0' transition in '"Ce to be 397.5 keV. The for the individual peaks.

RESULTS AND DISCUSSION A typical y-ray spectrum of the La fraction is shown in Fig. 2. The -y-ray peaks belonging to La isotopes are )O 2 ke» T, 2 261 13 5s indicated together with some grown-in Ce and Pr peaks. 170 2 keV . T, ? 21 3 * 3 9s The results obtained for the different nuclides studied will be discussed below, nuclide by nuclide.

""La Half-life. The calculated mean value of 42.1 ±0.7 s from the decay of the y-ray peaks at 397.5. 541.3, 585.1 and 844.9 keV is in good agreement both with the /3-ray measurement data of Amarel el a/. [3], 41 ±3s, and the value obtained by Ohyoshi el al.[5]. 42.4±0.6s. Decay scheme. The y-rays assigned to the decay of '•"La are listed in Table 1 together with the observed coincidences. The decay scheme derived from these coincidences is shown in Fig. 3. From three and four parameter coincidence measurements (including y-rays. X-rays and fragment masses) on products from the spontaneous fission of!* !Cf. Fig. 4. Decay curves of some u

Table I. Energies, relative intensities and coincidences observed for y-rays assigned lo '

Energy Rel. int. Observed coincidences Er (keV)

367.3 3 None 397.5 100 432.1(s), 541.3! 585. I (s) , 705.0(5). 735.Ms), 844.gis), 952.71 969.0(5), 1276.7(5), 1294.4(5), 1432.3(5)

432.1 397.5(5), 844.9i :)69.0!s)

5*11.3 42 397.5(b) 585.11=1 /35.4(s), 952.7(s), 969.ON) 585.1 12 397.5ls) 541 .3 .' s j, ?52.7'r-)

597-9 3 None 705.0 5 397.5(5)

735.4 11 397.5(5!, 541.3(5), 960.0(5)

752.1 3 Not qated 84 4. q 28 397- 432.!<<) 952.7 397. 541.3(5), 585.1(ml

969.0 397. 432.1U|, 541.3(s). 736.Ms). BMt.9(m) 1102.4 None 1276.7 None 1294.li 397-5(5) H32.3 397.51i) 1489.9 None 1674.2 Not qoled 1820.3 None 191.3.3 Nol Qatcd 2009.0 Not qatcd 2325.6 Not qaled 2867.1 Not qateri

The uncertainly in the ^f-ray intensities is estimated 10 bo ± I r* for the vlronq peaks

(above 10 relative intensity) and 120' for the weak ones.

The uncertainty in the Y"<*ay energies is estimated to he 10.5 keV be 1 ov; 200P 1-cV 'ind ±1 keV above 2000 keV. s(«=tronq) indicates that the peak is hiqhf ikar 3 'standard

IIS( \il "< ^! •'. ti G. SKARNEMARK el al.

present energy data agree well with this result as well as the half-life determination. The decay curves are shown in with the findings of Ohyoshi et al.15]. Seyb[6) and Fig. 4. It should be mentioned that these y-rays have been Wiinsch el u/.[7]. except for the 260.9 keV peak[7] which attributed to ""La only on the basis of their half-life is not seen in the present work. (which is in good agreement with the half-life obtained The present decay scheme is an extension of the from the growth-and-decay curves of the l4'Ce peaks in scheme previously proposed by us[12]. Jn that paper we the La fraction), and that one cannot completely exclude compared the experimentally obtained levels with the the possibility that they belong to another lanthanide predictions of the variable-moment-of-inertia (VMI) nuclide with the same half-life. The final half-life model (17]. concluding that the 938 keV level and the 1523 keV level probably were the 4' and 6* members of the ground state rotational band. However, in the present Counts work we have found a 1523 keV transition. If this 2585keV.T ?«8OiO5s transition de-excites the 1523 keV level, it excludes the 4IOOkeV, T,,_.=90i05s spin assignment 6' to this level. There is, however, a weak coincidence between this y-ray and the 397 and 844 keV y-rays, thus implying that it de-excites a 2766 keV level instead of the 1523 keV level. This is also in agreement with the results of Monnand and Fogelberg|l9]. Thus the 1523 keV level might still be the 6' level. In order to make definite conclusions about the spin of this level, y-y angular correlation measurements are a subject of future interest.

Since the C?B-value has been measured to be 4.4MeV[l8], /^-feeding of higherlying levels than the 2641 keV level is energetically possible. Inspection of the y-spectra indicates peak energies even above 3000 keV, but with too low intensities to be uniquely assigned. Besides the more conclusive energy data in Table I, Table 4 indicates several y-ray energies which may be attributed to the 0 -decay of '"La.

14

Table 2. Energies, relalive inlensilies and coincidences observed for y-rays assigned lo ""I.a

Encrqy Rei. int. Observed coincidences t UeVI I (•') r

'•8 2 40 Not gated 70.2 67 165.3(5). 799.8(m), 840.9(m), 830.2(s), 932.3(5)

1 U . 1 <5 Not qaterl 1 IS .1, 100 215.3(5), 403.7(5). 799.8(s), 840.9(s) if.:• 3 10 70.2(s). 215.3(w), 403.7(5) 170 2 39 254.0(nO 189.0 54 None 215 14 I89.0(w), 254.0(m) 2 54 .0 7 None uc?.7 ?[) 7O.2(s). 117.1(w;, 118.4(s), 165.3 79«.8 35 7O.2(m), 118.1.11} 840. 9 <5

P90. : 19 70.2(5) 918. 2 5« None 932. 3 11 70.2(5). 118.4(0, 1&5.3(w) "59. 2 <5 Wot rjated i n 53.? 16 70. 2(r.) 1167. b <5 Not gated

The uncertainty in the Y-ray intensities is estimated to be *?C . b) The uncertainty in the Y-ray enerqies is estimated to be ±0.

I . - tl.rM1,.'it • . t,t. '-11.U- 1 . Decay properties of 144"l-WlLa 1491

assigned. 25.3 ± 2.6 s is in agreement both with the values 0-decay of '"La mainly feeds the ground state of ""Ce, of Seyb(6] (28±3s) and Faschingfll] (24 +5s). and the weak y-ray transitions make some of the Decay scheme, y-ray energies (with relative intensites) assignments uncertain. Besides the y-ray energies assig- and coincidences are listed in Table 2. From these data we ned to '"La in Table 2, Table 4 gives a number of have derived a decay scheme, which should be regarded candidates for the same decay proposed on the basis of as a preliminary version (Fig. 5). It seems as if the the measured half-lives.

36 MeV \ 181 \

r; ^ in ro .4 3 S s( 2643 2 1 1167 8 11J3 4 24 76 6

959 2 918 2

1829 5 1691 9 1674 2

1242 4

1102 4

235 5 1 i i 170 2 i 118 4 é 70 2

0 0

Fig. 3. Partial decay scheme of ""La. The scheme is based on the Fig. 5. Partial decay scheme of '"La. The scheme is based on coincidences listed in Table I. the coincidences listed in Table 2.

Table 3. Energies, relative intensities and coincidences observed for y-rays assigned to l4*La

Energy Rel. int. Observed coincidcnce5 E, (keV) I ()

183.3 258.5(s), 292.Mw>, 410.0(s) 258.5 183.3(5), 292.Mm), 380.l(s), M0.0(s), 503-2(s) 515.0(5), 666.Ms), 702.Ms), 785.M5), 11M.9(S) 292 .J* 1 Not gated 380.1 5 258.5(5), MO. 0(5), 5O3.2(s) 1)10.0 63 183.3(5), 258. 5(s), 292.Mm), 503.2(5), 515.0(5) I.II7.2 21 None 503.2 21 258.5(s), 380. 1(5), MO.O(s) 515.0 22 258.5(5), Mo.0(5) 666.1* 8 183.3!m), 258. 5(s). 702. Mm) 7O2.f| U 258.5(s), bib. Mm) 785."i 1, 258.5(5)

101*3 .A 3 None ?|i)1.9 9 None 2359.« 13 Not qated 1492 G. SKARNEMARK cl al.

Table 4. Unassigned -y-rays observed in the La fraclion

Energy (keV) Half- life- MJ)

kgk.k 25 i 11 ings. 3 60

523-1 i 2'- 645.1 25 i 5 671.8 io t 10

685.8 26 i q

7 '< 3 .6 36 3 5 761, .ft 2 li i 7

853-1 1 3 i 5 880.2 26 i 7

1216.7 n j 6 1238.8 3/ i It 1922.9 21- i 6 ?156.2 1(. i 7

2207-3 lifi ! !5 2352.A t 1 \

a) The h..lf-liv« listed WJ.

by "convent iona! " dec.iy ineasurc-mtMM* '*; . c-. the TOO techn i tjue h.ib nol been CPIO loyed ) . The unc t-r t.i i n t y in t hc T" r ny em.M U i e i'- ».-s ti mated to be ±0.^. ke\/ be'ov- ?0';'j KCV ,i"ri ±' kcV dbove 2000 keV. It shou Id be not i c cd i h^ i .i I i t hest' > i i'vs <»M.- very

w*lk "T(re])<5 inco.n„ri,o •„„• ,.... 397.5^' .in, , , „, Lu) tind (hat" botne "f then '".._ '.flnn^ to heavier Icnnth.iri^ (Nd and Prv) .

'*La Half-life. The two most prominent y-rays in the ""'La 0-decay, 258.5 and 410.0 keV. have been used for the half-life determination. The decay curves are shown in Fig. 6. The half-life value finally assigned. 8.5 ± 1.0 s. is somewhat lower than the value of 11 ± 1 s earlier found hy means of the two-detector-delay (TDD) method|l21. It is consistent with the results of Seyb[6] (8.3 s) and Faschingll Ij (15 i 10s). It should also be mentioned that some of the y-rays assigned to '*La show decay curves which indicate the presence of a weak, longer-lived component. These y-rays, e.g. the transitions at 183.2 and component. These y-rays. e.g. the transitions at 183.3 and 503.2 keV. are. however, relatively weak, and thus the Decay scheme. The y-ray energies and coincidences 1183 5 1171 7 obtained are listed in Table 3. and the decay scheme 1144 2 derived from these data is shown in Fig. 7. Cheifelz el 1043 1 a/.[l6] assigned energies of 502.3. 410.1 and 258.6keV to 960 9 the transitions belonging to the 6' -»4* -»2" -»0' cascade in '*Ce. It is reasonable to believe that these energies correspond to three of the most prominent transitions found in the present work at 503.2. 410.0 and 258.5 keV. We adopt the spin assignments 2' and 4' for the levels at 258.5 and 668.5 keV, respectively, but we consider the assignment 6' to the 1171.7 keV level as less conclusive, due to the relatively strong ^-feeding of this state as well as of the 2* and 4' slates. Thus, if the 1171.7 keV level is a 6' state, it can be fed only from the decay of an isomeric state in ""La. As previously mentioned, the decay of some of the y-rays of '"La indicate the presence of a longer-lived component. However, this evidence is too Fig. 7. Pariial decay scheme of l4ftLa. The scheme is based on the weak to allow any definite suggestion of such an isomeric coincidences listed in Table 3. state. Decay properties of '""'"La 1493

Acknowledgements—The authors are indebted to Professors G. Beta-Stability. Leysin 1970, CERN-Report 70-30, p. 1093 Herrmann, A. C. Pappas and J. Rydberg for their interest in our (1970). work. We are also indebted to Mr. R. Heimann for kind assistance 9. J. B. Wilhelmy, University of California Radiation Laboratory during the experimenis and to the staff of the Mainz TRIGA Report No. UCRL-18978 (1969). reactor. Mrs. E. Jomar prepared the drawings and Mrs. M. Carlson 10. J. L. Fasching, Thesis. Massachusetts Institute of Tech- typed the manuscript. We also gratefully acknowledge the nology, Cambridge, Massachusetts (1970). financial support from the Swedish Atomic Research Council, the 11. P. O. Aronsson, G. Skarnemark and M. Skarestad, /. Inorg. Bundesministerium fur Forschung und Technologie and the Nucl. Chem. 36, 1689 (1974). Norwegian Research Council for Science and the Humanities. 12. P. O. Aronsson, G. Skarnemark, E. Kvåle and M. Skarestad. Inorg. Nucl. Chem. Lett. 10, 753 (1974). 13. P. O. Aronsson, B. E. Johansson, J. Rydberg, G. Skarnemark. J. Alstad. B. Bergersen. E. Kvåle and M. Skarestad. /. Inorg. REFERENCES Nucl. Chem. 36.2397 (1974). 1. T. Bjornstad, E. Kvåle, G. Skarnemark and P. O. Aronsson, 14. P. O. Aronsson, Thesis, Chalmers University of Technology. Submitted to J. Inorg. Nuci. Chem. Goteborg (1974). 2. G. Skarnemark, E. Stender, N. Trautmann. P. 0. Aronsson,T. 15. N. Trautmann, P. 0. Aronsson, T. Bjornstad, N. Kaffrell, E. Bjornstad, N. Kaffrell, E. Kvåle and M. Skarestad, Radio- Kvåle, M. Skarestad, G. Skarnemark and E. Slender, Inorg. chim. Ada 23. 98 (1976). Nucl. Chem. Lett. 11.729(1975). 3. I. Amarel, R. Bernas, R. Foucher, J. Jastrzebski, A. Johnsson 16. E. Cheifetz, J. B. Wilhelmy. R. C. Jared and S. G. Thompson. and J. Tiellac, Phys. Lett. 24B, 402 (1967). Phys. Rev. C4. 1913 (19711. 4. A. Ohyoshi, E. Ohyoshi, T. Tamai and M. Shinagawa, /. Inorg. 17. M. A. J. Mariscotti, G. ScharfT-Goldhaber and B. Buck. Phys. Nucl. Chem. 34. 3293(1972). Rev. 178, 1864 (1969). 5. A. Ohyoshi, E. Ohyoshi, T. Tamai, T. Takemi and M. 18. M. Devillers. M. Fiche and J. Blachot, Private com- Shinagawa, / Nucl. Sri. Tech. 9, 658 (1972). munication. March < 1976). 6. K. E. Seyb. Jahresbericht 1973 (Institut fiir Kernchemie der 19. E. Monnand and B. Fogelberg. In Proc. Ird Int. Conf. on Universitat Mainz). Nuclei Far from Stability; Cargése 1976. CERN-Report 76-13. 7. K. Wiinsch. H. Gunther, G. Siegert and H. Wollnik, /. Phys. p. 503 (1976). A: Math.. Nucl. Gen. i. L93 (1973). 20. P. A. Seeger, In Proc. Int. Conf. on the Properties of Nuclei 8. B. Grapengiesser, E. Lund and G. Rudstam, In Proc. Int. far from the Region of Beta-Stability. Leysin 1970, CERN- Conf. on the Properties of Nuclei far from the Region of Report 70-30, p. 217(1970). PAPER P.V. .14. pr 5929-1934 Pergamon Press. Printed in Greai Britai

DECAY PROPERTIES OF SOME NEUTRON-RICH CERIUM ISOTOPES

T. BJORNSTAD and E. KVÅLE Department of Nuclear Chemistry, University of Oslo, Oslo 3. Norway

and

G. SKARNEMARK and P. O. ARONSSON Department of Nuciear Chemistry. Chalmers University of Technology. Fack. S-402 20 Gotehorg 5. Sweden

and

The S1SAK Collabo.ation

(Received 22 March 1977)

Abstract—v-y-coinciiience measurements have been performed on neutron-rich Ce isotopes using the fast radiochemical separation system SISAK in combination with a gas jet recoil transportation system. The results include assignments of new y-rays and proposal of partial decay schemes for 3 min ' !Cc, 56 s 'Ce and 48 s "te. The existing decay scheme of 14 min ' Ce has been verified, except for a few transitions.

INTRODUCTION paration system. The first way was chosen by Seyb(8.9] The light neutron-rich lanthanides such as La, Ce and Pr who utilized an off-line separation technique where the are situated in the shape transition region between measurements could start ~8s after the end of the spherical nuclei around N = 82 and nuclei with sub- irradiation. His investigations included half-life deter- stantial deformation beginning around N = 90. Know- minations and assignments of some y-ray energies to ledge about the decay properties of these nuclei is there- 147Ce and 14»Ce. fore of considerable interest for the theory of nuclear The second way was selected by Aronsson et a/., who structure. developed the fast, on-line chemical separation technique So far. however, only a few authors have published S1SAK[10,11], which is based on liquid-liquid extraction nuclear data on nuclides in this region, mainly due to separations. difficulties in achieving a proper fast separation pro- To contribute necessary experimental data to this ra- cedure. ther unexplored rare earth region, we started a series of Already in 1943, Hahn and Strassmann[l] isolated a experiments on heavy La, Ce and Pr isotopes produced C'e-nuclide decaying with a half-life of about 15 min by in neutron-induced fission of U[12-14] employing the precipitation from fission products. The nuclide was SISAK separation technique. For Ce, these experiments identified as l46Ce. Later on, Mavkowilz et al. [2] isolated led to half-life determinations and the first assignments 3 min 'JS0e from a fission product mixture by employing a of y-rays to the nuclides l47Ce and ""te. as well as a liquid extraction prodecure. half-life determination of the previously unreported nucl- 15O However, no thorough research was performed in this ide Ce[13J. region until Hoffman et al. [3-6] made their investigations A short time ago, we were able to connect the SISAK about ten years ago. This group isolated Ce from fission system to a gas jet recoil transportation system[15], thus products by an extraction procedure and performed care- making reactor irradiations possible. With this ar- ful studies of '*Ce, including y-y and /3-y coincidence rangement, we obtained much stronger samples than measurements. The result was a detailed decay scheme. with the previoulsy used target system[!2,14]. This facil- They also constructed a partial decay scheme for 145Ce itated the y-y coincidence measurements on U5Ce. based on y-y coincidence data. From milking experi- l4*Ce, l47Ce and l48Ce on which we report in this paper. ments they obtained the half-lives of 147Ce and ' Ce; they were, however, not able to measure any y-rays EXPERIMENTAL belonging to these nuclides. Irradiations and chemical separations. The irradiations were By means of an electromigration technique, Ohyoshi et performed in the Main? TRIG A reactor. The neutron flux amoun- at. [7] performed a re-investigation of l45Ce, and to some ted to about 10" n cm 2 s ' und the target used (450 «ig :"U) was ex'snt also l4*Ce. They proposed a new decay scheme connected to a gas jet recoil transportation system described for M5Ce including one more level than the scheme given more in detail in Ref. [15], The same target has been used for by Hoffman et al. These results were, however, not more than 100 h without any measurable loss in efficiency. based on coincidence data but only the sum of y-ray The chemical separation system used for the isolation of Ce isotopes is shown schematically in Fig. I. The jet gas (1:1.4 energies. mixture of QH* and N;) carrying Ihe fission products is All the above mentioned experiments have been of the thoroughly mixed with a ftnw of 1 M HNO,. After a degassing discontinuous type. To improve the experiments one step removing the jet gas and the noble gases, the liquid is could either employ a fast and fully automatized off-line contacted (in CD with an organic phase consisting of technique or make use of a fast on-line chemical se- 2MHDEHP (bis-2-elhylhexylorthophosphoric acid) in kerosene

1929 1930 T. BJORNSTAD et ai

1 M HNO3 Gas jet 0.4 M H2SO4 from target 0.2 M K2Cr207 1 M HNO3 1 M HNO3 1 M HNO3 0.1 M H2SQj 0.05MNH2SO3H 0.05 M K2Cr2O? 0.05 M H2O2 Fig. 1. Flow sheet showing the chemical system used for the isolation of Ce isotopes. M = mixer. Dg= degassing unit, C1-C3 = mixer-centrifugal separator units, E = extraction column, D = detectors, FP = fission products.

(Shellsol-T). In this step, species like Zr, Y and part of the Nb and used without purification. Shellsol-T was used as or- are extracted into the organic phase, while the light lanthanides ganic diluent. The PVC beads were supplied by Kema Nord AB. remain in the aqueous phase. After this step, Ce (III) is oxidized Sundsvall. Sweden. Ionexchanged water was used throughout. to Ce(IV) by making the solution 1 M in HN03, 0.1 M in H2SO4 Measuring equipment and data evaluation. The measuring and 0.05 M in K2Cr2C>7. From this solution, Ce (and part of the equipment consisted of one 16 K and two 4K Intertechnique remaining Nb) is extracted (in C2) with 0.3MHDEHP in analyzers together with standard coincidence electronics with the kerosene. In C3, Ce is reduced ans stripped by 1 M HN05, possibility of spectrum stabilization. The resolving time of the 0.05 M H2O2 and 0.05 M NH2SO,H: then H2SO4 and K,Cr,O7 are coincidence system was 25 ns. Two Ge(Li)-detectors were used, added in the proper quantities to reoxidize Ce, which is then one with a relative efficiency of 23% and an energy resolution at absorbed on a HDEHP/PVC column serving as a source for the 1332 keV of 2.3 keV, and the other with an energy resolution of y-y coincidence measurements. Nb remains in the organic phase 1.75 keV at 1332 keV and 0.72 keV at 122 keV, and a high effi- and does not interfere. On this column, no other activities than ciency in the low energy region. To detect low energy transitions Ce and a small amount of grown-in Pr could be observed. (2:20 keV) we used fiat, thin-walled counting cells of polypro- Chemicals. Most of the chemicals used were of p.a. grade and pylene. All the coincidence data were transferred event by event manufactured by E. Merck, Darmstadt, Germany. The HDEHP to magnetic tape and evaluated by means of appropriate com- was supplied by Farbenfabriken Bayer AC Leverkusen, Germany puter programs.

•3 s

;ill i i i

S ji " SS s i o „ ~i [;/ si Un si s ;|; 11 Ir 200 400 600

A % s ••3 A S i i i i

1200 1400 1600 leoo Channel numtwr Fif. 2. y-Riy singles spectrum of neutron-rich Ce isotopes. The spectrum was recorded on-line during 30 min. Decay properties of some neutron-rich cerium isotopes 1931

RESULTS AND DISCUSSION and l48Ce. These properties are briefly discussed below, As an example of the chemical purity of the Ce nuclide by nuclide. fraction, a Ce y-ray singles spectrum is shown in Fig. 2. l45Ce. The coincidences obtained for the 3 min ""Ce The data obtained have facilitated a detailed study of are shown in Table 1. From these data, a partial decay the decay properties of the nuclides ""Ce. '""Ce. ""Ce scheme has been deduced, as shown in Fig. 3. In this

Table I. Energies, relative intensities and coincidences observed for y-rays assigned to I4*Ce

Energy Rel.ini.t Er(keV) I, 1%) Ohserved coincidences}

62.7 14.1 207.7(s). 232.01s). 284.6(s). 351. Ksl. 423.61s), 439.8(5)./24.3(s). H47.9(s) 207.7 1.6 232.0(s),284.6(s),655.9(s) 232.0 5.2 62.7(s), 207.7(s). 284.6(s). 492.2(s). 555.0(s) 284.6 11.8 n:.7(s),2O7.7(s).232.O(sj. 351 Ksl. 423.61s), 439.8(s).5l2.3(s).655.9(m) 347.2 2.8 207.7(s).439.8(s),5l2.3(s) 351.0 5.6 436.1(s) 351.1 4.6 62.7l.sl, 284.61s). 512.3(s). 859.51s) 423.6 12.7 62.7(s), 232.0(m). 284.61s). 436.1(5). 724.3(s) 436.1 2.4 347.2(s) 439.8 12.3 62.71s), 2MMa). 347.2(s). 423.61ml 492.2 4.0 62.7(si,232.0(s).655.9(s) 512.3 2.3 284 6(s),347.2(s), 351.1(8) 555.0 3.0 232.0(5) 655.9 2.5 2u7.7(m),492.2(m). 555.0(5) 724.3 100.0 423.6(s) 859.5 4.2 351.11s) 1147.9 12.3 62.5(s) 1210.6 1.9 None

tThe uncertainty in the -y-ray intensities is estimated to he 110% for the strong peaks (above 10% relative intensity) and ±20% for the weak ones. tThe uncertainty in the y-ray energies is estimated to be ±0.5 keV. s ( = strong) indicates that the peak is higher than 3

5 J! ^ » it "

g» w © W «? W SS S C •• ? £ 1 2 :- o s „ g n K n 859 6 7871 r3- S1 'S

O IV r. S S8 S ø Æ •" 5

T- r* * S S S r— 3510 347«

«2 7 \ 00

Frg. 3. Partial decay scheme of l4lCe. The scheme is based on the coincidences listed in Table I. 1932 T. BJORNSTAD el al.

scheme, the levels at 62.7, 787.1, 859.6 and 1210.7 keV few weak transitions namely the y-rays at 360. 468 and are in agreement with the results of Ohyoshi et al. [7] and 491 keV proposed by Hoffamn et at. We have also found Hoffman et al.[4,5] No evidence for the level at 300 keV evidence for a 35.0 keV y-ray deexciting the 35.0 ke V level proposed by Ohyoshi el al. has been found. The level at proposed by Hoffman et al. [6]. This y-ray was not ~35OkeV, proposed both by Ohyoshi and Hoffman has observed by the latter group, probably due to a com- been split up into two levels at 347.4 and 351.0 keV, bination of the low intensity and its situation in the close respectively. We also propose at new level at 555.0 keV, vicinity of the light lanthanide y-ray energies. decaying to the 347.4 keV level via a rather strong l47Ce and l48Ce. It is not possible to assign y-ray 207.7 keV y-ray and to the 62.7 keV level and the ground energies specifically to 141Ce or 148Ce solely on the basis state via a 492.2 keV and a 555.0 keV y-ray respectively. of their half-lives, because these are very similar (56 s ""'Ce. Although 14 min '*Ce was carefully studied by and 48 s. respectively). In an earlier work [14] we have Hoffman et al. [6], we have performed a re-investigation therefore reported on coincidence measurements which of (his nuclide. The results obtained are shown in Table 2 facilitated such assignments. In the evaluation of these and the decay scheme in Fig. 4. Our scheme is a verifica- experiments, all peaks showing coincidences with the tion of the scheme derived by Hoffman et al. except for a strong 269 keV peak, or with peaks in coincidence with the 269 keV peak, were attributed to l47Ce. Analoguely. the 292 keV y-ray peak was used fov the assignment of Table 2. Energies, relative intensities and coincidences observed y-rays to Ce. The assignment of these two strong for y-rays assigned to ''"'Ce

Energy Rel. int.t Table 3. Energies, relative intensities and coincidences observed E, (keV) I, (.%) Observed coincidencest for y-rays assigned lo M Ce

35.0 — 317.KS) Energy Rel. int.t 52.2 5.2 264.91s) Ey(keV) 1, (%) Observed coincidences* 87.0 2.7 264.9(s) 98.7 12.6 218.5(s) 92.9 51.1 198.9(s). 269. l(s), 374.4(s), 254.4(w). 439.8(s). 101.0 4.4 251.2(s) 452.3(w),832.2(s) 106.3 0.3 None 178.4 5.3 198.9(m),289.9(m) 133.7 21.8 218.5(s) 198.9 25.2 178.4 218.5 41.3 98.7(s). I33.7(s) 2899 19.3 178.41m), 254.4(w) 251.2 10.0 lOl.O(s) 362.0 6.2 439.8(5). 832.2(s) 264.9 38.3 52.2(s),87.0(s) 374.4 43.4 92.9(s> 317.1 100.0 35.0(s) 439.8 92.9(s), 269. l(s). 362.0(s) 352.1 0.6 None 452.3 28.0 92.91s) 369.8 5.4 98.7(m), 106.3(m) 467.3 29.8 None 415.9 5.5 None 832.2 15.7 92.9(5). 269. l(s).362.0(s| 503.1 6.2 None 1194.2 <3 None

tThe uncertainty in the y-ray intensities is estimated to be tThe uncertainly in the y-ray intensities is estimated M be ±20%. ±20%. tThe uncertainty in the y-ray energies is estimated to be tThe uncertainty in the y-ray energies is estimated to ne ±0.5 ke V. ±0.5 keV. For comments, see Table 1. For comments, see Table I.

146,n*' 7[/2 = l4min i

W u? 9 A A ffl - o. « <•) tf) 0> O r- (O m * <*> <& & W g. w W S g S S. .5 S

3521

^_ M15 =^ 1337 W10 "^— «70 i 3S0 '— 00

I46P| Fig. 4. Partial decay scheme of l4*Cc. The scheme is based on the coincidences listed in Table 2. Decay properties of some neutron-rich cerium isolupe y-rays to "*'Ce and te, respectively, was based on The doublet nature of the 269 keV y-ray peak makes their half-lives. the mass assignment of the two components difficult. In the present investigation, the detectors used had a However, we have solved the problem in the following better resolution than in the experiments mentioned way: The 242 keV y-ray peak is still attributed to "C"e above. This revealed that the 269 keV peak is nol a single due to its short half-life (45 ± 5 s.l. The 292 keV y-ray is peak, hut a closely spaced doublet. The energy difference in coincidence with a BOkeV y-ray. which in turn is between the two peaks is about 0.7 keV. coincident with the 269 keV component having the higher

.._ __ 1194 2

1 O Cl 00 tf n» N CH O C 1 M .__ (^ « c 1 3 S 0 1 5 1 i U ^ N ^ <"1 *~ 1 1 i i 1 544 2 i i i 467 3 i i i i 362 0 i , 289 R ' i *

• i J 92 9 T I47pr Fig. 5. Partial decay scheme of K7C'e The scheme is based nn the coincidences lisied in Table 3.

121 2 116 8 98 5 90 4

I48P, f-'ig ft Partia) decay scheme of UPtCc The scheme is ba^ed on the coincidences listed in Table 4. 1934 T. BJORNSTA'J t-f al

Table 4. Energies, relative iniensiiies ane coincidences ohserved fnr y-rays assigned te ' *C

Energy Rel. inl.t E,(keV) 1, <%> Observed coincidences}

74.5 Expected transition 90.4 14.8 105.2(s;, I95.7M. 269.7M. 325.01s). 374.41s) 98.5 75.9 191.7(ml. 291.8N. 269.7(s). 325.01s). 3h9.3|s|. 422.1« s I 105.2 29.5 90.4IM. 195.71s). 273 8(s). 325.0(s,i 116.8 18.3 273.8(sl 121.2 72.2 233.7[m).269.7(si. I^M.'WM 130.1 3.4 195.7(m>. 269.7isl. JW.xisi. WIlNwi 168.3 4.3 Not gated 191.7 7.7 98.5

+The uncertainty in the y-ray intensifies is estimated !o he ^20^ tThe uncertainty in the y-ray energies is estimated to he ±0.5 keV. For comments, see Table ! energy. i\.. makes the assignment of this y-ray to Ce 4. D. C. Hoffman and O. B. Michelsen. Kjeller Report KR-76 probable but not unambiguous. (1965,. y-rays in coincidence with the 269.7 keV y-line have 5. D. (". Hoffman. () li. Michelsen and W. R. Daniels. Ark. Fvj then been assigned to l4*Ce. while those coincident with 36. 211 (1966). the 269.0 keV y-ray have been assigned to l47Ce. 6. D. C. Hoffman. F. O. Lawrence and W R. Daniels. P/ni. Rev. 172. 1231 ii%8) The assignments are in good agreement with recent 7. A. Ohyoshi. [; Ohvoshi. T. Tamai and M. Shinagawu. J. findings of Blachot et al. [16], who used a mass separator Inorg. Nucl r hem. H. 3293 (1972) to obtain their samples. 8. K. ii. Seyb. Jahresberichl 1972 (Inslitut fur Kernchemie der J The coincidences obtained for ' Ce and '""Ce are L'niversitiil Mainz). BMFr-FB K 73-22 ||973|. presented in Tables 3 and 4, respectively. The cor- 9. K I: Sc\b. Jahrisbericht 1973. Institut fiii Kernchemie tier responding decay schemes are shown in Figs. 5 and 6. llnivarsitiit Main?. Mainz. 10. P. O Aronsson. H. R. Johansson. J. Rvdherg. Ci. Skar- Acknowledgements—The authors are indebted to Professors G. nemark. J Alstad. B. Bergersen. E. Kwile and M. Skarestad. Herrmann. A. C. Pappas and J. Rydberg for their interest in our J. Innrg Nut I. Chem. 36. 2397 (19741. work. We are also indebted to Dr. N. Kaffrell, Dr. N. Trautmann II P (). Ariipssnn. Thesis. Chalmers lni\ersiu of Technolog>. Dipl. Chem. E. Stender and Mr. R. Heimann for k'nd assistance Gbteborj? 119741. during the experiments and to the staff of the Mainz TRIGA 12. P. O. Aninsson. (i. Skarneniark and M. Skaresiad. / Innrg. reactor. We also gratefully acknowledge the financial support Nucl. Chan 36. 16M (1974). from Ihe Swedish Atomic Research Council, the Bundes- 13 P. O. Arun^nn. (i Skarnemark and M Skarestad. Ituir^. ministerium fur Forschung und Technologie and the Norwegian Nucl. Ciifm. l.m. 10. 499 119741. Research Council for Science and Humanities. 14. P. O. Aron.ison. G. Skarncmark. E. Kiåle and M. Skareslad. Inorg. Nucl. Chem. leu. 10. 753 11974). REFERENCES 15. N. Trautmann. P. (>. Aronsson. T. Bjiirnsl.ul. N. Kaffrell, E. 1. O. Hahn and F. Strassmann. Natimiss. 31. 499 (1943). Kvåle, M. Skarestad. G. Skarnemark and E. Slender, limrn 2. S. S. Markowitz, W. Bernstein and S. Katcoff. Phvs. Rer. 93. Nucl. Chem. I.eu. II. 729(19751. 178 (1954). 16. C. Duvilliers. (.". liche and J. Blachot. Centre d'Hludes 3. D. C. Hoffman and W. R. Daniels. J. Innrg. Nucl. Chem. 26. Nucleaire. (irenohlc. Private Communication (March 19761. 1769(1964) PAPER P.VI. Nuclear Instrumentsand Methods 171 (1980) 323-328 © North-Holland Publishing Company

AN IMPROVED SYSTEM FOR FAST, CONTINUOUS CHEMICAL SEPARATIONS ("SISAK 2") IN NUCLEAR SPECTROSCOPIC STUDIES

G. SKARNEMARK, P.O. ARONSSON *, K. BRODÉN, J. RYDBERG Department of Nuclear Chemistry, Chalmers Unhersiry of Technology, S-412 V6 (ioteborg,. Sweden

T. BJORNSTAD Department of Nuclear Chemistry, University of Oslo. Oslo S, Norway

N. KAFFRELL, E. STENDER and N. TRAUTMANN Institut fur Kernchemie, Universitet Mainz, D-6 500 Mainz, FRC

Received 1 November 1979

An improved rapid, continuous solvent extraction system ("SiSAK 2") is described. The system is connected to a gas-jet installed at the Mainz reactor. It allows single or multistage chemical separations of liquid phases by means of specially designed centrifuges within ~1 s per stage. The application of this system to study short-lived nuciides is exemplified for neutron-rich lanthanum and cerium isotopes produced by fission.

1. Introduction energy carried by nuclear reaction products them- selves for their separation in electric and magnetic Though fission products have been studied for fields. Recoil separators yield no selective element almost 40 years, there are still nuciides with half-lives isolation. shorter than a few minutes, possessing insufficiently The SL'putation of single elements from complex investigated or unknown decay properties. Therefore, reaction product mixtures can be accomplished with considerable efforts have been devoted to the devel- chemical procedures. For a long time, chemical opment of fast and selective procedures to isolate methods had the disadvantage of being rather slow. these nuciides. The methods used so far have mainly This limited their use to nuciides with half-lives of at been mass separation 11—3] or chemical procedures least seveul minutes. However, to perform faster (4,5]. The former method; with the advantage of a chemical separations, new techniques for off-line pro- rapid and unambiguous isobaric separation, has been cedures have been developed (4,9| allowing studies of applied to a variety of elements for which suitable nuciides with half-lives down to ~0.5 s. The disadvan- ion-sources are available. Rapid diffusion and evapo- tage involved is the discontinuous performance, i.e., ration of the desired element out of the target system an experiment has to be repeated several times to is a prerequisite and a separation with respect to the yield reasonable counting statistics. Such methods atomic number may be achieved if the diffusion— can often not compete with continuous procedures, evaporation and/or the ionization are selective, e.g., especially in detailed coincidence experiments. as for the alkali metals [6]. The demand for volatility For fast, continuous chemical separations we limits the use of conventional mass separators; thus it developed the SISAK technique [IO--I2| which uses has not yet been possible to isolate e.g., Zr, Nb and multistage solvent extraction procedures where the other non-volatile elements with such facilities. An separalion of the two phases is accomplished by cen- exception are the recoil separators |7,8] which uti- trifuges. A set-up of four centrifuges was installed al lize, instead of evaporation, the charge and kinetic the Mainz reactor in combination with a gas-jet recoil transport system [I.11. The shortest half-lives acces- * Present address: Ringhals Power Station, Varobacka, sible with this system were 3-5 s. Sweden. The present paper describes an improved system

323 324 G. Skarncinark et al. i Continuous chemical separation

"SISAK 2". in which smaller centrifuges are used. directly to the inlet of the centrifuge. The outlets are This new system allows on-line nuclear spectroscopic p/ovided with pressure gauges and throttle valves for studies on nuclides with half-lives down to 1 s. Since manual operation. The valve units have been designed gas-jets have widely been applied at accelerators, the to give a minimum hold-up time. The present system technique outlined in this paper can also be used for installed at the Mainz reactor consists of four centri- products from charged particle induced reactions. fuges mounted in one compartment to reduce the transport time between the centrifuges to a mini- mum. 2. Experimental The "SISAK 2" control panel is built in modules to allow simple replacement of failed parts. It is 2.1. Gas-jet system and degassing unit equipped with controls for the centrifuges and pumps and instruments for remote temperature reading, as For the continuous transport of fission products well as a level alarm unit connected to resistive gauges out of the reactor to the wet chemistry system situated in storage vessels (low level alarms) and waste "SISAK 2", a gas-jet recoil transport system with containers (high level alarms). chloride clusters dispersed in nitrogen was used [14], Transport yields of 60 to 107<- for fission 2.3. Nuclear detection equipment products produced by irradiation of 215U and 239Pu targets with thermal neutrons were achieved even The nuclear detection systems presently used in over an operation period of several weeks. The reac- combination with the SISAK equipment consist of tion products are dissolved from the clusters in a Ge(Li) detectors with standard electronics and allow static mixer (Kenics Corp., USA) in which the jet gas measurements of single 7-ray spectra, multispectra is extensively mixed with a liquid phase, usually an analysis, multiparameter coincidences and y-y angu- acid solution. lar correlations. A system for measuring ^-ray spectra In order to get rid of the noble gases and to obtain is under development. a good phase purity, the gas-liquid mixture is fed For ihe investigation of very short-lived nuclides. into a degassing unit working like a cyclone. Inside the liquid phase containing the activity is pumped the degasser are baffles to break the rotation of the through a cell placed in front of the detectors. In tangentially injected liquid. The liquid, collected in a order to study longer-lived isotopes, the cell can be conical bottom, is then transferred continuously with filled with ion exchange resin or extraction cliro- a pump to the first extraction step of the "SISAK 2" matographic materials which retain the activity by system; the upper outlet of the degasser is connected adsorption. to a suction system to remove the noble gases together with the carrier gas of the gas-jet. 3. Results 2.2. The "SISAK2"system 3.1. Delay properties of the "SISAK 2" system The most important improvement of "SISAK 2" is the new H-10 centrifuge [15], a scaled down version The delay in a two-stage SISAK set-up was deter- of the H-32 centrifuge [16]. Compared with the lat- mined by operating the reactor in the pulse mode and ter the volume of a H-10 centrifuge is reduced by measuring the activity at the detector sile with a almost one order of magnitude and amounts to only multichannel analyzer operated in the multiscaling 12 cm3, whereas the maximum throughput remains mode. The width at half maximum of the neutron unchanged or can even be increased to an upper limit pulse was ~-30 ms. In the system used, shown in fig. of 23 cm3 s"1 per phase, corresponding to a mean 1, an element was extracted in Ihe first centrifuge hold-up time of —0.25 s for each of the two phases. step Cl from aqueous phase I inlo the organic phase In most experiments described below, the H-10 cen- and back-extracted in the second unit C2 into trifuges were run with flow rates between 10 and 20 aqueous phase 2. cmV. The time profile of the "SISAK 2" system, given In the "SISAK 2" application, each H-10 centri- in fig. 2. was obtained with lechnetium activity fuge is equipped with a static mixer connected separated from fission products. This chemical procc- C. Skamemark et at. / Continuous chemkal separation 325

G*s-i«t

Aqueous ph««> 1 Aqusous phas* 2 lig. I. Schematic diagram of the "SISAK 2" set-up used for delay property determinations. DG = degassing unit, C], C2 = centri- fuge steps, D = Ge(Li)-detector. dure, which involves extraction of Tc as TCO4 from is obtained after 5.2 s. The corresponding values for

O.I M KBrO3 and O.I M HNO3 into 0.05 M Alamine- the gas-jet without the SISAK system are 0.7 and 1.1 336 in CHCI3 and a subsequent back-extraction with s, respectively. 2 M HNO3, will be described in a forthcoming paper [I7j together with continuous separation methods 3.2. Study of short-lived La- and Ce-isotopes for Zr, Nb, Br and I. From the elapse of the curve, one recognizes that the first activity reaches the The "SISAK 2" system was first applied to the detector after ^2.8 s while the maximum count rate isolation of neutron-rich isotopes of La, Ce and Pr

Counts/0.1 min

1000

BOO

0.0 IJO 30 4.0 SO • 0 70 •0 •0 WO Thm («I Fig. 2. Time profile of the technetium activity separated from fission products and measured with a Gc(LiHetector after palling the SISAK set-up shown in fig. 1. 326 G. Skarnemark et al. / Continuous chemical separation

from fission products, using chemical procedures state. The chemical procedure for cerium involves its already utilized for on-line separations [12,18]. La oxidation to Ce(IV) with subsequent extraction into and Pr are isolated by extraction into di-(2-ethyl- HDEHP from ! M HNO3 and a stripping process with hexyl)-orthophosphoric acid (HDEHP) from a nitric a reducing agent. These experiments yielded informa- acid solution (pH 1.4) and back-extraction with 1 M tion on the decay of i47La and I48La. A (2.2 ± 0.3) s

HNO3 after oxidation of cerium to the tetravalent activity in the lanthanum fraction was assigned to

too

too

Fig. 3. (a) Decay curve of the y-ray peak at 117.9 keV. (b) Growlh-and-dccay curve of the 269.1 kcV y-ray peak. Both yrays have been measured in the La-fraction. C. Skarnemark et al. / Continuous chemical separation 327 mass 147 because there is good agreement between 148La has also been reported in ref. [19] with an the half-lives obtained in measurements of individual energy of 158.7 keV. In addition to these activities, 145 146 7-rays and from the growth-and-decay curves of peaks 7-rays of La (TU2 = 25.2 s), La (Tln = 8.5 s) 147 150 belonging to the daughter product Ce. As an and Pr (r1/2 = 6.2s) could be identified in the example, fig. 3a shows the decay curve of the com- spectrum. plex 7-ray peak at 117.9 keV. The component decay- When we started these investigations no detailed ing with a half-life of (25.7 ± 4.5) s belongs to 145La data had been published on the decay of 149Ce and and the shorter one decaying with a half-life of 1S0Ce. Recently, 7-ray energies and half-lives belong- (2.2 ± 0.3) s is attributed to I47La. A half-life of 2.5 s ing to isotopes of the mass chains 149 and 150 were for 147La follows from the growth-and-decay curve of reported by Pfeiffer et al. [21] from experiments at the strongest 7-ray peak of 147Ce at 269.1 keV as the Lohengrin separator. shown in fig. 3b. Data on 147La have also been repor- The most intense short-lived 7-rays present in the ted by Devillers et al. [19] working with the fission Ce fraction separated with the "SISAK 2" system fragment separator Lohengrin who obtained a half- exhibit half-lives of (5.7+0.5) s and (4.8 ± 0.5) s. life of (4.0 ± 1.0) s. In the SISAK experiments, the The 4.8 s activity could be assigned to ls0Ce from the half-life of 147La was determined by running the sys- growth-and-decay curve of ' S0Pr [22], while the 5.7 s tem in a cyclic mode with subsequent decay mea- activity was attributed to 149Ce. This latter assign- surements. ment could not be based on the growth-and-decay 149 Figure 4 shows a 7-ray spectrum of short-lived La- curve of Pr due to the low 7-ray intensities of that isotopes. The assignment of the 7-ray energies to nuclide. The half-life values are in agreement with I47La is based on half-life determinations. Another Pfeiffer et al. [21] who obtained (5.0 ±0.5) s and 149 150 short-lived activity with a half-live of ~1 s and the (4.8 ± 0.6) s for Ce and Ce, respectively. strongest peak at 158.5 keV was also seen in the Since the half-lives of 149Ce and I50Ce are quite lanthanum fraction. This activity is attributed to the similar, 7-7 coincidence measurements were neces- decay of 148La in accordance with the energy differ- sary for the unambiguous assignment of weaker ence between the 2* -> 0* levels in 148Ce measured by 7-lines. Cheifetz et al. [20]. The most dominant 7-line of

SOOO n— i r 1 —i 1 1 1 14 6

10 •

å \ IA \« • Ifi tN ti N *"_ > 1 z o )< *^IO • _i • ui 2500 IJ jr jr * 14 6 10 Tg « • «• f) .Jf« r» -• • r* 1™» r- «I ib* «s ni n m N is 22 70. 2 L M 2 La - mm O n r* I ID ss sis ss 43« . IS i W i • AL I ' 1 i tmo ( f n|n i ( i - AV" rJ*L f *>* 500 1000 CHANNEL NUM«£H Fig. 4.7-ray spectrum of the La-fraction. To obtain this spectrum, the third and fourth spectra of a decay measurement (recorded 4-8 s after start of counting) were subtracted from the first two spectra (recorded 0-4 s). 328 G. Skarnemark et al. / Continuous chemical separation

The authors are indebted to Professor G. Herr- (81 P. Armbruster, Proc. 3rd Int. Conf. on Nuclei far from mann and A.C. Pappas for their interest in this work, stability, Cargcse 1976, CERN-Report 76-13,p. 3. to Mr. R. Heimann for his kind assistance during the (9] J.M. Nitschke, Nuclear Chemistry Division Annual Report 1973, Lawrence Berkeley Laboratory, Univer- experiments and to the staff of the Mainz TRIGA sity of California, Berkeley, 1973, Report LBL-2366,p. reactor. H. Bratt, L. Båtsvik, L.E. Ohlsson and H. 446. Persson were responsible for the mechanical construc- [10] P.O. Aronsson, B.E. Johansson, J. Rydberg, G. Skarne- tion work. mark, J. Alstad, B. Bergersen, E. Kvlle and M. Skare- stad, J.inorg. Nucl. Chem. 36(1974) 2397. Financial support from the Swedish Natural [11] P.O. Aronsson, Thesis, Chalmers University of Tech- Science Research Council, The Norwegian Research nology, Goteborg, 1974. Council for Science and Humanities and the Bundes- [12] G. Skarnemark, Thesis, Chalmers University of Tech- ministerium fiir Forschung and Technologie is grate- nology, Goteborg, 1977. fully acknowledged. [13] N. Trautmann, P.O. Aronsson, T. Bjornstad. N. Kaffrell, E. Kvile, M. Skarestad, G. Skarnemark and E. Stender, Inorg. Nucl. Chem. Letters 11 (1975) 729. [14] E. Stender, N. Trautmann and G. Herrmann, submitted References to Radiocliem. and Radioanal. Letters. [15] J. Rydberg. H. Persson, P.O. Aronsson, A. Selme and G. [ 1) S. AmieJ and G. Engler (eds.), Proc. IXth. Int. Conf. on Skarnemark, Hydrometallurgy, in press. Electromagnetic isotope separators and related ion [16| H. Reinhardt and J. Rydberg. Acta Chem. Scand. 23 accelerators, (North Holland, Amsterdam 1976). (1969)2773. [2] W.L. Talbert Jr, Int. Conf. on the Properties of nuclei [17] K. Brodiin, G. Skarnemark, T. Bjiirnstad, D. Eriksen, I. far from the region of beta-stability, Leysin 1970, Ilaldorsen, N. Kaffrell, E. Stender and N. Trautmann, to CERN-Report 70-30(1970) p. 109. be published [3] HX. Ravn, Phys. Rep. 54 (1979) 203. [18] P.O. Aronsson, G. Skarnemark and M. Skarestad, J. [4] N. Trautmann and G. Herrmann, J. Radioanal. Chem. Inorg. Nucl. Chem. 36(1974) 1689. 32(1976)533. 119] C. Devillers, C. Hiche and J. Blachot, Centre d'Etudes [5] N. Trautmann, Proc. 3rd Int. Conf. on Nuclei far from Nucleates, Grenoble, private communication 1976. stability, Cargcse 1976, CERN-Report 76-13 (1976) p. [20 [ F. Cheifctz. J.B. Wilhetmy, R.C. Jarcd and S.G. Thomp- 30. son, Phys. Rev. C4 (1971) 1913. [6| R. Klapisch and R. Bernas, Nucl. (nslr. and Meth. 38 [21J B. Pfeiffcr, J.P. Bocquet, A. Pinston, R. Roussillc, M. (1965)291. Asghar, G. Baillcul, R. Decker, J. Greif, H. Schrader,G. [7| P. Armbruster, M. Asghar, J.P. Bocquet, R. Decker, H. Sicgert. II. Wollnik, i. Blachot, E. Monnand and I'. Ewald. J. Greif, E. Moll, B. Pfeiffer, H. Schrader, F. Schussler, J. Physique 38 (1977) 9. Schussler, G. Siegert and H. Wollnik, Nucl. Instr. and [22] P.O. Aronsson, G. Skarnemark and M. Skarestad, Inorg. Meth. 139 0976)213. Nucl. Chem. Letters 10(1974) 499. PAPER P.VI I. Nuclear Instruments and Methods 185 (198]) 175-180 175 North-Holland Publishing Company

AN AUTOMATIC DEVICE FOR SAMPLING OF THIN ASSAYS OF SHORT-LIVED RADIONUCLIDES IN A LIQUID FLOW *

S.I. NOV1K **, T. BJORNSTAD, J. ALSTAD Department of Nuclear Chemistry, University of Oslo, Blindern. Oslo j. Norway K. BRODÉN and G. SKARNEMARK Department of Nuclear Chemistry, Chalmers University of Technology, S-412 96 Gatenborg, Sweden

Received 23 December 1980

A semicontinuously working apparatus for isolation of tracer amounts of radioactive nuclei from a continuous liquid flow (10-15 ml/s) has been constructed. The radioactivity is isolated on thin, porous plastic membranes impregnated with ion-cv change resin or extraction agents, or on a thin preformed precipitate. Chemical yields of ~20% have been achieved for Br. Zr. Ce and Tl at on-line separation conditions. The time sequency for the different operations involved can be optimized with respec! to a certain nuclide by proper programming of the timer that controls the sequence. In combination with the fast, continuous chem- ical separation system SISAK the new apparatus allows ^-measurements and delayed neutron measurements as well as X-ray and low energy y-ray measurements on nuclei with half-lives down to a few seconds. The feasibility of the method with respect to (3-measurements is demonstrated on nuclei with well known Øø-values.

1. Introduction 7-rays (>20 keV). To take better advantage of the nuclides delivered by the SISAK system, we have Studies of properties of nuclides far from the ^-sta- developed an apparatus which produces thin samples ble region has many important theoretical aspects, suitable for /?-. X-ray, low energy 7-ray and delayed and several experimental techniques have been devel- neutron spectroscopy. oped to provide the necessary experimental data Several techniques for making thin samples from a [1,2]. One of these techniques is the fast, continuous continous liquid flow have been considered, e.g. elec- chemical separation system SISAK [3,4], which is trospraying and electroplating. Neither of these meth- based on single or multistage solvent extraction pro- ods, however, is able to handle the high liquid (low- cedures. Phase separation is accomplished by use of rate involved, nor is the deposition rate high enough to specially designed liquid-liquid centrifugal separators, allow studies of nuclides with short half-lives. We called H-centrifuges [5,6]. Presently, the SISAK have therefore chosen a semicontinuous method in equipment can deliver radiochemically pure samples which the radionuclide under investigation is of nuclides with half-lives down to less than 1 s. The absorbed on collectors containing ion-exchange resin, technique has been employed for 7-ray singles, yy(i) extraction agent or a preformed precipitate. The coincidence and yy angular correlation measurements apparatus has been throroughly tested under off-line on several nuclides [7-9]. as well as on-line conditions, as described below. A drawback connected to this technique is (hat the short-lived nuclides are delivered in a liquid phase (aqueous or organic), which limits the feasibility for 2. Experimental spectroscopic measurements to relatively energetic 2.1. Collators

Experiments show that the cation exchanger im- • This work is based partly on the Thesis of Svein I. Novik. University of Oslo. 1980. pregnated Acropor * SA-6404 (Dowex 50 W X 8) and •• Present address: Institutt for Energiteknikk. 1750 Halden. Norway. * Gclnian product

0 029-554X/81/0000-0000/S02.50© North-Holland 176 5./. Novik et al. /Automatic sampling device

Table 1 air. The procedure for the preparation of the collec- Properties of ion exchanger collectors tors is similar to the one described above, except that pretreatment with deionized water is omitted. Collector Acropor Acropor SA-6404 SB-6407 For some applications, a preformed precipitate (cation) (anion) offers a specific way of accumulating selectively the nuclide under investigation. Thus Br~ can be caught Thickness (mg/cm2) 6.63 5.76 on a AgCl-precipitate. In this case, the collector con- 2 Resin weight (mg/cm ) 1.87 1.56 sists of an AgCl-ioaded paper filter, on PVC-backing Capacity (meq) 4.80 x 10"3 2.49 X 10"3 covered with another paper filter to avoid damage (8 mm source diameter) from the liquid flow. This type of collector is too thick (—20 mg/cm2) to allow good (^-measurements, but it can be used for low energy 7-ray and neutron the anion exchanger impregnated Acropor SB-6407 spectroscopy. (Dowex 1 X 8) are suitable for use in the apparatus. These collectors offer high mechanical strength in 2.2. Collection device combination with good chemical resistance against e.g. traces of organic phase in the aqueous solution A schematical picture of the main collection sys- pumped through. They consist of acrylonitrile polyvi- tem is shown in fig. 2a. Approximately 50 collectors nylchloride copolymer membranes reinforced with are piled up in a storage cylinder. A pneumatic piston nylon fibers. Table 1 shows the main properties of pushes the lowest-lying collector into the collection the cation and anion exchanger collectors, respec- position. Another pneumatically operated piston then tively. The collectors are prepared by punching cir- presses the upper part of the collection cylinder with cular discs (diameter = 25 mm) out of Acropor 1000 N against the lower one, the collector disc lying sheets. The ion exchanger resin is then allowed to in between. The high pressure prevents leaks during swell in de-ionized water for a few minutes. There- the collection process. The viton o-rings on both col- after, the collector discs are stretched and mounted lection heads limit the source diameter to 8 mm (cf. between two conical steelrings as shown in fig. 1. fig. 2b); this measure corresponds to the dimension The inner diameter of the steelring is 20 mm to •jf the Æ detector in the /3-telescope to be used for reduce the mass in close vicinity to the source (the ^measurements [10]. diameter of which is 8 mm) in order to minimize the When switching a pneumatically operated stop- spectrum distortion caused by absorption and scat- cock, the liquid flow from the SISAK separation sys- tering of the radiation. tem is fed into the collection device. To minimize the The ion exchangers have been efficiency-tested by pressure variations, suction is applied by connecting adsorbing radionuclides from aqueous solutions: the eluate waste-barrel to a vacuum pump. This mea- 2 H4Ce4+_ I44pr3^ 204-j^ "ZrO * and ""TcOi. sure is necessary to maintain good phase separation in To prepare filter membranes impregnated with an the H-centrifuge delivering the feed solution. After extraction agent, a filter sheet is inserted into a solu- collection, the sample is dried by passing hot air tion of for instance ~2 M HDEHP (di-2-ethylhexylor- through it, and then a pneumatically operated arm thophosphoric acid) in toluene, and allowed to dry in kicks it out of position into the transfer channel. Compressed air is applied to forward the sample into the counting position (cf. figs. 2c and d). After the Steel ring counting period, the sample is released by withdraw- ing electromagnetically the sample-supporting bar. The duty cycle can be optimized by loading a new sample as soon as the first has left the collection posi- tion and releasing the counted sample shortly before the new one arrives. The entire operation sequence is or controlled by a programmable timer. Filter membrane f'ig. 1. Design of the collector discs. The collector membranes are mounted on a frame of two steel rings

Compressed air

Steel spring Piston with airtight shield In direction counting position for return of Pressureoperated piston piston (titanium) Fig. 2c. Cross-section of the transfer channel at the position of the inlet for the compressed air, (11) in fig. 2a. When the magnetic valve opens for the compressed air, the piston (right-left hatching) is pressed down. A channel opens for Incoming liquid the air in direction of the counting position, while the direc- tion of the collection house is blocked by the piston sup- ported with rubber tighteners.

Fig. 2b. Cross-section of the collection part of the sample preparation apparatus, (5) in fig. 2a. The housing and the baseplate are made of PVC (right-left coarse hatching). The movable upper and stationary lower filtering heads are made of corrosion resistant titanium (left-right hatching), while quid to the "o"-rings are made of vitron. vacuum barrel 178 S.I. Novik et al. i Automatic sampling device

.'• • ' -•' .' / / ' < /- 7 '. // -x > / Sample holder .n detection posi( on ' / Sample recoil inhibitor

//// • / ,. ., / J»u Sample in f ' / ' / detection posit io

Direciion o' sample release

Fig. 2d. Detail of the sample holder in the counting position, (13) and (14) i» lig. 2a. The sample positioning is very reproducible. The steelring hits and tests on the stop and release bar. A steel spring prevents any recoil out of position.

3. Results from test experiments /3-speclroscopic studies should not exceed ~-l min due to the increase in sample weight caused by capillary 3.1. Performance effects resulting in absorption of water from the sur- rounding moisture. Tests have been made to investigate the recovery of the radionuclei on the collectors, the on-line 3.2. Delay properties [11,12] performance of the system and the feasibility of the method with respect to ^measurements. An important question in connection with this The yield depends on the chemical conditions in apparatus is for how short-lived nuclei it can be used. the passing liquid. For each product to be collected, It is here convenient to divide the transport time the optimal combination of oxidation state, pH, ion for a radioactive species from the target to the detec- strength (and type), ion exchanger (or extraction tor into four parts, namely the time necessary for the agent) and collection time has to be evaluated. This is transport from the target through the SISAK system mainly a task for off-line experiments. However, the to the thin sample apparatus, the collection time, the conditions necessary for a proper on-line separation drying time and the transfer time from the drying po- in the SISAK system itself have to be maintained, and sition to the detection site. For a normal SISAK set- will thus affect the choice of chemical composition at up, the first transport time can be estimated to 5-7 s. the collection position. depending on the number of centrifuges involved. In off-line tests, chemical yeilds of ~20% were The collection time can theoretically be very short, found for 144Ce(IV) using either Acropor SA-6404 but for practical reasons we have to set the lower (cation exchanger) or a collector doped with HDEHP limit at approximately 2 s. The drying of the filter needs at least 3 s and the transport from (lie drying for passage of 500 ml 0.1 M HNO3 solution with a speed of 7-8 ml/s. Similar yields were obtained for position to the detector less than 0.5 s. Thus the shortest transport time from the target to detector Zr(IV) (absorption from 0.3 M HN03 on Acropor can be estimated to 10—12 s. Thus we estimate the SA-6404) and Br" (absorption from 0.1 M H2SO4 on a preformed AgCl-precipitate) in on-line experiments. shortest-lived nuclides which can be studied with this The test experiments also showed that a sample apparatus to have half-lives in the range 3 5 s. So far, the most short-lived nuclidc isolated with SISAK in drying time of 3 s at an air temperature of 5150° C l00 was necessary. The measuring time when performing combination with this apparatus is 7 s Zr. 5./. Novik el al. /Automatic sampling device 179

3.3. The collection time 3.4. Accuracy of the method

The optimum collection time per sample depends The applicability oi' the method to Qp-measure- upon the half-life of the nuclide to be studied, and on ments has been tested on some long-lived nuclides 144 204 the chemical composition and purity (for particles) with wellknown E0max, e.g. Pr and Tl. Conver- 137 of the liquid. The overall capacity of the collector has sion electrons from the two sources Cs (Ee- = 624 been checked by adsorption measurements of a long- keV) and 207Bi (£•„-= 481.7, 975.6 and 1682.2 keV) lived nuclide as a function of collection time. were used as energy calibration standards. The results The resulting count rate versus time curve can be obtained from the Fermi-Kurie plots of the recorded fitted with an expression composed of a sum of 0-spectra are: "growing in" terms of the type [] - exp(-i////)]- I44 Each of the terms can not be interpreted as express- Pr: ^„»x = GP = (29501 30) keV, 204 ing a physical process, but the sum is rather a conve- Tl: A>max=<2(j = (762 nient way of mathematically reproduce the gross These numbers are to be compared to the literature behaviour. Including the decay term the counting rate values of 763.4 keV [13] and 2996 keV [14]. respec- per sample as a function of the collection time may tively. As an example, the ^-spectrum for 144Pr and then be written: the corresponding Fermi-Kurie plot are given in fig. 3. exp(-Xf). (1) t=l where i//,- and R\0^ are constants for each apparent process ;' [^,- can be regarded as analogous to the radioactive decay constant X, and R.\o) as the count rate at saturation for each (unphysical) process /]. Their values can be determined in a computer fit. Formula (1) has a maximum where d/?(f)/df = 0 which gives:

0) S /?' [exp(-^r)-( j + X) - X] = 0 . (2)

An analytical solution of t from eq. (2) is not easily achieved. The optimum time is most conveniently 200 iOO 600 800 1000 found by a graphical method. The sum eq. (2) is cal- CHANNEL NUMBER culated for a few /-values chosen so that both positive and negative values are obtained for the sum. A smooth curve connecting the points intersects the abscissa at t = fopt. This curve is fairly well approxi- mated by a straight line in a limited time interval around /op[. Normally m = 2 or 3 is sufficient to fit the filtra- tion curve. An expecially simple expression is achieved in the case where the curve can be suffi- ciently well represented by a single term. Eq. (2) then reduces to:

X) = X , (3) 20 ENERGV (Mev) with the solution: Fig. 3. (a) 0-spcctrum of 14

4. Conclusions [2] R.A. Meyer and E.A. Henry, in Nuclear spectroscopy of fission products, (Institute of Physics Conference Series The experiments show that radionuclides delivered Nr. 51, London, 1980.) p. 59. [3j G. Skarnemark, P.O. Aionsson, I. Rydberg, K. Brodén, by an on-line chemical separation system can be T. Bjornstad, N. Kaffrell, E. Stender and N. Trautmann, adsorbed on a thin collector with an acceptable chem- Nucl. lnstr. and Meth. 171 (1980) 323. ical yield, and that the kinetic effects involved in the |4J P.O. Aronsson, B.E. Johansson, J. Rydberg, G. Skarne- absorbtion are negligible. They also show that the mark, 1. Alstad, B. Bergersen, E. Kvåle and M. Skarc- SISAK system and the thin sample apparatus work stad, J. Inorg. Nucl. Chem. 36 (1974) 2397. [51 J. Rydberg, H. Persson, P.O. Aronsson. A. Selme and well together. Finally, the /^measurements clearly G. Skarnemark, HydrometaUurgy 5 (] 980,' 273. indicate that it is possible to obtain quite accurate |6] H. Reinhardt and J. Rydberg. Acta Chem. Scand. 23 endpoint energies even if the thickness of the source (1969)2773. is ~5 ing/cm2. 17) J. Stachel, N. Kaffrell, E. Stender, K. Summerer, N. Traulmann, K. Brodén, G. Skarnemark, T. Bidmstad and 1. Haldorsen, Radiochim. Acta 26 (1979) 127. The authors are indebted to Prof. A.C. Pappas and 18] K. Summerer, N. Kaffrell, E. Stender. N. Trautmann Prof. J. Rydberg for their interest in our work, to K. Brodén. G. Skarnemaik, T. Bjornstad. 1. Haldorson Mrs. U. Knitz for her kind assistance during some of and J.A. Maruhn, Nucl. Phys. A339 (1980) 74. the experiments and to the staff of the Mainz TRIGA [9] N. Kafl'reii, E. Stender, K. Summerer, N. Trautmann, K. Brodén. G. Skarnemark, T. Bjornstad. 1. Haldorsen reactor. Thanks are also due to the Drs. N. Traut- and /.A. Maruhn, in Nuclear spectroscopy of fission mann, N. Kaffrell and P. Peuser for valuable discus- products (Institute of Physics Conference Series Nr. 51, sions and suggestions. London. 1980) p. 265. Financial support from the Norwegian Research 110J P. Peuscr, Thesis (University of Mainz, Main/, 1980). Council for Science and the Humanities and the [11 ] N. Trautmann, P.O. Aronsson, T. Bjornstad, N, Kaffrell, E. Kvåle, M. Skarestad, G. Skarnemark and E. Stender. Swedish Natural Science Research Council is grate- Inoru. Nucl. Chem. Lett. II (1975)729. fully acknowledged. (12] R.J. Silva. N. Trautmann. M. Zcndel. P.K Dinner and E. Stender, Nucl. Instr.and Mcth. 147 (1977) 371. [13] Nuclear Data Sheets 27 no. 4 (1979). References [14] Nuclear Data Sheets 27 no. 1 (1979).

[1] N. Trautmann and G. Herrmann, J. Kadioanal. Chem. 32(1976)533. PAPER P.VIII. Nuclear Instrument.; and Methods 1K8 (1981) 375 387 375 North-Holland Publishing Company

A CONTINUOUS ON-LINE METHOD FOR FISSION YIELD MEASUREMENTS WITH THE COMBINED GJRT-SISAK TECHNIQUE

Tor BJORNSTAD Department oj Nuclear Chemistry, University of Oslo. Blindern. Oslo 3, Norway

Received 20 March 1981

The principles of a method for fission yield measurements with the combined GJRT-SISAK technique are outlined. Only the simple case where parent effects can be neglected is considered in detail. The article includes derivation of a relation between the production rate and the counting rate of a nuclidc. The expression contains various efficiency parameters and the system depen- dent delay probability function /HO- The way to experimentally determine these parameters and the delay probability curve, as well as the fitting of an analytical function to the curve are shown. The method is general in the sense that the main principles are independent upon the nuclide under study, fissile target mate- rial and bombarding particle used.

I. Introduction

Since the first observation of the fission process in 1939 [1] much effort has been expelled in characterizing and describing it with the aim of a full understanding of the mechanisms working. To day a large amount of detailed data exists on mass and charge distribution, on angular and energy spectra of the emitted particles. But no single theory lias yet succeeded in unifying a substantial portion of the informa- tion available on fission. There is a constant need of new and in some cases more accurate data. To arrive at our present state of knowledge about fission, radiochemistry has been a necessary and higlily po- werful tool. But the chemical separation procedures used have mostly been manually operated, and therefore rather time consuming, limiting the number of nuclides possible to study to those lying close to stability. However, in the later years a number of new separation techniques has been invented. Mass {A) separators, even capable of achieving a separation in the proton number (Z), and automatically fast operating semicontinu- ous, or fully continuous chemical separation systems have been constructed. These systems car be connected on- line to the production unit to form semicontinuous or fully continuous on-line production-separation systems. These new techniques largely extend the number of nuclides accessible for study to include still more short-lived species. The combined GJRT—SISAK technique is one of these new techniques. It offers the possibility of separating nuclides with half-lives down to 0.5-1.0 s.

2. A brief recapitulation of the combined GJRT-SISAK technique

The GJRT (Gas Jet Recoil Transportation)-technique has been known for some years. It offers the possibility of transporting nuclear reaction products from the production site to the separation or direct measurement, by stopping and catching the species recoiling out of the target in a gas or a gas mixture, which continuously sweeps the target. Some of the large number of reported GJRT-systems are found in refs. [2-7J. The SISAK (Short-lived Isotopes Studied by AKUFVE tecluiique) system is a fully continuous multistage solvent extraction system based on H-centrifuges [8,9] for phase separation, the so-called AKUI-VE technique 1101. It is intended to deliver radiochemical pure samples of, in principle, any element of interest. The SISAK technique is described in refs. [11-I4|.

002')-554X/8l/0000-0000/S0:.5O© 1981 North-Holland 376 7". Bjurnstad I Fission yield nicasiirctnents

In early 1975 a GJRT system (N2/C2H4 gas mixture) was successfully connected to the S1SAK system at the Mainz TR.IGA reactor [15]. This reactor can be operated in the pulsed mode, thus making time behaviour mea- surements of the mass flow in the experimental system possible. Such measurements are the basis of the fission yield measurement method described in the present article.

3. Principles of the method

3.1. Derivation of a formula for the production rate of a nuclide

The distribution in transport time for products transported from production to detector position is expressed by the delay probability function usually denoted by p(t)dt and first introduced by Winsberg [16]. It expresses the probability of finding a delay between t and / + df. This function is a helpful notation in the process of con- verting the number of recorded counts of a nuclide to the production rate. In the following such a mathematical expression will be derived for the simple case where there is no parent effect.

A nuclide with decay constant X is formed in the target. Let the production rale be Rp. During the interval t to t' + åt' the number of atoms formed is dJVT=/?pdr'. (1) The recoil yield, i.e. the fraction of the total amount produced of a nuclide which recoils out of the target and into the jet gas, is denoted |. The fraction of the recoiled nuclides which reaches the end of the capillary (decay not included) is denoted •q and represents the CJRT efficiency. The parameter r? depends on the target and target chamber arrangement as well as on the length of the capillary tube. Accordingly the number of atoms reaching the end of the tube neglecting decay) is dA'ci^KpSrjdf1. (2) The chemical yield from the moment of mixing the jet gas with the first solution to the moment of measuring is denoted v. The number of atoms in the measuring position at an arbitrary lime /' when both decay and delay probability are included, is then given by dVc =Rvlnvåt exp|-X(/" - /')] p(t" - t'). {.?)

Assuming continuous production at a constant production rate Rr. the number of atoms in the counting position at an arbitrary time t" will 1

Nf= fdNc=Rpfrv [ exp|-X(r" - /')] p(t" - t') åt' . <4> b

It is experimentally shown that T? is time independent. Introducing the desintegration rate Dc

and the counting rate Rc R'c = eD'c = eW'c', (6) where e is the counting efficiency of the nuclide in concideration, gives r" R'c=RpZerit>\ f cxp|-X(/" - f')| p(t" - t')At'. (7) b The number of counts recorded in the time period /" to t" + d/" is dS = R'c'dt". (8) T. Bjornstad / Fission yield measurements 377

The total recorded number of counts in a counting period which equals the production period from 0 to r is

S= f 4S = Rpl-erii>\ f If exp[-X(r" - 01 Pit" - t') åt' \dt" . (9) o o ' b i

If Rc is plotted as a function of time the curve would qualitatively look like the one in fig. 1. After a time tz the counting rate Rc becomes approximately constant and remains constant throughout the counting period

TIME

1 ig. 1. Qualitative picture of the counting rate, Rc, as a function of the time from start of the counting, indicating the time tz wlierc åRJit < z.

(assuming that Rp is constant during the counting period). The time /, is determined by the half-life of the nu- clide and the shape of the delay probability curve, and is defined mathematically by t t {R'c'-R c-)[R (?

where the value of z may be taken as small as desirable. This is equivalent to (for /" > tz) t" <2 f exp[-A(r" t')\ p(r" - /') it' - f exp[-A(/2 - /)] p{tz - t') it' o b ...... s^, ^ (. 1 1 >

J exp[-\(t" - t')\p(t" - t')it'

If an analytical expression for p(t) is available, tz can be found by numerical integration of the two integerals. Likewise, the constant value of the integral on the right hand side of the inequality (12), C, can be calculated. If the counting is started a time tz after start of production, the following equation for the totally recorded number of counts is valid: (13)

where tc is the counting time. Then

Rp=SI$eni>\Ctc. (14)

An approximate value of tz may be found by experimentally recording a curve like fig. 1. 3.2. Determination of fission cross sections

Knowing the amount of target material (N is number of atoms) and the flux tø) of the bombarding particles, the cross section (a) is defined by

RP=oN, (IS) 378 T. Bjornstad / Fission yield measurements

Aqueous solution

Detector

Aqueous solution

Fig. 2. Example of an arbitrary flow diagram for the chemical separation system. Hie arrows indicate the flow directions of the liquids and gases. Cl, ..., C4 are units consisting of one static mixer for thorough mixing of the aqueous and organic phases, and one H

o = Rp/d>N. (15a)

In cases with unknown

Oxlos = {Rp)xl(Rp)s, (16)

°x=\(Rp)xWPh\°s. (16a) Two cases are considered: l)x and s isotopes, then the efficiencies r? and v are the same for the two nuclides, and so are the delay prob- ability functions p(t). By introducing eq. (14) and letting (tc)x = (/c)s one gets

(17)

II) x and s non-isotropic:

Ox=-. (18)

Here the constants Cx and Cs contain different delay probability functions. The two nuclides are then sepa- rated and measured in the same experiment, e.g. according to the schematic separation scheme in fig. 2. This approach requires the evaluation of a delay probability function for two elements.

3.3. Determination of efficiency parameters of the system

The recoil yield $ depends on parameters like the mass, ionic and nuclear charge of the recoiling nucleus, the absorber material type and thickness and the energy of the bombarding particle. The value for £ for a nuclide with unknown formation cross section can in general probably be estimated to better than 30% by extrapolation T. Bjornstad I Fission yield measurements 379

from known or measurable lvalues for nuclides in the neighbourhood to the nuclide of interest (a corresponding error will then, of course, be introduced in the final result for the reaction cross section). However, in fission, for instance induced by thermal neutrons in 235U, the recoil ranges vary in aveiage from 6 mg/cm2 to 13 mg/cm2 for mass numbers from 156 to 77 [17]. A typical target in our experiments at the TRIGA reactor in Mainz contains from 0.1 mg to 0.5 mg, corresponding to 0.127 mg/cm2 and 0.637 mg/cm2, respectively, which is a factor 10- 100 lower than the ranges. Accordingly, the recoil efficiency £ is close to 50% (the other half recoiling into the target backing) for these irradiations, and there is no need to worry about uncertain estimates. For the sake of generality £ is, however, all the way kept as a parameter in the formalism.

So far there is no evidence that the 17-values are element dependent when using the N2/C2H4-jet. Accordingly the GJRT efficiency can be found by measuring the activity of a suitable nuclide on a catcher foil mounted close to the target relative to the activity of the nuclide collected at the exit end of the capillary. Of course the chemical yield (separation efficiency) v is element dependent. In addition, for a given chemical separation scheme, v is sensitive for variations in temperature and liquid flow rates. It is therefore essential to keep constant running conditions of the system during the measurements. The efficiency v is composed of the chemical yields of each of the separation steps in the system •

where vllG expresses the fraction of the cluster-bound nuclides in the jet gas that is dissolved in the liquid, the remaining part being swept off together with the jet gas. By measuring the counting rate of the nuclide emerging

from the end of the capillary, Rcl, and the counting rate of the nuclide in the two liquid phases after the first centrifuge, /?()rgl and Raql, the efficiency fDt; can be calculated by (20)

The extraction efficiencies vci, fc-2.... are found by measuring the counting rate of the nuclide in both phases after extraction, Rori and Raq. For the extraction from the aqueous into the organic phase the efficiency is expressed by

"c:i=«<*gi/(«orgi+^i). (21) and for the back-extraction from the organic into the aqueous phase we have

"CZ ~ ^3i|2/(^»rg2 + ^aq2) = ^ai^/^orgl. (22) and so on throughout the extraction steps. In cases where difficulties may arise in measuring JJ and v separately

the product i\t> may be obtained assuming that Rp, |, e, X and C are known for a long-lived isotope of the ele- ment in question. According to eq. (13) Si - 5, TIV —

where 5, and S2 are the totally recorded number of counts after the counting times tj and t2 respectively (start counting at a time t0 > tz). The total counting efficiency e is composed of two parts - the absolute intensity of the detected characteristic radiation, /(abs)i and a part including the counting geometry and the internal detector efficiency, ep. e = '(»bS)e[>. (24) In summary the following measurements are necessary for evaluation of a fission cross section: Determination of a delay probability function, determination of the GJRT efficiency and the chemical yield of the separation system and the determination of the counting efficiency of the detection system. The following sections show how to determine the delay probability curve, and the fitting of an analytical function to the curve. 380 T. Bjonistad / lusshti yield measurements

4. Methods to determine the delay probability curve

4.1. Previous work

Various methods may be used for determination of transport times in on-line systems. Grover [18] has modi- fied some of the formulas describing molecular flow to fit the transport of rapidly decaying radioactive gases. Oron and Amiel [19] and Winsberg [20] have described a method used at the separator SOLIS [21] in order to determine mass transport times. By means of a fast neutron beam shutter the irradiation time of the target is well defined, and the growth of the activity on a stationary collector during irradiation and the decay of the same activity after irradiation are measured. From the resulting data the delay properties of the system expressed by the delay probability function p{t) may be found as shown in ref 20. Rudstam [22—24] has carried out thorough theoretical investigations of the delay properties of on-line mass separation systems, taking into consideration various transport mechanisms like diffusion and desorption processes. Hagebo et al. [25] and Ravn et al. [26] describe an experimental method for determination of the integrated delay function applicable at the ISOLDE facility [27,28]. After reaching saturation in the production of a nuclide at the collector position, the proton beam is deflected off the target and the activity of consecutive samples measured using sampling times much shorter than the half-life of the nuclide collected. Ref. 26 describes how to derive the delay probability function for certain elements released from molten targets using this approach.

4.2. Delay measurements in the combined OJRT-SISAK system

As shortly mentioned in the introduction a GJRT-SISAK system is at present installed at the Mainz TR1GA reactor. This reactor can be operated in the pulsed mode with a pulse width of ~-30 ms (fwhm) and a corre- sponding neutron flux of ~5 X 10'3 n • cm"2 'S'1. Thus an activity, well defined in time, can be produced in the target. This pulse starts the counting equipment, and the activity at the measuring position is recorded as a function of time either by multianalysis or multiscaling. In a typical experiment counting intervals as short as 0.5 s are necessary in order to allow the construction of a delay curve. In such short periods, too poor statistics are obtained by gating at single y-ray peaks. Accordingly gross 7-ray counting is applied. In order to reveal the "true" shape of the delay curve, measured values must be corrected for dead-time and decay, qualitatively illus- trated iii fig. 3. Correction methods will be discussed in the following.

4.3. Corrections to the recorded delay curve

4.3.1. Dead-time correction

The number of counts 5" and the effective counting time r are iccorded simultaneously by the electronic equip- ment during the 0.5 s clock-time counting intervals. The counting rate is assumed constant within one interval. The dead-time corrected number of counts is then s'= 5' 0.5/T. (25)

• L/ncorr#ct«d -- - Corracted for diadtimt - -Corrected for a dcadttmt and decay COUN T

TIME Kip. 3. Qualitative illustration of the mass delay curve indicating the effect of the necessary corrections that have 10 he applied 10 the measured curve. T. Bjornstad / Fission yield measurements 381

4.3.2. Decay correction When applying gross 7-ray counting the activity is normally composed of several decaying nuclides. It is there- fore essential to record a decay curve of the activity used. The dead-time corrected delay curve must then be cor- rected according to this decay curve. The decay corrections may be done either semigrapliically or computerized. Below a few procedures are out- lined for the simple case where growing-in effects can be neglected. It is practical to define three index labels: (' is the time index starting from the first point in the peak, / defines the nuclide, and p defines the experimental point number (p = 1 is the first one). Fig. 4 illustrates the definition of the indexes;' and p. The decay measurements are carried out by first letting an activity pulse pass the actual chemical separation. At a well defined time the liquid flow is stopped, and gross 7-ray multiscaling measurements performed on a liquid volume element. The counting intervals are preferably identical to those used for recording the delay curve. As to the exact starting point of the measurements, two cases must be separately considered: Case 1. It is desirable that the midpoint in the first counting interval corresponds to / = 1 in the delay curve. But the countrate at / = ] in the normal counting position is ~0. In order to assure sufficient activity for the measure- ment of the decay, the detection position must therefore be moved "closer" to the separation apparatus. How- ever, the counting is started at the time corresponding to i = 1. The recorded decay curve is then qualitatively sim- ilar to the one illustrated in fig. 5a. Case 2. If practical hinderances make the method in case 1 inapplicable, the decay measurements can be done in the normal counting position, and started at a time corresponding, for instance, to the delay curve maximum (at / = 5 in fig. 4). Die resulting decay curve will then qualitatively look like fig. 5b. When extrapolating the curve back to / = 1, care must be taken for possible short-lived activity contributing significantly only to the first part of the decay curve.

Decay curve , "case 1"

i=1 3 5 7 9 II 13 15 T7 19 TIME LABEL

b.

Dec. O Occay corrected UJ l- <

O 3 O Cl

. I 3 5 7 9 II 13 15 17 19 TIME LABEL TIME LABEL lip. 4. Qualitative picture of the mass delay curve illustrating the definition of the indices; and p. lig. 5. Qualitative illustration of the decay of the activity used in the measurement of itie mass delay curve, (a) Start measure- ment at / = 1 (case 1). (b) Start measurement at / = 5 (case 2). 382 T. Bjornstad I Fission yield measurements

4.3.2.1. Semigraphical correction procedure Case 1. The measured counting rate (cp 0.5 s) in the points 2, 3 4, ..., p of the delay curve are denoted Rl,R\, R\, ..., Rf. The same counting rates corrected back to / = 1 are R\, R\, R\ ..., R^. The counting rates for the points in the decay curve corresponding to the times labelled 1 ... i are denoted /?? ...Rf. Then the following rela- tion is valid:

R^^Rf. (26, «<• The /?d-values are taken directly from the decay curve, and the corrected delay curve is readily derived. Case 2. Suppose that short-lived activity which would contribute mainly to the first points in the delay curve (p < 5 in the example) is absent. Then the extrapolation of the decay curve back to / = 1 may be done rather safely, and eq. (26) can be applied.

4.3.2.2. Computerized correction procedures Case 1. Let rjj- symbolize the counting rate of nuclide / at the time labeled i and related to the point p in a delay curve which is smoothed with an analytical function. The measured total counting rate at any time labelled i, is then expressed as /

/Jfs/ff=if,+if!1 + ... + ffy=£/f/, (27) where Rf and Rf are the experimental and fitted total counting rates respectively. This expression may be written as /

Rf =tfi exp(-X, tj) + r?2 exp(-X2f,-) + ... + rf,- exp(-X/f,) = £ rf/ exp(--Xyf,-). (28) /-i By exchanging the index p with the index d, this expression represents the decay curve. A computer fit of eq. (28) to the decay curve gives the best values of the parameters rf, ... /-^ and X, ... Xy. The corrected counting rate to be found is expressed by /'

/?? = r?, + r?2+ ...+/-£= S/I/. (29)

From the eqs. (28) and (29) and the following relations valid for all/,

the following expression forÆ? is derived:

R'1= — Rf , (31) i S Ai

which is equivalent to

Ri'--éRf • (.Ma)

In botli these equations the parameter values on the right hand side are known for all p and /'. Case 2. Let x be the time label of the first experimentally recorded point (/ = x) in the decay curve (x = 5 in Tig. 5b). By exchanging the index 1 with x in the cqs. (28), (29) and (30), and the time tt by f/-f.v, and taking into T. lijornstad / Fission yield nicasurcnwHts 383 account that

'•'\/ = (32) an expression identical to eq. (31) can be derived. Note that all the decaying components contributing to the recorded delay-curve are here supposed to be clearly detected at the time labelled x.

5. Examples of experimental determination of delay probability curves

A number of delay curves has been determined experimentally for various chemical systems with variation of parameters like the measuring position, speed of the different liquids, length of tubes and the shape and volume of the measuring cell. As an example delay curves measured by short-lived isotopes of lanthanum isolated from fission products of 235U are presented below. The chemical separation scheme is described in ref. 29. To facilitate the understanding, the procedure is outlined step by step: a) The delay curve is measured in the following way: The counting cell is a tank volume of —100 ml. A plug flow (see section 6.1) delay line with a volume of 28.3 ml, corresponding to 4.3 s delay, is inserted in front of the cell. The whole experimental sequence is controlled by an electronic timer, and the time sequence for the various events is shown in fig. 6. The data are accumulated in the gross 7-ray multiscaling mode in the energy region 100-1000 keV, and with counting intervals of 0.5 s; b) The main contributors to the number of counts in the delay curve are known to be '"""La (42.1 s) and I46La (8.5 s) with a small contribution also from I4sLa (25.2 s) [29]. Accordingly, the dead-time corrected counting number also have to be corrected for decay. The decay curve is accumulated according to case 1. The delay line is disconnected (which in practice means that the detector is closer to the separation apparatus). When the decay measurements start, the liquid has to be at rest. This is achieved by a valve operated by the electronic timer. The measuremenys is started at a time corresponding to the first channel in the delay peak. The whole time sequence is shown in fig. 6. c) By applying the decay data in the semigraphical correction procedure described in section 4.3.2.1 the cor- rect form of the delay curve is found, and illustrated in fig. 7. The results from a similar measurement, keeping all the parameters constant except for the counting ceil wluch is changed into a short piece of tubing (5 cm, plug flow cell, see section 6.1), approximately a "volume ele- ment source", are given in fig. 8.

MASS DELAY 4.3s Start of sequence CURVE Start measurement

I n-pulse Delay 4.3 s -n—T- A 0 0.1 0.5 18.1 DECAY CURVE Start of sequence Start measure- ment

n-pulse Liquid flow .' off

177*6 "Te.i TIME AXIS(s) I ip. 6. The time sequence for the various events when accumulating a mass delay curve, and the corresponding decay curve. 384 T. Bjornstad / Fission yield measurements

1UUU —r 1 • • • 1 1 " ' ' 1 ' ' ' ' 1 ' ' ' » 1 ' ••* ' T ' r-i-T-T—

. • : Decay cur,* of La isotop ••s the nlass delay curve

5000

• """ "*»•*•«»...

0 - -

a : Uncorricltd tipa'trr •ntal rnoii . "o. delay cor« ti. -

0 ° ' Dtcaycorreclid »xptnmen 0000 _

"o S* 0.

• S 000 " 8

n. "°°0o e °o0;

1 t TtM£ (s)

Fig. 7. Mass delay curve (and the corresponding decay curve) measured on a tank-volume cell (-100 ml) by short-lived La iso- topes isolated from thermal n-induccd fission products of 23S V. The dashed curves are "hand-fits" to the experimental points.

i-'ig. 8. Mass delay curve (and the corresponding decay curve) measured directly on a small piece of tubing (-0.75 ml) by short- lived La isotopes, thus approximating a volume element source (see also text to fig.7) .

6. Mathematical description of the delay curve

6.1. Comments on the form of the measured delay curves

Two idealized flow patterns, plug flow and back-mix flow, are of particular interest. Plug flow assumes that a volume element of the fluid moves through the vessel with no overtaking or mixing with earlier or later entering volume elements. Back-mix flow assumes that the fluid is perfectly mixed and uniform in composition through- out the vessel. True laminar flow is present when there is no mixing between two adjacent infinitesimal fluid layers in the radial direction of the tube. The flow velocity varies and is parabolic in its form with the highest velocity in the tube middle and zero velocity at the tube walls. All patterns of flow other than plug and back-mix flow may be called non-ideal flow patterns [30J. In process equipment, even with proper design, some extent of non-ideal flow remains due both to molecular and turbulent diffusion and to the viscous characteristics of real fluids which result in velocity distributions. The dimensionless Reynolds' number [31] nay be used to predict the now velocity profile in a pipe. This number depends upon the four parameters density and viscosity of the fluid, linear velocity of the flow and diameter of the lube. Plug flow is obtained when Reynolds' number >230O. For Reynolds' number <23OO there might be incomplete mix- ing of the fluid in the radial direction, and laminar layers near the tube walls may occur. True plug flow In tubes yields nearly Gaussian distribution curves [32]. Back-mix flow can be obtained in a vessel with thorough mixing, and yields a distribution curve that decreases exponentially in time [30). Laminar parts in the fluid will bring about a slower decrease in the distribution curve than expected front the two flow patterns mentioned above mainly due to the slow exchange of fluid elements between the laminar and turbulent (plug) flow regions [32], The chemical separation equipment consists of several different units, each of them influencing the mass flow. T. BjornstaJ / Fission yield measurements 385

The 30 ms broad (fwhin) neutron pulse will not contribute much to the broadening of the delay peak. However, the reaction chamber [ 15J may be regarded partly as a back-mix vessel. The glas flow in the capillary is presum- ably laminar. In the degassing unit the liquid is swirling around on the glass walls and a certain degree of laminar flow is created also here. The flow pattern in the static mixers may be composed of both laminar, back-mix and plug flow regions. Aronsson [ 12] has shown that the flow pattern in the centrifuges is composed of back-mix and plug flow regions. With the liquid speeds normally used in the SISAK system, plug flow is attained in the tubes. Essentially fig. 8 reflects the influence of the entire production/separation system on the mass distribution, while the extra broadening and increase in screwness as documented in fig. 7 is due exclusively to the counting cell which must be regarded as a back-mix volume. In summary one can say that the average flow pattern in the combined GJRT-SISAK system is not ideal, and that the deviation from pure Gaussian distributed delay curves is ascribed to well recognized back-mix and lami- nar flow regions.

6.2. Fitting of an analytical function to the measured delay curve

A mathematical model of the mass flow behaviour in the SISAK-system, containing exclusively physical parameters, would be valuable with respect to predicting the form of the delay probability curve for a given experimental setup and selected running conditions. Evaluation of such a model would however require much effort in a direction that is outside the frame of tliis work. In order to evaluate the double integral in eq. (9) by computerized numerical integration, it is sufficient that p(') can be described by an analytical function. Procedure 1. The measured delay probability curve may be represented by a step function (fig. 9) p(f,---»•/; ,) = ^,forf,->/>r/_1, (33) where qt is a constant for each step number i. The step length on the time axis, tt — r,-_ j .can be varied over the time axis so as to make the difference between the integral of the step function and the measured curve as small as desirable. Procedure 2. More convenient would it be to represent the curve by a continuous analytical function. The curve shape suggests a Gaussian distribution with some tail function superimposed on the right side. Tail functions nor- mally used in fitting 7-ray peaks [33,34] have been tried with small modifications, using the minimization pro- gram MINU1T [35]. However, the fits obtained were not satisfying, especially in the tail region. A substantial im- provement is obtained when exchanging the Gaussian part with a log-normal distribution [36]. The tail excess can be properly fitted with a function composed of a "growing-in" term multiplied by a "decay" term. The total composite function finally used in the fitting procedure of the measured delay probability curves is:

1 2 f In 2 It- r0 p - 1 \~} [>(t)=yoexp - —• p ln|— + lM +A.,A.'2{1 •-exp [-X2(/-- r2)]} exp[-X,(/ - t2)] . (34)

= where the parameters to be fitted arey0 the ordinate value at the peak maximum.p = (t2 - to)l(to - t,), a mea-

O a.

TIME (t) Y'lf.. 9. Qualitative illustration or 11 step-function fit to the delay curve. 386 T. Bjornstad / Fission yield measurements

1 i — 1000 - o o : Experimental f values • 1 o\ : Computer fit with eqn.(36)

- 500- \

w * U5 r z 5; sooo - LU

• \ Os-o o i : . J...... ?7> . 20 t, t0 '2 30 0 TIME(s) Hg. 10. The same curve as given in fig. 7 with the computer fit (fully drawn curve) based on eq. (34). The parameter values found

are>'o = 12857,p = 3.58, W = 36.49, f0 = 29.54,K2 = 0, \, = x2 = 0.

Fig. 11. The same curve as given in fig. 8 with the computer fit (fully drawn curve) based on eq. (34). The best parameter values

found arey0 = 938, p = 1.42, W = 5.40, f0 = 24.68, K2 = 970, \, = 0.247, \2 = 0.085.

sure of the peak skewness, W = t2 - tu the full with at half-maximum, t0 = the time at the peak maximum, A'2, X and X2 = parameters for the tail function.

Here K, = 0 for t < t2 and 1 for t > f2. The times f0) 11 and t2 are defined in fig. 11. Initial values of v(>, P, H' and t2 are easily found from a plot of the delay curve. Figs. 10 and 11 show the measured delay probability curves from figs. 7 and 8 respectively, properly corrected, and the fits to these curves with eq. (34). The fits are acceptable. Accordingly, by inserting the expression for pit) given in eq. (34) in the normalized form [by

demanding that °°jop(t)åt = 1 ], eq. (9) can be solved by numerical integration.

7. Concluding remarks

The fission yield measurement method outlined here in the present article is, in principle, general with respect to the nuclide under study, type of fissile material and bombarding particle. It can even be employed in high energy spallation and fragmentation reactions provided that the kinetic energy given to the products is high enough for a substantial portion to recoil out of the target. As mentioned above, the emphasize has been put on a relatively detailed description merely of the principles of the method by assuming all the way that parent effects can be neglected. This restricts the presently derived formalism (section 3) to be applicable solely on the following types of measurements: a) Determination of primary reaction yields of nuclides shielded by a (3-. Examples may be 86mRB (1.02 min) and llOgAg (24.6 s). b) Determination of primary reaction yields of nuclides shielded by a "long-lived" nuclide (the decay yield of the "long-lived" nuclide during the time from production to separation from the nuclide of interest must be negli- gible). Examples of nuciides in this category are 97mNb (53 s) and 10SmRh (45 s). c) Determination of cumulative reaction yields for nuclides whose parents are short-lived enough to be consid- ered totally decayed at the production site. Examples are 131>l32Sn (59 s and 40 s). Extending the method to include also parent effects during the mass transfer is possible by using the necessary growing-in functions in the mathematical formalism, in a treatment parallel to the one given by Rudslam (24J, but adjusted to fit the present experimental technique. 3ut since it is hardly feasible to construct a simple, gen- eral formalism, this work should be carried out for each actual experiment that might come up. T. Bjonistad I h'ission yield measurements 387

The data shown in the figs. 7, 8, 10 and 11 are taken from a joint SISAK experiment, and the permission to use the data from my collaborators, Drs. K. Brodén, N. Kaffrell, I. Rudstad-Haldorsen, G. Skamemark, E. Stcncler and N. Trautmann, is kindly acknowledged. Special thanks are due to Dr. M. Skarestad for having critically read the manuscript, for good advice and fruitful discussions. Financial support has been obtained from the Norwegian Research Council for Science and the Huaminitics.

References

111 O. Halin and I . Strassmann, Naturwiss. 27 (1939) 11. |2| J. Åysto, P. Puumalainen and K. Valli, Nucl. Instr. and Meth. 115 (1974) 65. [3] J. Aystii, S. Hilkbrand, K.ll. Hellmuth and K. Valli, Nucl. Instr. and Meth. 120 (1974) 163. [4] K.L. Kosanke, M.D. Ldmiston, R.A. Warner, R.B. Mrestonc and Wm.C. McHarris, Nucl. Instr. and Mcth. 124 (1975) 365. [5] H. Jungdas, H.G. Wilhelm, W. Westmeier, W. Kornahl, Y. Tares, D. Molzahn, R. Brandt and H. Wollnik. Nucl. Instr. and Mcth. 137 (1976)93. [6J H. Wollnik, Nucl. Instr. and Meth. 139 (1976) 311. [7] R.J. Silva, N. Trautmann, M. Zendel, P.I'. Dittner, E. Stender and H. Alirens, Nucl. Instr. and Meth. 147 (1977) 371. |8] H. Reinhardt and J. Rydberg, Acta Chem. Scand. 23 (1969) 2773. [9| J. Rydberg. H. Persson, P.O. Aronsson, A. Selme and G. Skarnemark, Hydrometallurgy 5 (1980) 273. 1101 J. Rydberg, G. Reinhardt and J.O. Liljenzin, Ion Exchange Solv. Extr. 3 (1973) 111. 1111 P.O. Aronsson. U.K. Johansson, .1. Rydberg, G. Skarnemark, J. Alstad, B. Hergersen, I.. Kvåle and M. Skarestad, J. Inor^. Nucl. Chem. 36 (1974) 2397. [12] P.O. Aronsson. Thesis, Department of Nuclear Chemistry, Chalmers University of Technology, Golebore (1974). [ 131 G. Skarncmark, Thesis, Department of Nuclear Chemistry, Chalmers University of Technology, Gbteborg (1977). 114] G. Skarnemark, P.O. Arunsson, K. Brodén, J. Rydberg, T. Bjornstad, N. Kaffrell, K. Stender and N. Trautmann, Nuci. Instr. and Mcth. 171 (1980) 323. (151 N. Trautmann, P.O. Aronsson, T. Bjornstad, N. Kaffrell, E. Kvalc, M. Skarcstad, G. Skarnemark and K. Stender, Jnorg. Nucl. Clicm. Lett. 11 (1975) 729. (161 L. Winsberg. Nucl. Instr. and Meth. 95 (1971) 23. [17| B.C. Harvey, Ann. Rev. Nucl. Sci. 10 (1960) 235. [18| J.R. Grover, J. Inorg. Nucl. Chen). 31 (1969)369. [ 191 M. Oron and S. Aniiel, Proc. Int. Conf. on Klectromagnctic isotope separators and the techniques of their applications, Publ. BMBW-l'B K70-28, eds. H. Wagner and W. Walcher (Bundesministerium lur Bildung und Wissenschaft, Marburg. 1970) p. 87. [20] L. Winsberg, Nucl. Instr. and Meth. 95 (1971) 19. [21 ] S. Amicl, Ark. l-ysik 36 (1967) 9. [22 J G. Rudstam, Nucl. Instr. and Meth. 38 (1965) 282. [23 ] G. Rudstam, CERN YeUow Report 70-3 (1970) p. 123. 1241 G. Rudstam, Research report L1--65 (1975). [25) E. Hagebii, A. Kjelberg, P. Patzclt, G. Rudstam and S. Sundell, CERN Yellow Report 70-3 (1970) p. 93. [26] H.L. Ravn, S. Sundell, L. Westgaard and F.. Roeckl, J. Inorg. Nucl. Chem. 37 (1975) 383. |27) A. Kjelberg and G. Rudstam (eds.), CliRN Yellow Report 70-3 (1970). [28| H.L. Ravn, L.C. Carraz, J. Dcnimal, K. Kugler, M. Skarestad, S. Sundell and L. V/u^piard, Nucl. Instr. and Mctli. 139 (1976) 267. [29] G. Skarnemark, P.O. Aronsson, T. Bjornstad, li. Kvale, N. Kaffrell, E. Stender and N. Trautmann, J. Inorg. Nucl. Chem. 39 (1977) 1487. [301 O. Lcvenspiel and K.B. Bischoff, in Advances in chemical engineering, eds. T.B. Drew, J.W. Hoopcs, Jr. and T. Vermeulen, Vol. 4 (Academic Press. New York, 1963) p. 95. [31 ] J.M. Key, Introduction to fluid mechanics and heat transfer, 2nd ed. (Cambridge University Press, Cambridge, 1963) p. 65. [32] G. Taylor, Proc. Roy. Soc. A223 (1954) 446. [331 K.I:. Brockmann, Thesis, Institute of Physics, University of Oslo (1974). |34| J. Kern, Nucl. Instr. and MelJi. 79 (1970) 233. [351 I'. James and H. Roos.CKRN Computer, 6000 Series Program Library, Long Write-up 1)506, D516 (1971). |36| D.H. Siano. J. Chem. Education 49 (1972) 755. PAPER P.IX, Nuclear Instruments and Methods 186 (1981) 391-400 North-Holland Publishing Company

NEW TARGETS FOR ON-LINE MASS SEPARATION OF NUCLEI FORMED IN 600 MeV PROTON AND 910 MeV 3He REACTIONS

T. BJØRNSTAD1, L.C. CARRAZ3, H.Å. GUSTAFSSON2, J. HEINEMEIER3, B. JONSON2, O.C. JONSSON2, V. LINDFORS2, S. MATTSSON2 and H.L. RAVN2 1 Department of Nuclear Chemistry, University of Oslo, Norway 1 The ISOLDE Collaboration, CERN, Geneva, Switzerland 1 Institute of Physics, University of Aarhus, Denmark

The performance of the previously reported UCVgraphite cloth. VC. La and Pb targets, as well as (he newly developed ThCrgraphite, UCrgraphite, ScC^-graphite. and CaBf, targets is discussed. Production-yield curves for the elements Li, Na, Ar, K. Ga, Rb, In, Cs, Hg, XI and Fr are given for 60() MeV protons and «10 MeV 'He. It is shown that the latter projectile gives higher yields for the neutron-deficient isotopes produced in fragmentation and spallation processes.

1. Introduction also from irradiations with a 910 MeV 3He beam. Finally, in section 4, we summarize the present In recent years great progress has been made status and mention some possibilities for further in the fast-target techniques for the production developments. of intense and pure beams of mass-separated radioactivity. Especially, the development of high-temperature targets (1700-2200°C) has 2. Experimental given the field a great push forward and permit- ted the study of nuclides with halflives down to 2.1. Design of the target and ion-source unit milliseconds. Typical examples of these targets are the refractory metal powder targets such as The ISOLDE target unit is described in detail Nb, Ta and Hf, and also the refractory carbides elsewhere [3,4]. It consists of a target material such as VC, UC2-graphite cloth and LaC2- container which can be heated to 2400°C. The graphite cloth [1.2]. evaporated nuclear reaction products are al- The high temperature and the short solid-state lowed to diffuse to the ion source via a transfer diffusion path make it possible to achieve a fast tube, which may be either heated or cooled. A release of the reaction products which is decisive high chemical selectivity is obtained by combin- for the yields of short-lived nuclides. The ing different target materials with positive sur- production rates may become very high since the face ionization, negative surface ionization, or long range of the relativistic bombarding parti- plasma discharge ionization. cles allows the use of thick targets. As one link in a long chain of related articles, this paper deals 2.2. General target material treatment with new target developments which have led to substantial improvements, both with respect to To ensure good working conditions of the ion increased intensity of the separated beams and to source and the separator, it is important to keep the number of elements available for study. the pressure in the ion source below —lC'Torr. In section 2.3 we give the details of a new This requires high purity of the materials used in method for the preparation of carbide targets. In the entire assembly. Furthermore, a sintering section 3 the results of the survey of the and oulgassin«; procedure which lasts for more produced elements arc discussed in terms of the than 24 h at temperatures of I500-2000°C is isotopic yield distribution as obtained from necessary. This is normally done at an off-line 600 MeV proton reactions and, in most cases. target pump stand, which has quick connectors

0029-554X/8l/IKKH)-0()'K)/$2.50 © North-Holland VI. TAROF.T THCHNIOUKS 392 T. Bjømstad el al. I New targets for on-line mass separation

identical to those of the on-line separator. In reasons a 5 cm target of only 12 g/cm2 of Th was some cases the target material alone is reacted in used in the present experiment. a vacuum oven, which has previously been des- The VC target. Pure VC powder was com- cribed [5]. Temperature calibration of the target pressed into tablets and loaded into a graphite- container as a function of the applied current is lined Ta target container, as described for ThC2 performed simultaneously with a two-colour opti- in the previous section. After outgassing at cal pyrometer through a glass window in the 1700°C, the density of the tablets was measured vacuum chamber. This is a necessary procedure to be 4 g/cm3; this corresponds to 70% of the since a thermocouple connected to a high-tem- specific density of pure VC. The target thickness perature target lasts for only the first 5-6 h of was 62.3 g/cm2 of V.

on-line conditions. The ScCrgraphite target. The ScC2-graphite target was prepared in the same way as the 2.3. Targets ThC2; it had an Se:C molar ratio of 1:3.8, which allowed for an Sc target thickness of 12.8 g/cm2. The UCrgraphite cloth target. The develop- The CaB6 target. Commercially available CaB6 ment of this target is described in detail else- powder was loaded into the graphite-clad Ta where [2, 5, 6]. Here we give only a brief remin- target container and compressed manually to der that a graphite tube, filled with -380 UC2- about 50% of the specific density of pure CaB6. graphite cloth disks of 15 mm diameter, was in- The Ca target thickness obtained was 7.5 g/cm2. serted in a 20 cm long tantalum target container. The Pb and La targets. These two targets, The graphite tube served the purpose of increas- described elsewhere [3], are here referred to only ing the lifetime of the Ta container, which for the comparison of Hg and Cs production 3 otherwise would react with the UC2-graphite rates from He and protons. The target cloth. In recent targets this reaction has been thicknesses were 140 g/cm2 of La and 170 g/cm2 delayed by means of two Ta-foil layers inside the of Pb. Ta container; these prevent direct contact be- tween the UC2-graphite cloth and the container 2 2.4. Beams from the CERN synchrocyclotron wall. The target thicknesses were 13-17g/cm of (SC) U. The ThCrgraphite target. A new general The beams available at ISOLDE from the method for the production of carbide materials CERN SC are ~4/iA 600 Me V protons, was used in the preparation of this target matrix. ~0.5p/iA* 910MeV 3He2*, and -O.IpjtA 15 5 Graphite powder and ThO2 were mixed in a ball 86MeV/u "C\ Other heavy ions such as N *, mill. From this mixture, 2-3 mm thick tablets 16O6+, and ^Ne6* have been accelerated at the with diameters of 14 mm were formed in a com- synchrocyclotron, but their relatively low inten- pacting press which applied a pressure of ~5 sity makes them less interesting at present. Only ton. The tablets were placed in a graphite tube those results obtained by the first two beams and reduced to ThQ in the vacuum oven at mentioned above will be given here. The use of ~ 2000°C. The tablets consist of a solid solution "C* as a production projectile has been rather of ThC2 in excess graphite, and have a high extensively studied, and the results will be pub- porosity due to the voids left after the compac- lished in a forthcoming paper [7]. ting procedure and from those created by the reduction process. The advantage of this pro- 2.5. The production yield measurements cedure over the previously described graphite- cloth impregnation [2] method is that it allows In this article the performance of the different thicker targets to be made within a given con- targets will be presented only by the measured tainer length. The currently used material has a production yields. Their release properties are Th:C molar ratio of 1:3.8; when placed in a discussed in another paper [7]. standard Ta target container,. it allows for Th 2 target thicknesses of ~ 50 g/cm . For practical • I p/iA * 6.24 x I»12 parliclcs/s. V T. Bjømstad et al. I New targets for on-line mass separation 393

The production yield is defined here as the matrices are discussed. The target and ion source measured intensity of the mass-separated beams temperatures as well as the target thicknesses are ',-.. the collector tank. The yields were determined summarized in table 1. by in-beam beta-decay measurements at satura- tion for the very short-lived products. For the 3.1. Li, Na, K, Rb, Cs, and Fr from UCrgraphite long-lived products, collection and successive cloth singles counting or multiscaling of the beta activity were used in order to determine the The production yields for the alkali elements correct component from the complex decay cur- are shown in figs. 1-6. Among the alkalis, Li is ves. The most abundant radioactive ion beams the most difficult to ionize; and for fast release, were measured directly in a Faraday cup. In the temperature of both the target and the ion order to average short-time variation in the source must be high. Even at a line temperature bombarding particle flux and in the separation of 2500°C, the Li-production yield increases and detection efficiency, it was found necessary linearly with the temperature. to determine the intensity ratio between two For the fragmentation products (Li and Na) adjacent masses rather than the absolute value. there is a trend towards higher production yields The method will be outlined more closely in a with 3He than with the protons (figs. 1 and 2). forthcoming paper [7]. This trend appears to be washed out at K (fig. 3), For the various measurements, the ion beam and in the fission product region (Rb and Cs) it is was brought into a low-background counting not present at all (figs. 4 and 5, where for com- area and collected on an aluminized mylar tape parison the isotopic yields from thermal neutron- for transport of the activity. The beta particles induced fission [12] are also shown). The were counted in a thick efficiency-calibrated 4ir measured yields for 'He are here even somewhat particle scintillator. For gamma rays, a 17% lower than those for protons. Ge(Li) detector was used. Charged particles In the region of spallation products (Fr) the were recorded by silicon surface barrier detec- mass distribution with 'He appears flatter than tors. The delayed neutrons were measured with for protons, and the yields for the most neutron- eight 'He proportional counter tubes embedded deficient isotopes have even higher absolute symmetrically in paraffin-wax, yielding a detec- values (fig. 6). This variation is illustrated in fig. tion efficiency of 14%. The P„ values used were 7, where the ratio yfHe)/y(p) is plotted versus taken from other papers [8-11]. the mass number. A substantial gain in produc- Intercalibrated (identical) Faraday cups were tion yield could be achieved for the lightest used to check the beam transmission in the beam isotopes with 'He instead of protons as the line. The yields are calculated by the usual formula: Table I Experimental parameters for production of some elements from UC;-graphite cloth target eae,IB[\ - exp(-A/Mn)][exp(- xA-'[l-exp(-Af „,)], Element Target temp. Ionizer temp. Target cou (°C) amount rc> (g/cnr) where: S = counts in time fox,,«; e,j = detector efficiency; e, = beam line transmission (70-80%); Li 2000 2500 13.6 /B= branching ratio; <„,„ = collection time; Na 1800- 2000 2000-2200 13.6 'dec.y = waiting time; !«,„„, = counting time. K 1800-2000 2000-2200 13.6 Rb 1750-1850 1750-2000 16.4 Cs 1750 2000 16.4 Fr 1800 2000 13.6 3. Results and discussions Ga 2000 2400 15.1 In 2000 2500 15.1 In this section the production yields for the Tl 2000 2500 15.1 various elements from the different target

VI. TARGET TECHNIQUES 394 T. Bjifmstad et al. I New targets for on-line mass separation

1 1 —i r—i—

ciotn ('0 "; g'f n'O . om UC-g' aphite 109 cloth ('0.0 g/crr 2 U)

O -

O

- - j£ ID4

§103

- O 1 ppA w-

103 - O o 1 PM" He f I • i 1 10" - 6 7 8 9 10 11 17 MASS NUMBER Fig. 1. Production yields of Li isotopes from 600 MeV pro- ton-induced and 910 McV 'He!*-induced reaction in UfV graphite cloth. The yield of A - 7 was measured by ion Fig. 2. Production yields of Na isotopes from 600 McV pro- detection, A = 8 by 4vf) counting, and A = 9 and 11 by ton-induced and 910 MeV ^He^-induced reactions in VC:- delayed-neutron counting |8]. graphite cloth. The yields of A = 20. 21. and 25-27 were measured by 4ir/3 counting, A - 24 by y counting, and bombarding projectile, assuming the same beam A = 27-32 by detection of delayed neutrons. The P„ values intensity. The relative effect on the production used are based on the new value for 'Li of (50 ±4)% (8.9]. yields of the two projectiles can be tentatively understood in terms of higher energy deposition The problem with contaminant alkalis is more and transfer of angular momentum by the 'He pronounced in the In production, especially for than by protons. As 124. where Cs (and some Ba) isotopes are produced in relatively large amounts. The yields 3.2. Ga, In and Tl from UCrgraphile clolh for A > 130 are not measured, although nuclear spectroscopy experiments have been successfully As the first ionization potential for Ga, In, and performed on A = 132 [13]. Since Ga and In are Tl is around 6eV, they can be ionized with fission products, they are not expected (in ac- efficiencies of the order of 10% in a tungsten cordance with the results presented above for Rb surface-ionization source. The resulting yield and Cs) to be produced in larger amounts by the curves for Ga, In and Tl are given in figs. 8-10, 3He beam than by the proton beam. For Tl respectively. isotopes, produced in deep spallation, we would With this ion source, Ga can be produced expect the 'He projectiles to be better than the almost radiochemically pure (contamination fac- protons. The experimental results given as the tors <310~3). For the heaviest isotopes (A =s 80), ratio YCHe)/Y(jp) in fig. 11 reveal that the iso- Rb and Sr may appear as contaminants, but the topes with A < 187 do have yields from 'He measurements can always be designed to favour bombardment that are 10-100 times higher than Ga because of the big difference in half-life. the corresponding proton-induced yields. The production yields for A > 78 have not yet been properly measured, but a neutron mass 3.3. TJ and Fr from ThCrgraphite scan, presented in fig. 8, showed activity up to -4 = 85. When comparing U and Th as targets for the T. Bjpmstad et al. I New targets far on-line mass separation 395

1 1 1 1 1 1 1 1 1 1 1 1 ' i 1 i i i i i 111 i 10" Rb trom UC {16 t. g'cm

-A i o8». 1- V o • - /' * * o

10« f I'0' 4 3 io« y V gras - \ 5-0' — *

,03 1 MA p* 3 O 1 p^A He A 4 2 1 H P on Nb-targtt IO - ir From ther mal neutron irraflidtion IO1 oi U

40 45 50 „0 1 i i i i 1 . . LU lill 11, i MASS NUMBER 70 75 90 85 90 95 100 MASS NUMBER Fig. 3. Production yields of K isotopes from 600 MeV pro- 3 2 ton-induced and 910 MeV He *-induced reactions in UC2- Fig. 4. Production yields of Rb isotopes from 600 MeV pro- graphite cloth. The yields of A = 37, 38. and 44-50 were ton-induced and 910 MeV 'Her*-induccd reactions in UQ- measured by 4ir/3 counting. A = 42 and 43 by y counting, graphite cloth. The yields of A = 79, 81, 86. and 90-94 were and A 3* 49 by detection of delayed neutrons, for which the measured by 4vfi counting, and A>93 by detection of f„ values have recently been measured |10|. delayed neutrons. The P„ values for A s 98 111 ] are all normalized to the new P„ value of "Li [8] while those for A = 99, 100, and 102 are estimated. We also give the yields production of Tl and Fr, the latter matrix, as for proton-induced spallation in a Nb-powder target (50 g/cm2) [ 11 and the strongly cumulative yields from thermal shown in figs. 12 and 13, gives higher absolute 1

VI. TARGET TECHNIQUES 396 T. Bjømstad et al. I New targets for on-line mass separation

I ' I I I | I I I I | I I I I | I I I I | I I I I | I I I I | I w'"

CS trom JC - gropMe cloth (16 £ g/cm2 U) 10* Fr Irom UC-graphile cloth (10.0 g/cm^

era' O 5 Q 105 UJ Sid5 >- z'0* O I 103

1 H* P' 1 PPA W I pA p* on La-target From thermal neutron irrad>oliOn cH 10°

i I i i I i 115 120 125 130 135 UO U5 150 I ' i ' ' ' ' ' i ' I i ' ' i I i i r I MASS NUMBER 210 220 MASS NUMBER Fig. 5. Production yields of Cs isotopes from 600 MeV proton-induced and 910 MeV -'He^-induced reactions Fig. 6. Production yields of Fr isotopes from 600 MeV pro- ! in UCrgraphile cloth. The yields of A = 123, 125, and ton-induced and 910 MeV 'He *-induced reactions in UC;- 140-146 were measured by 4ir/3 counting, A — 127, 129. and graphite cloth. The yields of A = 202-213 and 219-221 were 132 by y counting, and A = 146-152 by detection of delayed measured by a counting, and A ^ 222 by 4irp counting. neutrons. The P„ values are estimated, and the yields nor- malized to the f3 measurement at A = 146. Also shown are the yield curve from proton-induced spallation in molten La (120g/cm2) [1] and the strongly cumulative yields from as Fr from UC - graphite cloth thermal neutron-induced fission of V [I2|. • 10 i i F production of K has been attempted from a VC \ target, so a direct comparison between the two i matrices is not possible. \ i •

According to spallation cross-section cal B culations using the formula of Silberberg and # • Tsao [14], the production of neutron-deficient Ar - isotopes would be favoured by a factor of 3 to 10 TI O CL \ when using Sc as a target element instead of V, B assuming normalized conditions. a v B

V Also, a direct comparison of the two matrices YIEL D B B is not possible here, because no attempt has been made to produce Ar from the Sc target. /: JCTIO N S ^" -— *-" r 3.6. Arfrom CaB6 i.

i i . i 1 i • , , i , , i Nuclei with N

iO8 ' 1 2 GQ from UC - graphite , • Tt from UC -graphite cloth (JO g/c m U) " cloth (15) q/cm2 U) in«.

s Neutron mass £10' * O •* y io* > |i03 10? £ O • 1 gA p* D II \f\ o 1 ppA 3He2*

1 . . 1 . . , i I P* 185 190 195 MASS NUMBER Fig. 10. Production yields of Tl isotopes from 600 MeV pro- TO* 2 75 80 ton-induced and 910 MeV 'He *-induced reactions in UQ- MASS NUMBER graphite cloth. The yields of A = 184-186 were measured by a detection, and A ** 185 by the

Tl from UC-graphite cloth ' 1 ' In Ironri UC-graphite cloth (15.1 g/cm2 u)

!10° / " - -*- o 10' - / •'' " "~ -

\ • In, high spin isomer, 1 pA p +

; a In, low spin i orner, 1 yA p * * * Ag. 1 pA p • - i -

, J . . 1 . . . 1 , 110 120 130 MASS NUMBER 10» ISO 185 190 195 200 Fig. 9. Production yields of In isotopes from 600 MeV pro- MASS NUMBER ton-induced reactions in UCz-graphile .cloth, all masses measured by the 4n0 detector. Also shown arc the yields of Fig. II. Ratios of the 'He2* to proton-induced production three Ag isotopes measured simultaneously. yields of Tl isotopes from UC:-graphite cloth.

VI. TARGET TECHNIQUES 398 T. Bjømstad et al. I New targets for on-line mass separation

20

_ TI from ThC; -c (12.4 g/cm 2 Th) _ I o g

o-o' o a 10 - o - 9 P' 38 UJ 7 oob - • _ "6 p , ?•* §5 3 10 -- o' ,*' - o S3 D10* a o lil £ o i • i pA p* • ; / o i ppA 3He2 .:

11 102 : i, ,,, 1 , , 180 185 190 195 200 190 MASS NUMBER MASS NUMBER Fig. 12. Production yields of Tl isotopes from 600 MeV pro- Fig. 14. Ratios of the 'He2* to proton-induced production 3 2 ton-induced and 910 MeV He *-induced reactions in ThC2- yields of Tl isotopes from ThC2-graphite. The curve is con- graphite, with target and ion source temperatures of 2000°C structed from the two dashed lines shown in fig. 12, and and 24O0°C, respectively. The comments in the text of fig. 10 normalized at A = 194. The normalization is done because of are also valid here. uncertainty in the absolute position of the yield curve from 'He2*.

enhancement factor of 40 to 200 for the produc- 1 2 tion of ^Ar as compared with V as a target. A Ar trom VC (54 3 g/cm V) 9 2 study of off-line release, high-temperature 10 - and Ca Bg (7.5 g/cm Ca) stability, and corrosion showed CaB6 to be suit- 10« - o

7 10 - / - t c Fr (rom ThC2-C (12.4 qlcm Th) 6 o |io /O o 8 10 A /

z '• D = 8 - p103 X 1 pA p' a. A 102 o o 1 ppA 3Hf2> - 0 A p' on CaBg-target 10'

10° -

• 205 215 220 io-' liti • 1 1 1 1 1 1 MASS NUMBER 40 MASS NUMBER Fig. 13. Production yields of Fr isotopes from 600 MeV pro- Fig. 15. Production yields of Ar isotopes from 600 MeV Ion-induced reactions in ThCj-graphite wilh a target and ion proton-induced and 910 MeV 'He'* induced reactions in VC. source temperature of 2000°C. The low temperature on the Also shown are the two yield points at A = M and 35 from ion source was used in order to limit the simultaneous proton-induced reactions in CaB„. The yield of A = 32 was ionizafion of the neighbouring element Ra (in addition, see measured by ^-delayed proton detection, and the other the text of fig. 6). masses by 4ir0 detection. T. Bjømstad el al. I New targets for online mass separation 399

10' 1 1 1 ' 1 ' 1 able. Unfortunately, this choice gives less chance •BIO«"- of release, because CaS is a rather stable K trom -c compound at high temperatures. In an oxygen- (12.8 g/cm^ Se) - containing Ca compound, however, we could « 0 hope for release of sulfur as SO or SO2. ;,o7 The on-line experiment revealed some : !O6 o - difficulties due to a rather strong and long-lasting

> 105 outgassing, probably because of impurities in the 0 target material. The operation of the plasma ; 10' 1 ^A p« ion-source became difficult since it was con- 3 0 10 nected to the target via an uncooled line. The 102 i 1 1 35 40 maximum target temperature was 1500-1600°C. MASS NUMBER The two measured yield points of ^Ar and 35Ar Fig. 16. Production yields of K isotopes from 600 MeV pro- plotted in fig. 15 should be considered to ton-induced and 910 MeV 3He2*-induced reactions in ScCi- represent a lower limit. If we normalize to the graphite. The yield of A = 35 was determined by in-beam /} same molar fraction of the target element, the detection, while the others were measured by the 4ir/3 detec- gain in using Ca is a factor of 80 to 90. As tor. A tungsten surface ionization source was used at a temperature of 2000°C. expected, S and P were not seen (Ca3P2 is also quite stable at high temperatures). Chlorine, which we hoped to produce as CaCl+, was not seen, probably because of a too low line tem- perature. J 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 .1 1 1 1 i 1 i i i ._ 1012 Cs from molten La ( 1£O gycm? (_a ) -I J I 1 I [ '<<<|' T I I J 1 • I T | I • I <-\ 10" _ _ o Hg from molten Pb (170 g/cm? Pb) 1010 8°88 - • • 8 10« o° • -

•S 108 _ - o

7 _ O 10 -

O • 106 _ _ o

5 _ 10 _ TIO N 3 _ o i 10* -

3 10 - -

1O2 • I »JA p- _ o 1 PM» 3He2' 1 IO ~ o - IO1 -

10° 1OU -

IIS 120 125 130 135 140 , I i • . i I . . . . I • • i . I MASS NUMBER 180 I«S 190 196 200 205 210 Fig. 17. Production yields of Cs isotopes from 600 MeV MASS NUMBER proton-induced and 910 MeV 'He!*-induced reactions in Fig. 18. Production yields of Hg isotopes from 600 MeV molten La. The yields of A = 118-137 were measured by proton-induced and 910 MeV 'He:*-induced reactions in means of the Faraday cup, A - 116-117 by the 4n/3 detector, molten Pb. The yields of A » 186-205 were measured by and A = 114 by /3 -delayed protons means of the An)} deteclor, and A - 177-185 by a detection.

VI. TARGET TECHNIQUES 400 T. Bjttmstad et al. I New targets for on-line mass separation

3.7. Hg from Pb and Cs from La tion rate can be established for the spallation and fragmentation products. The yield curves shown in figs. 17 and 18 indicate that 3He-induced spallation gives higher References yields for the neutron-deficient products. The effect is less pronounced in these two systems, [1] H.L. Ravn. L.C. Carraz, J. Denimal. E Kugler, M. where the targets are only two protons away Skarcslad, S. Sundell and L. Westgaard, N'jcl. Instr. and from the product. Meth. 139(1976) 267. [2| L.C. Carraz, S. Sundell. H.L. Ravn. M. Skarestad and L. Westgaard. Nucl. Instr. and Meth. 158 (1979) 69. |3| HL. Ravn. S. Sundell and L. Westgaard. Nucl. Instr. 4. Conclusions and outlook and Meth. 123 (1975) 131. |4) H.L. Ravn. Phys. Rep. 54 (1979) 201. All the target materials described here, except [5] L.C. Carraz. I.R. Haldorsen. HL Ravn. M. Skarestad (so far) CaB , have been used in on-line produc- and L. Westward. Nucl. Instr. and Meth. 148 (1978) 217. 6 |6] G. Rudstam. Nucl. Instr. and Meth. 139 (1976) 239 tion of nuclides for studies of nuclear properties. |7] T. Bjomstad. H.-Å. Gustafsson. B. Jonson. O.C. Jons- The potential of these targets to produce several son. V. Lindfors. S. Mattsson. A.M. Poskanzer. H.L other elements is large and has still not yet been Ravn and D Schardt, to be published in Z. Phys. fully examined. For instance, the production of |8| T. Bjornstad. H.-Å. Gustafsson. P.G. Hansen. B. Jon- the halogens Cl, Br, and I from the U target has son. V. Lindfors. S. Mattsson, A.M. Poskanzer and H.L Ravn. submitted to Nucl. Phys. A (1980). become routine [15]. With new developments in |9] E Roeckl, P.F. Ditlner, C. Détraz. R. Klapish. C. the ion-source field and in the chemical Thibault and C. Rigaud. Phys. Rev. C10 (1974) 1181. evaporation technique [3,16], we expect many [10] The ISOLDE Collaboration, to be published in Phys. more elements to become available as pure Lett. beams within the next 3 to 4 years. The oxide- |il| E. Roeckl. P.F. Dittner, R. Klapish, C. Thibaull. C. containing materials are especially interesting Rigaud and R. Prieels. Nucl. Phys. A222 (1974) 621. |12| L. von Rcisky, J. Bonn. S.L. Kaufman, L. Kugler. E.-W. with respect to the evaporation of some refrac- Otten. J.-M. Rodriguez-Giles. K.P.C. Spain and D. tory elements that have volatile oxides. Weskott. Nucl. Instr. and Meth. 172 (1980) 423. One of the new targets developed at ISOLDE 113) T. Bjornstad. L.-E. De Geer, GT. Ewan, P.G. Hansen. is a TaQ-graphite target for the production of B. Jonson. K. Kawade, A. Kerek, W.-D. Lauppe, H. Lawin, S. Mattsson and K. Sislemich. Phys. Lett. 91B neutron-rich isotopes of the light noble gases and (1980) 35. for a number of other elements. On-line tests [14] R. Silberbcrg and C.H. Tsao. Astrophys. J. Suppl. show very promising results, and in some Series No. 220(1), 25 (1973) 315. experiments TaQ will probably replace the U 115] B. Vosicki, T. Bjomstad, L.C. Carraz, J. Heinemeier targets. and H.L. Ravn, these proceedings. [16] P. Hoff, L. Jacobsson, B. Johansson. P. Aagaard. G. The results from using the 910 MeV 'He beam Rudstam and H.U. Zwicky, Nucl. Instr. and Meth 172 are encouraging, and an overall higher produc- (1980)413. PAPER P.X. /. Plus. A - Atoms and Nuclei 303. 227-233 (19811 A Atoms and Nuclei Springer-Verlag I%1

Comparative Yields of Alkali Elements and Thallium from Uranium Irradiated with GeV Protons, 3He and I2C

T. Bjørnstad*, H.-Å. Gustafsson, B. Jonson, O. C. Jonsson, V. Lindfors**, S. Mattsson, A. M. Poskanzer***, H. L. Ravn, and D. Schurdt The ISOLDE Collaboration, CERN, Geneva. Switzerland

Received July 22, 1981

Mass-separated ion beams of the alkali elements Na, K., and Fr, and of the element Tl, were produced by bombarding a uranium target with 600MeV protons, 890 MeV 3He2', and 936MeV I2C4 + . Isotopic production yields are reported. In the case of the I2C beam, these are thick target yields. Absolute cross-sections for the proton beam data were deduced by normalizing the delay-time corrected yield curves to measured cross-sections. For products farthest away from stability, the 3He2< beam generally- gives the highest yields.

1. Introduction

The present work was initiated by the new possibil- pects of the utilization of heavy ions for the pro- ities of accelerating heavy ions at the CERN Syn- duction of rare nuclear species. chro-cyclotron (SC). In addition to the 600 MeV pro- ton and 910 MeV 3He2* beams, the SC is also able to accelerate 12C4t ions to an energy of 86A MeV. 2. Experimental Techniques Since neutron emission is usually favoured over the emission of charged particles, the heavy-ion reac- Data for available projectiles at ISOLDE from the tions may be suitable for the production of a variety CERN 212cm Synchro-cyclotron are summarized in of very neutron-deficient nuclides of Z higher than Table 1. The beam intensities were measured by a the target element. The purpose of the present ex- secondary emission chamber which was calibraied periment was, however, to investigate the production against the reaction 2"Al(X,x«yp)24Na [2]. of nuclei lighter than the target, and to see if the The reaction products separated in the ISOLDE more complex projectiles have higher cross-sections electromagnetic mass separator were brought so as to significantly broaden the isotopic yield dis- through a beam-handling system onto a movable tributions. aluminized mylar tape [3]. The collected activity A survey of comparative yield measurements for the was transferred to a thick, 40rum diameter. An plas- elements Na, K, Tl, and Fr, produced by bombard- tic scintillator [4] where the /i-partieles were count- ment of 12g/em2 uranium targets, was carried out a( ed. The detection efficiency, shown in Hg. 1. was ISOLDE [1]. The 12C ions are completely stopped in such thick targets, and therefore these measure- ments arc thick-target yields. Relative to protons T«Me 1. Summary of beam data at the ISOLDI: target and 3He, the effective target thickness is thus re- Projectile Beam intensity Incident I'nergy loss duced considerably. Here we shall discuss some as- beam energy in target (MeV| (McV| * Present address: Department of Nuclear Chemistry, University .il Oslo. N-1000 Oslo 3. Norway MX) .10 ** Present address: Department of Physics. University of Hel- JHe! * 0.5 S90 :on sinki. SF-00170 Helsinki 17. Finland ut4. a, 9.1ft *** Present address: Lawrence Berkeley Laboratory. Berkeley. C'A 14720. USA " I PM A = I particle microampere = f>.24 x W" panicles s

0.140-219.VH I 0.10.1 «227 SOI.40 228 T. Bjørnstad el al.: Comparative Yields ot Alkali Filements and Thallium

ration yields are thick-target yields normalized to 1.0 1 puA. The proton beam results are shown in Figs. 1 1 1 <> 1 1 2--5. Within the experimental accuracy the isotopic - distributions, shown in these figures, reveal no struc- ture due to odd-even effects. The Fr and Tl yields 0.8 - have their maxima at the neutron-deficient side of / ' p stability as expected for spallation products [10]. LU The yields for Na and K are peaked at the neutron- u rich side, but closer to the stability line in accor- fco.6 - '0 - dance with the fragmentation model [10]. The yields 111 from the 'He and '~C irradiations are presented in Figs, 7-14 as ratios to the proton-induced yields. The presented distributions will be further discussed 1 1 1 1 1 0.4 in Sect. 4.

• y9,max , MeV

Kig. 1. The efficiency ol the An plaslic detector obtained by using 3.2. Delay-Times and Cross-Secliims standard beta-sources The time distribution for the release of Na from a measured by means of standard /(-sources. The 4n UC,-graphite cloth target was determied by measur- plastic scintillalor was also used for the Tl isotopes, ing the intensity of the :5Na beam, by means of a which decay by isomeric transition or electron cap- channeltron detector, as a function of time after ture. In these cases, the absolute yields are estimated switching off the proton beam. The recorded curve to be uncertain by a factor of 2 to 3 owing to the was corrected for decay and general background. varying efficiency of the /(-detector. The relative yields for the different projectiles are, however, not af- fected. For the detection ci x-particles. silicon sur- 1 ' 1 face barrier detectors [5] were used either placed in the beam behind a carbon collector foil or in com- 8 bination with a tape-transport system [6]. The neu- -10 iron-rich nuclides, which are characterized by the Na 10°- emission of /j-delaycd neutrons, were identified with a 4n neutron counter [7] calibrated with a 48g _ o -106 sample of uranium. For a few nuclides the gamma- 2 rays were measured with a 17% Ge(Li) standard io - 0 efficiency detector [7], The observed counting rates in were corrected for decay losses by using the formula o . s o 4 presented in [S] in order to obtain the saturation -10 » 4 yields. To eliminate the effect or short time vari- • o 10" U- o - ations in the bombarding beam intensity and the « 0 separator efficiency, most of the yields were obtained LU 2 as ratios between two adjacent masses. The satu- -10 • • 6 ration yield ratios determined were then normalized w - to one absolute vield measurement for each element. -10° 20 25 30 35- 3. Results I , . . . i . MASS NUMBER

5.1. Mensural Yield a Hy.. 2. Production \ields of Na isotop» I illed c rcles are nonn i/ed saturation uelds I sec Icvll I lie l'„ \ allies used are noim The presented production yields from the proton i/cd in Hie ncn /; mine for "I I of (5(H4|.. |7| I he cro J section scale on the nujit.liam) ;i\is is nortnn li/ed lo (il'i nh and He irradiations are normalized to a beam in- M tensity of I puA and a target thickness of lOg cm2 measured for Na 112| Hits c.ile applies lo tin I2 \ields (open circlesi according lo the lexl. Tile pi ills al masses . of uranium, i.e. (hin-targe( yields. Since Ihc C and 34 are wilhin parentheses. hecaKsi' lite correct (cm for the beam is completely stopped in the target, its satu- daughter activities is not laken inlo account T. iijornsuid et ;tl.: Comparative Yields of Alkali Flemcnts and Th.iihurn 229

• i • • i 1 I • ' • I • • ' I • • I • I ' ' • • I ' • • ' • 10°- 7 -10 • K

2 5. W - a 5 8 O -10 30 9 O S (A - V> 9 10"- 3 * m -10 0 o S *

-101 3 * ?- cr

40 45 50 *- 1 . , . . 1 .

MASS NUMBER

Fit;. 3. Production yields of K isotopes. See caption of Fig. 2. The Pn values used are taken from [9]. The cross-section scale is normalized to 0.35 mb for 44K, assuming the same cross-section MASS NUMBER at the maximum of the yield distribution as (hat measured for Sc Fig. 5. Production yields of Kr isotopes. See caption of 1-ii.v 2. The cross-section scale is normalized to 0.20mh measured for ;i*Fr [14]. The error bars arc from (he uncertainties in the /'-wilui.' ;md the normal temperature variations of the target

s m m m g m g m m m

-106 TI - 1.0 25Na from UC- - • graphite cloth a. - • • a. • • a -10" • UJ 0.1 • -102

UJ >- -10° 185 190 19S | . I . S 10 15 20

MASS NUMBER TIME, s

Hg. 4. Production yields of Tl isotopes. Sec caption of Fig. 2. No Hg. 6. Fractional yield A'(fl defined as the adnit> per sample decuv correction is applied to the points (sec tcxtl. The letters m after beam-off relative to the equilibrium ;icliul> nicasuicd while and c indicate the mctastable and the ground slate. rcspec(i\cl>. henm-cm. for Nil front CC.-graphile clolli at ahoul 2.IX)()'( Hie ! he same detection efficiency as for bcla-particles wus used curve is a fit to the experimental poms {see text I and for variable background due lo the electrons line delay measurements. [II]. show that the same from the decay of :'Na collected close to the de- //-value is roughly applicable also for K and Ir. tector. The data are shown in Fig. 6. The parameter The //-value determined was used to correct the sa- /( for the diffusion mechanism(s) in solids, described turation yields for decay losses in the target in ac- in [il]. was fitted to the experimental points in Fig. cordance with the formalism in [II]. Al tempera- 6. giving a value 0.26s"1. Earlier results from on- tures around 2.(HH)0C. mainly nuclides wilh half- 2.10 T. Bjørnslad el al.: Comparative Yields of Alkali Elements and I hallium lives shorter than 30s are affected, because at 'his higher yield at the neutron-deficient side as well us temperature a rather fast release of products is nor- at the neutron-rich side of the distributions. The iv.'.My observed. The 20 ms isotope 219Fr is a good higher yields are tentatively understood in terms of case for testing the relevance of the diffusion cor- higher energy deposition in the target nucleus. For rection. The delay-corrected yield of 219Fr, shown as production purposes the 3He beam will become an open circle in Fig. 5, falls within a factor o[ 2 of even more attractive in the near future because a the 22OFr yield. This nuclide has about the same beam intensity of the same order as the proton beam formation cross-section but, owing to its long half- is within reach at the SC. An unexpected high yield life, no delay correction is necessary. To increase the was observed for 2lNa with '2C as the projectile. To corrected yield of 219Fr to a proper value, the dif- investigale whether this could be attributed to the fusion time in the target-ion source system must be reaction 12C on carbon in Ihe target, a separate much longer than that found in the present work. experiment with ' 2C on a pure graphite target was For the most short-lived Na isotopes, the saturation performed. The result obtained was in agreement yields were increased by a factor of 10-15 because of with integration of the 30 to 80MeV 12C on I2C delay losses, and for the K isotopes the correspond- data [15. 16], showing that the ''Na is produced ing increase is between 3 and 7. All the Fr isotopes near the end of the range of the beam in the target. are almost unaffected by decay losses, except for This experiment points to the possibility of perform- '"Fr for which the saturation yield was increased ing such reactions in thick targets, which ma> be by a factor of 20. The diffusion time for Tl was not interesting in order to produce neutron-deficient nu- measured for the actual target system, but earlier clei in the region 7. <20 where suitable high-temper- experiments have shown that it is much longer than ature targets for proton and JHe bombardment are for Na. K, and Fr. In this case it was not possible to hard to find. In the I2C on I2C experiment a small apply the /(-value obtained for the alkalis, and there- contribution at masses A > 24 was also observed, fore the Tl yields were not corrected for delay in the originating from reactions with the Ta target con- target system. tainer. The yields depend very strongly on the performance The ratios for the fragmentation product potassium. of the actual target-ion source system and, in order K, are shown in Figs. 9 and 10. The trend of higher to keep the experimental conditions approximately yields from the 'He and 12C irradiations is not as the same, the presented data for p and 3He were pronounced as for Na. but still there is a gain in obtained by using the same target unit. Due to tem- yield at both the neutron-deficient side and the neu- perature differences, the performance of various tar- gets mainly affects the sh'>rt-lived nuclides because 1 ' i • • I ' • • • i of their strong sensitivity to decay losses in the - 100 target. Occasionally 10-100 limes higher yields have been observed for these nuclides. This means that Na precise cross-section measurements are difficult to • perform at an on-line separator such as ISOLDE, but it is very suitable for relative yield measure- ments. It may still be interesting to estimate approx- imate cross-sections far away from stability, since it is very difficult to obtain this information from tech- niques other than on-line measurements. In order to - 10 obtain absolute formation cross-sections, the delay- corrected yields have to be related to cross-sections measured, for instance, by radiochemical methods oc [12-14]. The cross-sections given in Figs. 2, 3, and 5 o are determined to a precision of a factor of 2 to 3, depending on the uncertainties in the delay-lime corrections and the normalization cross-sections. -1 20 25 30 35 t . 4. Discussion MASS NUMBER

For the fragmentation product Na, the yield ratios Fig. 7. Ratios nf the 'Me to (he proton-indueed saturation yields in Figs. 7 and 8 show that both JHc and 12C give a of Na isotopes normalised to the same beam intensity T Biørnstad et al.: Comparative Yields of Alkali Elements and Thallium 231

1 1 • i • ' ' ' I ' I ' ' ' 1 ' ' - 100 -10 Na K • • • • • • • • O •

-1 - 10 - • Si rx Q o UJ > > -01 - 1 40 45 50 20 25 30 35 , I . , . . 1 . . , . 1 ... . . i . . I , I . . . . I MASS NUMBER MASS NUMBER Fig. 10. Ratios of the K isotopes. See caption of Kig. S Kig. 8. Ratios of the ' 2C to the proton-induced saturation yields of Na isotopes normalized to the same beam intensity

I 1 ' I • ' - 10 - K

• - 1 • o I cc o UJ

>• >

- o1 40 45 50 . I . . . . 1 ... MASS NUMBER MASS NUMBER Fig. 9. Ratios of Ihc K isotopes. Sec caption of Kig. 7 Fig. 11. Ratios of the TI isotopes. Sec caption of Kig. 1. Kir the letters m and f, sec caption of Fig. 4 tron-rich side of the distributions. The argument used stead of protons, because the higher total eiiergy above for the Na ratios is also applicable in this transferred to the system favours the evaporation of case to explain the higher yields produced in the many particles. The effect is shown in Fifs. II and 'He and I2C irradiations. 12. The experimental ratio illusirated in Fig. II For the deep spallation product Tl, higher yields are shows that the yields for A < 187 arc 10 to W0 times expected when using 3He or llC as projectiles in- higher from 'He than from protons. When usiii{! 232 T. Bjørnstad el al.: Comparative Yields of Alkali Elements and Thallium

rh ' é 1 rh å ' r-, , rh ' was lo produce Am isotopes by irradiating VC,- - 10Ci * graphite cloth by 12C. which involves a transfer of TI three protons to the target nucleus. In this experi-

• ment a thermal ion source was used. The detection system was optimized to measure both alpha and fission-fragment energies. The most typical charac- teristic of the Am isotopes would be spontaneous • 1 , , | , , , , | , 1 -10 - 1'' "I ' 1 ' ' ' ' 1 ' - 10

1 • Fr

>-

-1 - -1 185 190 195 , I . . , , I , MASS NUMBER

Kig. 12. Ratios of the TI isotopes. See eaption i I-ig. S. Kor Ihc O letters m and g see eaption of lig. 4. _) UJ

IJ C instead of protons as projectiles (he effect is not - 0.1 so pronounced, but still there is a gain in yield, as 200 210 220 230 shown in Fig. 12. The measurements of the Tl iso- • i . . . I . , . . I . . . . i . . . . I . . . . 1 . . . . 1 . topes were not extended to the neutron-rich side of stability because of contamination from the isobaric MASS NUMBER Fr isotopes. Fig. 13. Ratios ni the fr iMiiopcs. Sec cuption of l'ig. 7 In the case of the close spallation product Fr (Figs.

13 and 14). the higher bombarding energy of (he 11 12 . 1 . . . 1 ' ' 1 " • i • • • • l • • 'He and C beams does not favour the yields. This -1 is analogous to the case where the proton energy is increased from medium to high energy, where the Fr cross-section decreases for close spallation products • but increases for deep spallation and fragmentation products [10]. In the fission region the cross-section a. is hardly affected by the higher bombarding energy, >• in agreement with observations for C's and Rb iso- topes produced in 910 McV 'He irradiations [8]. -0.1 Kor the most neutron-deficient nuclides. shown in Fig. 13. there is a gain by using 'He as projectiles, • while nuclidcs closer to stability are not favoured. 2 Because of coniamination from neighbouring masses in the separator, this gives cleaner conditions when studying nuclear properties in the very light Fr iso- topes. The low ratio shown in Fig. 14 is not only a i consequence of the rcaciion mechanism but also an effect of the small effective target thickness for '2C. -0.01 while the proton and 'He yields arc thin-target 200 210 220 230 . i . . . . 1 . . . 1 .... i . > ields. During the ' 2C experiment an attempt was made to MASS NUMBER produce elements heatier than the target. The aim li(>. 14. K;il»"> of i IH- I i IWIMJX» See i.ipli/'» «f I ip X 1 Bjornslad ut al.: Comparative Yields of Alkali Elements and Thallium

fission, but no such events were observed. It is possi- Dam. Ph . Hagberg. 1:.. Jonson. II.: Nuel. lustrum. Methods ble that relatively high-temperature stable Am com- 161.427(19791 und;: 'Aere formed, thus preventing the release of . Bjonistad. T.. Gustafsson. H.-A.. Hansen. P.G.. Jonson. B.. Lindrors. V. Matlsson. S.. Poskan/er. A.M.. Ravn. H.I..: Am from Ihe target. However, some weak alpha Nucl. Phys. A 359. I (19X11 peaks were observed in the expected region of en- Bjornslad. T., Carraz. L.C.. Gustafsson. H.-A.. Hagberg. 1 .. ergy for Am. Assigning these alpha peaks to Am, an Heinemeier. J.. Jo B.. Jonsson. OX".. I.indfor-.. V.. Mall- son. S.. Ravn. H.L A1. Instrum. Methods 1S6. 391 II9S1I estimated upper limit of 30|ib for the production of 4 •pAm could be made, which roughly agrees with . The ISOI.DI-.-C'ollaborauon: The "K beta-strength function. I2 in Proc. 4th Int. Conf. on Nuclei far from Stahilit). HeUuigor. the data obtained by a low-energy C beam [17]. 19S1 [CI-RN «1-09 11981)] p. 387 The present work shows that the results obtained Wolfgang. R.. Baker. I-..W.. Caretto. A.A.. ( umming. .1 ».. with the 3He beam are very encouraging and higher Iriedlander. G.. Hudis. .1 : Phys. Rc\. 103. 394 H95hi production yields are established, especially for the tarra/. I.C.. Sundell. S.. Rain. II.I.. Skarcstad. M.. Wesi- deep spallation and fragmentation products. For gaard. I..: Nucl. lustrum. Methods 15». (>9 11979) Haldorsen. I.. I-.ngelsen. S.. l-.rikson. D.. Hageho. I. Johansen. production purposes the 'He beam will become T.. Pappas. AC : Production of :4Na and ;vMg b\ interac- even more attractive with the planned higher beam tion of ( with 15 in WHlMeV protons. J. Inorg Nucl. C hem intensity. The 12C beam does not seem to offer any (in press) advantage for production of elements lighter than Ravn. H.I.: .1. Inorg. Nucl. Chem. 31. lss;3 |19(,9| the target. The higher energy available does not l.avrukhma. A.I... Rodin. S.: Radioklunnya 2. S3 Il9hii| Conjeaud. M.. Ciary. S.. Harar. S.. Wielec'/ko. J.P.: Nucl. Plus compensate for its low intensity and shorter range A309. 5I5|J97,X) as compared to 'He and protons. Nambrodiri. M.N.. C'hulik. I.T.. Nalowit/. J.B.: Nucl. Pins A 263. 491 (1976) Hyde. H.L.. Perlman. I.. Seaborg. G.T.: The nuclear properties of the heavy elements. Vol. I. p. 350. I-ngle\\ood Cliffs. New References Jersey: Prentice Hall. Inc. 1964

T. Bjørnslad 1. Ravn. HL. Carraz. L.C., Denimal. J.. Kugler. I:.. Skareslad, H.-Å. Guslalsson M.. Sundell, S.. Wi-.igaa.rd. L.: Nucl. lustrum. Methods 139. B. Joiison 267 |I97(>> O.C. Jonsson 2. Tuyn. J.W.M.. Deltenre. K.. Lamberet, C Roubaud. G.: V. Lindfors t I-'RN Internal Report HS-RP TM 80-5 (19801 S. Mallsson 3. Lindahl. A., Nielsen. O.B.. Rasmussen, I.L.: Prnc. Int. C'onf. A.M. Poskan/er on the Properties of Nuclei far from the Region of Beta HL. Ravn Siahilin. Leysin. 1970 [CERN 70-30 (1970)]. Vol. I. p. 331 IX Schardt 4. Ram. H.L.. Sundell. S.. Wcslgaard. L.. Rocekl. I-:.: J. Inorg. The ISOLDE Collaboration NUL-I. ("hem. .17. 383 (19751 CI.RN V Hagberg. I.. Hansen. P.O.. Hardy. J.C.. Jonson. B.. Maltsson. CHI21I Geneva S.. I idemand-Pelersson. P.: Nucl Phys. A 29.1. I 119771 Switzerland PAPER P.XI. CERN 81-09 20 July 1981

A LIQUID SALT TARGET FOR SELECTIVE PRODUCTION OF NEUTRON DEFICIENT ANTIMONY ISOTOPES

AT ISOLDE

O. Gloraset , T. Bjørnstad, E. Hagebø, I.R. Haldorsen and V. Hjaltadottir

Department of Chemistry, University of Oslo, Blindern, Oslo 3, Norway

and S. Sundeli

ISOLDE group, EP-division, CERN, 1211 Geneva 23, Switzerland

Abstract

A target system designed for selec- containing elements with Z>52 are not ideal tive on-line mass separator production of as the volatile reaction products Xe and I neutron deficient antimony isotopes in easily diffuse to the ion source and become high energy proton-induced spallation re- ionized together with the elements of inte- actions is described. rest. ThiSjis for instance shown by Glomset and Hagebø for a cesium-containing target The target material consists of matrix. The interest therefore focussed on telluriumdioxide (TeO7), potassiurachloride matrices containing tellurium as target (K,C12) and lithiumchloride (Li,Cl2) in element. the molar ratios 2 9 : 25 : 46. The mixture constitutes a eutecticum with a melting Since indium, tin and antimony form point at 347+5 °C. volatile halides, halogen containing mater- ials were expected to be possible candidates The material shows acceptablestabi- as target material. Early experiments lity at vacuum conditions (10 ~l"o torr) have tested the release properties of tel- during prolonged operation at 45 0 C. In luriumtetrachloride as a fine powder. vacuum chamber release tests of products However, in spite of promising results from from irradiated samples one finds that 50% vacuum chamber tests, on-line isotope sepa- of the antimony activity has evaporated rator experiments gave an average delay after 2.6 min heating at 42 0 C, while half-time of 18 min for the two elements there is practically no release of other antimony and tin. elements (i.e. tin and indium). Thermo- chromatographic experiments show no deposit The delay in powder matrices is mainly in quartz tubes of the released products governed by the diffusion rate in the solid in the temperature range 430 C to 20 C. and/or the surface desorbtion rate. Higher The products are firstly condensed in a rates are observed at increased tempera- liquid nitrogen trap. tures. But a high vapor pressure of the target matrix itself may lead to consider- Careful tests of the target system able loss of material, which in turn may were carried out at the ISOLDE-off-line affect a proper operation of the ion source. isotope separator in order to estimate the The maximum permissible operating tempera- best overall on-line running conditions. ture for telluriumtetrachloride was 100 °C. Diffusion in liquids is generally much Subsequently the target system was faster than in solids at comparable tempe- tested on-line at ISOLDE with 86 MeV/amu ratures. In order to improve the release C as the bombarding projectile (protons rate, a search has been conducted for a were not available at the actual time). tellurium-containing halide system with Release yield curves and release delay sufficiently low vapor pressure above the times were recorded. At 420 C delay half- melting point to ensure a stable operation times of ^2.6 min were found for antimony. of the mass separator. The preparation No tin and indium were seen. This is in and performance of such a material at off- accordance with results from the vacuum line and on-line conditions is reported in chamber experiments. The production yield the present paper. of, for instance 119-Sb, with 2.5 pA 600 MeV proton bombardment of a normal size- target of -><32g/cm , is estimated to •x.2.5-10 2. Off-line experiments atoms/s. 2.1 Vacuum chamber teats

1. Introduction 2.1.1. Target_materlal_greBaration

Over the last few years a considerable The system finally chosen as subject effort has been devoted to the development for further studies consisted of tellurium- of an ISOL-target for selective production dioxid (ToO,), potassiumchlorlde (K,C1-) of indium, tin and antimony from spallation and lithJumehloride (Ll-Cl.) in the molar reactions induced by high energy protons. ratios 29:25:46 respectively. This mixture The scientific justification is obvious is reported to.have a cutectlc melting point sine? these elements lie in the region of at 347-352 C ', and is prepared in the the magic number Z«50. Target materials following way: After some hours storing of the very hygroscopic lithiumchloride at a) Present address: 110-120 C to remove any trace of humidity, Department for Health Physicss, the three compounds are mixed, and pounded Rikshospitalet, Oslo 1, Norway. - 732 - in a morter. The mixture is then melted in measurements after each heating interval a quartzship at approximately 800-900 C in are summarized in Fig.2 which gives the dried nitrogen atmosphere, the mixture fraction of activity left, F(t), as a func- appears white-yellow and crystalline. It tion of the heating time. is hygroscopic, and is preferably stored in a desiccator. The melting point was measured to be 347+5 C in accordance 2 with the literature value . 10 -

2.1.2. Target_material _p_regaratign

Fig.l gives the vapor pressure curves of the pure components in the mixture, and of the possible product compounds . At xS temperatures below some 450 C the matrix -}- _;_ vagor pressure can still be kept below 1 10 torr (which is necessary for a stable 10 ** O : source). At this temperature the vapor COMPONENT pressure of the probable products are r r rather high. Hence, when the product com- LONG pounds are present at the liquid surface, SHORT COMPONENT the evaporation is expected to be fast. -

in" i i i 1111 1 1 1 1 1 1 1 1 6 9 12 HEATING TIME (mm)

Fig.2 The,factional release of antimony ( Sb) at 400 °C as a function of the heating time.

Liquid targets have previously been reported to show characteristically a rather simple release behaviour ' , where the release process can be described by a single exponential term

F(t) = e-Pt (1)

Here u is a constant believed to be related to the surface desorption step. This description does not, however, fit the data in Fig.2. It is necessary to 600 BOO 'COO introduce one extra delay component. About TEMPERATURE CO 10% of the totally produced antimony acti- vity is ruled by this long component. The Fig.l Vapor pressure curves of the probable origin of this delay is not easily recog- chemical reaction products of the nized. Somewhat speculative suggestions spallation produced elements indium» may be that the temperatur is kept much tin and antimony. lower here than in ref. s. 3,9) and a liquid diffusion time may play a role. Besides, since this is a chemical target, The thermal stability has been tested the necessary chemical reactions for for- gravlmetrically, and revealed an evapora- mation of the chlorides have to be completed tion loss rate of less than 1 permille per before evaporation can take place. hour for sample sizes of 2-3 g, a tempera- ture variation in.the range 390-470 C and a pressure of 10 to 10 torr. After However, by subtracting the log delay 48 hours at 400 C, the mixture is still a component from the data in Fig.2, a single liquid, thus underlining that the change in exponential component can be fitted to the chemical composition is not pronounced resulting points. during this period of time. The delay half-time is defined as the The reaction product release rate.was time needed for evaporation of one half tested in a vacuum chamber experiment '« of the original activity. From the measur- The probe nuclide was selected to be Sb ed delay curve in Fig.2 one finds a gross (60.3d). Small samples of the mixture were delay half-time of (t, )_- 2.6 min. The irradiated with 600 MeV protona for 72 hours, short component alone'ylelds (ti)D « 2.3 and allowed to decay for 4 months. The samp- mln, and the long component (t, FD> 30 min. les were successively transferred to small graphite evaporation containers surrounded 2.1.3. Ihermgchromatg2rap.hv by tantalum for ohmic heating, and mounted between two.electrodes in a vacuum chamber Thermochromatograph ic experlmentu test bench . Heating intervals of 3 min have been performed on the released pro- were applied tg,the samples. The pressure ducts where internal parts of the oven was 10 to 10 torr and the temperature consisted of pyrex glass or quartz tubes, 400 C. The results of the Y-spectrometric and a. linear temperature gradient between - 733 - 420 WC and room temperature was establish- achieved by shielding the line against ed. The results show that 100% of the re- radiative heat from the target and the leased products (both of Sb and Sn) were ion source by stainless steel heat found in the liquid nitrogen-cooled trap screens, and by a water-cooled copper- mounted at the outlet.of the chromatograph. tress on the line. The system is illu- The activity of tin ( Sn) was, however, strated in Fig.3. rather poor. 3.1.2. The_bombardin2_beam 2.2. Isotope separation The bombarding particle available On-line conditions has been simulated, at the CERN synchrocyclotron for^tbese at the ISOLDE of f-.'ine isotope separator . experiments was the 8 6 MeV/amu C """. The target and ion source arrangement used This beam does not give optimal condi- is described later. tions due to its short range (see below) and high ionizing effect. The Samples of 5-6 g of 600 MeV proton later will result in breaking of chemi- irradiated material were heated to 420 C cal bonds and local overheating, thus for about 30 min, and the separated activi- producing higher vapor pressure than ty collected on an end strip placed in the expected during proton irradiation. focal plane in the collector tank. Gamma - The high vapor pressure will in turn spectrometric measurements were performed lower the ionization yield and disturb on the target material before and after the extraction optics. It was found the separation, on the endstrip and on appropriate,* to work with a beam intensi- various parts of the target-ion source ty of 5-10 particles/s. assembly. The results show that about 75% of the original antimony activity in the 3.1.3. Production ^ield_measurements target is released. Only traces of antimony was found on the ion source internal parts. An ionization efficiency for antimony of Since the projectile has 2=6, it >4.5-6 % has been derived. Small amounts of is possible to form products with higher tin was also released from the target, but atomic number than that of the target practically nothing reached the collector element in appreciable amounts. Pro- strip. duction yield measurements of Sb, I and "-• isotopes are presented. The two j. .tter results from.proton transfer reactions from the C-ion, and will 3. On-Xine experiments normally not be observed in proton irradiations. 3.1. Experimental Indium and tin isotopes were not observed. 3.1.1. Tar2et_description The mass separated beam from the ISOLDE on-line isotope separator was collected on the aluminized side of The general layout of the target,is a movable thin plastic tape, and the similar to the normal ISOLDE-targets '. source subsequently transferred into Due to the corrosive behaviour of the detection position 50 cm away. All the molten mixture the target container and yield measurements were performed by the transfer line to the ion-source had Y-ray spectroscopy using a Ge(Li)- to be made of quarts. The line is con- detector with standard electronics nected vertically onto the middle of the connected via CAMAC to a HP-computer. A target container, bends then horizon- 47ig-plastic detector was utlized for opti- tally and finally vertically down into mizing the separator parameters on each the 4-on, source which is of the FEBIAD mass. The data were stored on magnetic type* ' . The target container is tape, and subsequently analyzed by the covered with an outer mantle of stainless computer code GAMANAL ' . steel, and the target is heated by passing DC-current through this mantle. In order to avoid too high trans- 31.4. Delay_;time measurements fer of unwanted species from the target to the ion source the transfer line has to be kept relatively cold. This is The ideal nuclide for on-line delay time measurements should be long-lived compared to the delay time in question so that decay-corrections can be neglect- ed, - be shielded in order to avoid parent corrections and have reasonably high production yield giving sufficient counting statistics. The best candidate among the antimony isotopes according to these criteria is Sb (16 min., 60.4 min). These measurements wore performed with the same expi rimontal setup as described in the preceoding section. The structure ot the time-sequonce for the different operations is illustrat- ed in Fig.4. The collection timo and sample transport arc controlled by elec- Fig. 3 Cross-cut of the target-ion source. tronic "flip-flops" driven by a crystal

- 734 - , IV 5b- SOTOPES - - s-

- s / / /

/ s / -- II. / / 5 IQ5 - / / / / Fig.4 Illustration of the time sequence s _ ^' _ used when accumulating the on-line / / , delay-curves for Sb. / / 5 /• / $ m _ clock. The counting after each collec- Z .o- - / tion is manually started on a light / / signal. The error in starting point / is estimated to be less than 0.5 s, and y hence without any practical importance. All the y-ray spectra were stored on magnetic tape, and subsequentlVc^nalysed X by the HP-computer code ISANLT '.

\ 1 1 t 1 i i 1 t ( 3.2. Results and discussion 112 IK 116 11B MASS NUMBER A 3.2.1. Production_i;ield_gf_antirnony_ isgtoges Fig.5 Curve I shows the production yields for Sb-isotopes from 1O-8O MeVV C - The yields are calculated as the induced (5*10 parts/s) spallation production yields (in the collector in a Te-based (17 atom%) target thank) at saturation by the formula material of thickness x32 g/cm . Curve II is curve I normalized to the averagejmajjmum intensity available (2) of the c" (after stripping in the glass wall 6+), 3*10 part/s. Curve 111 is curve I normalized to 1 part >\(6.24-10 part/s). Curve IV is curve l^raised with a factor of 3.75-10 estimated £or a 2.5 uA proton beam. where physical and experimental parameters. Also S = the total net integral in the y~peak shown in Fig.5 is the estimated yield t, = decay time curve for the (at present) full C - beam intensity of 3*10 particles/s. t ,. = collection time ''count = c°unting time 3.2.2. Production_yields_gf_xenon_and E = absolute counting efficiency at the ' energy E E = beamline transport efficiency (=80%) The xenon and iodine isotopes pro- I = branching ratio of the y-line duced in proton transfer reactions are detected simultaneously and their produc- Here the collection, decay and tion yields are derived. counting times were preselected for each All the Xe-yields are calculated by nuclide. The absolute counting efficiency the formula (2). The results are given was^(3gtermined using a calibrated source in Fig.6. Of the measured 1-isotopes of Eu, and the decay scheme information the only one that can be directly.,, was taken from various issues of "Nuclear,, calculated by the formula (2) is I. Data Tables" and from "Table of Isotopes" . The yield.fpr I .is calculated from the The quoted errors are composed of the observed Fe activety. For the manses statistical error in the net number of 117, 118 -rid 319 a correction has to be given by the code GAMANAL (a ), an overall made (although rather small) for the error of 5% in the efficiency curve (or. ) growth from the corresponding Xe-motheri; and 10% in the beamline transport efficien- both durinn collection and counting. cy (a t). (oqn.3) The yields are illustrated in Fig.5 and listed in Table 1 together with sone

- 735 - Table 1. Production yields of Sb-isotopes by C -irradiation (5-10 part/s) of a Te-based (17 atom?) target with thickness 32 g/cm<-.

IY(abs) Number of Prod. yield Average prod. Y counts atoms/s yield, atoms/s keV

110 23s 984.7 31.2 900 900 1 1 111 lit.Is 154.5 64.5 0.0485 300 299 115±3i 3.1-IO1 ±1.5- r 3.1-101 ±1 .5-10

1 1 1 112 53.5s 1257 95.0 0.0067 90 90 17*7 6.4-io + 4.8-IO 6.4-1O1 +4 .8•10

113 6 .7m 331 10.4 0.0247 300 299 325±46 1.5- +.3.3- 498 80.0 0.0160 1209±51 1.3- $ +1 .9-10' 1.35-1O3 ±1 .6 •io2

887 300 125+15 8.0-JO-; 114 3.5m 17.7 0.00925 299 1022 1300 100 0.00645 7.8-10"" W:\: 7.9-1O2 ±8 .7-101 3 115 31.8m 491 3.8 0.0163 300 597.5 152±73 5.5- 10? + 2.7 *,o 3 3 498 99.1 0.0160 1 .10-10'* ±1.5- 10 9.74-1O ±1 .3-,o 3 2 116 16m 933 24.7 0.00885 300 597 177+33 1.1- io .* 2. 5 ,o• 1.1-lO3 ±2 .5•io2

116 60m 99 32.0 0.055 300 597 3278+.169 7.4- + 1.1- 135 29.0 0.0535 2711±162 7.0- 3 + 1 .0-103 407 42.0 0.0195 1438-+.72 6.9- ±9.9- lo 103 0.0148 1534r67 7.9 -io3 ±1.1- 7.27-103 ±5 .2 -io2 543 52.1 ioJ 4 117 2.8h 158 86.1 0.0477 300 592.8 19279+538 4.98- in ±6.8- 103 4.98-1O4 ±6 .8•io3

118 300 588.2 1.2- ±1.9- 5.Oh 253 93.0 0.0307 1811+148 103 1050 98.0 0.0079 375±60 9. 1- ±1.9- 103 3 1229 100 0.0068 439±59 1.2- ±2.3- 103 1.1-10* M .2•io

s-x. • Xe-ISOTOPES (3) O I -ISOTOPES

S '0'

x aexp(x Lh i -i i^unt» , ,-X, 1 X, • d-exp(-X2tcounc)) " 2 10J

- [l- -^ i 2 1 rø 1 I *2-A, (l-«cp(-A2tcoU))

115 120 IZb MASS NUMBER A The subscripts 1 and 2 means the mother and daughter nuclide respective- ly. The results are given in Fig.6. Fig.6 Measured production yields of Xe- isolojies (•) ami J-igotopos (o) with the same conditions as described in 3.2.3. Estimate_of_the_Broduction_^ield the text to Fij;-5.

estimated to be T-2.7 g/ctn' (the glass full_intensit^_Brgton_beani wall of the target containur degrade the particle energy to v80 MeV/amu). The 12 4 When the C -ions pass through the total target thickness Is calculated by quartz wall of the target container, d-, » l'w/V , whore I is the length of the they areg^otally stripped of electrons tirget = 12 cm, w Is tlie total weight = into C -particles. Dy mnanSgijf Fig.7 27 g and V is thu volume of the,melt = (which gives the ranges-of C -ions 10 cm . This gives d_ = 32 g/cm . in different elements) , the range of Accordingly, only the first 1/12 of the these particles in the target mixture is total target thickness contributes to

- 736 - •s- \ • • CJJ'

''Ul ... -- ~- ''-•

\ \ "-- '-iV'c-ia',

• 0«.,. ,'.«• 1',

• • « • •

• T

40 60 ELEMENT (Z)

Fig.7 Range (in g/cm ) of C +-ions o£ 80 and 85 MeV/n in different elements. the activity production.

For 600 MeV protons the target can be considered as thin, and the whole length contributes equally to the production of the measured isotopes. If the production rate of the measured isotopes from C irradiation is constant over the carbon-ion range (rough estimatel and the reaction cross section for C -bombardment and protons are equal then the yields increase by a factor of,"'12 for luA p as compared Fig.8a Delay-curve for 5b (from 511 keV and to 1 part IJA C. An additional factor and b i:93 keV resp.). of 2.5 is gained by increasing the beam current to 2.5 uA p (the maximum permissible onto an ISOLDE-target at present). V-lines with known relative intensities, The measured production yields of belonging only to this isomer, i.e. the antimony isotopes then have to be 99 keV(32%), 407 keV(42%) and 542 keV(52%). multiplied by a factor of 3.75-10 to The shape of the decay curve could then be arrive at the estimated production established, and the proper corrections yields from 600 MeV proton bombardment with made in ths delay curves. The resulting a maximum beam intensity of 2.5 uA. The delay curves are shown (dashed) in results are shown as curve IV in Fig.5. Fig.s. 8a and b. While the vacuum chamber delay curve (section 2.1.2) expresses the remaining 3.2.4. The_delay_time_of_antimony^ (or fractional) activity in the sample after heating, denoted Fit), the presently derived curves result from a differential measure- The statistics obtained in the 1.5 min ment of the released activity, denoted f(t): counting intervals was rather poor, and the only y-lines possible to use for the fit) = - J-6F(t) delay-time calculations are the 12 9? keV <5t U) and the 511 keV annihilation radiation peaks. The latter is practically free from The vacuum chamber curves can be fitted general background..Heither do the two with a sum of two exponentials, and only nuclides lx Xe and I disturb. The. about 3 0% of the total amount of produced measured delay curves after "beam off are antimony nuclides are ruled by the long delay given in.Fig.s 8a and b. However, the component. Supposing the same delay properties nuclide Sb has two isomers with half- in on-line experiments, a long component lives of 16 min and 60 min. Both contribute will be depressed in the resulting delay to the two Y-lines, and this must be taken curves, and appear as a level hardly distin- into account when making the decay correc- guishable from the general background (with- tion of the measured delay curve. The con- in the counting statistics). However, the tribution from the 60 min isomer in the delay can not be expected to be identical in 1293 keV (with I , = 100«) at constant vacuum chamber and on-line experiments, and production rate Is calculated from other the present curves also show a much more pronounced long component which may be due Acknowledqement to relatively slow ad=orbtion and desorbtion processes on the moderate temperature quartz We would hereby like to express our grati- surfaced in the target container and transfer tude to the Dr.s. A. Knipper and line. The short component is, however, still C. Richard-Serre, and to the ISOLDE ^prtciable, and the derived half-time values collaboration for interest and practical of 2.4 min and 2.8 min (for the 1293 keV and assistance during parts of the experiments. 511 keV peaks, respectively), are in accordance with the vacuum chamber results. The errors are estimated to be +50». References

1) 0. Glomset and E. Hagebø, 4. CONCLUSION ISOLDE report 1.12.1972, unpublished results. The tellurium-based melted mixture has been thoroughly tested both in vacuum chamber 2) J. Alstad, B. Bergersen, T. Jahnsen, and off-line isotope separator experiments. A.C. Pappas and T. Tunaal, A target made of this mixture has been tested CERN Yellow Report 70-3, 17(1970). on-line under especially unfavorable condi- tions, (short range and high ionization densi- 3) J. Alstad, O. Glomset and E. Hagebø, ty) . ISOLDE internal report 1.11.1970, The target material has survived the unpublished results. tests in good shape. The reason for not observing tin and 4) P.G. Rustanov indium in the mass separator experiments Azerb.Onsk.Chim.Zurnal, 4_, 57(1962). is not fully examined yet. However, two cir- cumstances may be contributing: low vapor 5) Landolt-Bornstein, pressures and formation of chemical side- "Zahlenwerte und Functionen aus Physik, bands. The relatively low vapor pressures Chemie, Astronomie, Geophysik und of the probable chemical reaction products Technik", Vol.2, 6th ed.. Springer (see Fig.l) counteracts an efficient release Verlag, 1960. from the target. On the other hand off-line results show that part of the tin is re- 6) V. Hjaltadottir, leased (possibly as SnCl,), but can not Thesis, University of Oslo, 1976. be recovered at the Sn-mass on the collec- tor strip. Chemical sidebands are often 7) L.C. Carraz, I.R. Haldorsen, H.L. Ravn, formed in isotope separators. Unpublished M. Skarestad end L. Westgaard, ISOLDE results ' show that addition of Nucl. instr. Meth., 148, 217(1978). fluorine or chlorine-containing compounds in small quantities to the target matrix 8) H.L. Ravn, S. Sundeli, E. Roeckl and leads in several cases not only to an L. Westgaard, increa-.ed production rate at the proper J. inorg. nucl. Chem., 37, 383(1975). mass M, but to still higher rates at sidebands like MF and MF, or MCI and MCI,. 9) E. Hagebo, A. Kjelberg, P. Pat.zelt, Sidebands with a still higher number of G. Rudstam and S. Sundell, ligands may be formed and separated, CERN Yellow Report 70-3, 93(1970). depending upon the stability of high oxidation states of the central atom. 10) G. Andersson, B. Hedin and G. Rudstam, Such sidebands have not been checked Nucl. Instr. Meth., 2§, 245(1964). in the present work. As we believe that the target material 11) H.L. Ravn, S. Sundell and L. Westgaard, will stand a 2.5 pA proton beam, the Nucl. Instr. Meth., 1^2' 131(1975). estimated yields in Fig.5 establish that this target/ion source combination 12) R. Kirchner and E. Roeckl, may be a useful ISOLDE production target. Nucl. Instr. Meth., 12J7, 307(1975). It is not unlikely that the delay time of -\-2.5 min can be improved somewhat in proton 13) R. Kirchner and E. Roeckl, irradiations, since a more uniform tempera- Nucl. Instr. Meth., 131, 371(1975). ture can be achieved in the target melt. This opens up possibilities to extend the 14) R. Gunnink and J.B. Niday, research,into the poorly known region Lawrence Livermore Laboratory, Report beyond Sb. Spin and magnetic moment UCRL-51051, Vol.I-IV. measurements with the on-line ABMR- technique reguire, at present, yields 15) ISOLDE internal computer code. in the range 10-10 atoms/s^u', while lazer spectroscopy with the collinear 16) CM. Lederer and V.S. Shirley (eds.), technique ' 'can make use of beams "Table of Isotopes", John Wiley s Sons, down to 10 atoms/s . Recent interest 7th ed., 1978. at ISOLDE in performing Mossbauer studies on-line with the mass separator, 17) C. Richard-Serre, has been expressed . Experiments are CERN 72-19, 1972. already performed using In. The same low energy y-transition of 2 3.87 keV 18) T. Bjørnstad, H.A. Gustafsson, follows the decay of Sb, which can now be O.c. Jonsson, V. Lindfors, produced clean and in good quantities. A.M. Poskanzer and H.L. Ravn, ISOLDE internal report 9.5.80, un- published results.

- 738 - 19) C. Ekstrom, S. Ingelmann and G. Wannberg, Nucl. Instr. Meth., 148, 17{1978).

20) C. Ekstrom, private communication, 1980.

21) S.L. Kaufman, Opt. Commun., 12, 309(1976). 22) K.-R. Anton, S.L. Kaufmann, W. Klempt, G. Moruzzi, R. Neugart, E.W. Otten and B. Schinzlor, Phys. Rev. Lett., 40, 642(1978).

23) R. Neugart, private communication, 1980.

24) G. Weyer, proposal for Mossbauer experiments at ISOLDE, presented at the open ISOLDE meeting, December 1978.

- 739 - PAPER P.XII. Nuclear Instruments and Methods 186 (1981) 307-313 North-Holland Publishing Company

INTENSE BEAMS OF RADIOACTIVE HALOGENS PRODUCED BY MEANS OF SURFACE IONIZATION

B. VOSICKI, T. BJ6RNSTAD, L.C. CARRAZ, J. HEINEMEIER and HL. RAVN CERN, Geneva, Switzerland

A negative surface-ionization source has been developed for on-line separator use in order to make intense ion beams of nuclear-reaction produced halogens. It consists of a planar LaB6 surface onto which the mixed volatile nuclear reaction products are allowed to impinge. A transverse permanent magnetic field and an intermediate electron catcher electrode, inserted before the separator extraction gap, is used to separate the ion and electron fractions. The efficiency of the source is shown to be up to 50% for the halogens, Br and I. i.e. very similar to that obtained by the same process for positive ionization of alkalis on a tungsten surface. During periods of several weeks the source has been used for a number of on-line experiments and has allowed the identification of a series of new nuclei. Yields and halflives of 4i43Cl and ™."^"™««Br are reported.

1. Introduction experiments [8-10], to be as efficient and reliable as experienced with positive surface ionization of Up to the present time, methods for on-line the alkalis on a tungsten surface. The perfor- isotope separation of about 40 nuclear-reaction mance of the source is described in terms of produced elements have been developed at production yields, ionization efficiencies, and ISOLDE [1], For 22 of these elements intense delay-time distributions. The new region of ion beams, uncontaminated by isobars from nuclei far from stability made accessible by this neighbouring Z elements, are available for stu- technique is illustrated by the half-life deter- dies. This selectivity is achieved by proper com- mination of a number of newly identified nuclei. bination of target materials [2,3] plasma-dis- It now seems possible to extend negative sur- charge ionization [1,4,5] and positive surface face ionization to a number of elements less ionization [5]. By application of the negative electronegative than the halogens. For this pur- surface ionization effect a new class of isotope- pose, surfaces with work-functions down to separated beams can be obtained, making use ~ 1 eV, as reported in the literature, may be of the high selectivity of this process that forms needed. The very high yields of negative mole- negative ions of the halogens. This was demon- cular ions formed by surface ionization may be strated early on by Venezia and Amiel [6]. In an alternative method. their off-line separation of iodine, ionization yields of only 10~4% were obtained, presumably owing to the high work-function of the graphite 2. General principles of the negative surface ion- emitter used. Also in the on-line separations of ization process Br and I, by Reeder et al. [7], low yields were obtained from their LJO2-graphite surface. The degree of negative surface ionization is The present paper describes how the use of described by an expression similar to the well- low work-function thermoionic emitter surfaces known Saha-Langmuir formula allows the development of a negative surface- ionization source for on-line separation of Cl, Br, a =£-exp| n kT (1) and I with efficiencies of up to ~ 50%. In com- So -)• bination with different target materials the where g. and go are statistical weight factors source has proved, in a number of on-line describing the number of states accessible, res-

0029-554X/81/0000-0000/$02.50 © North-Holland V. ION SOURCES TECHNIQUES 308 B. Vosicki et al. / Production of intense beams of radioactive halogens pcctively, in the negative ion or the neutral should not react appreciably with halogens, and atom, weighted by their Bolzmann's factor at the has a low resistivity. actual temperature, A is the electron affinity (eV), 4>o the "clean" surface work-function (eV), /(£") the Schottky effect [11] describing the 4. Construction and characteristics of the target diminution in 4> with increase in the electrical and ion-source combination field E, 0 the degree of monolaycr surface coverage, k the Bolzmann constant, T the ab- The principle and construction of the source is solute temperature (K), and c a constant that shown in fig. 1. The negative ion emitter consists may be positive or negative. Partial surface of a porous pellet, 2 mm in diameter and 2 mm coverage with a low or a high work-function thick, contained in a cavity in a tantalum holder. material may, respectively, decrease or increase The pellet is prepared simply by pressing the dry the effective work-function [12,13]. In cases LaB6 powder by hand into the cavity, followed where the Schottky effect and the surface by a sintering in vacuum at 1800°C for 15 min. coverage effect can be neglected, a high degree The pellet holder is inserted into the extraction of ionization is obtained when 0 < A. end of the tantalum transfer lube which forms For proper operation in an on-line isotope the connection to the target. The volatile nuclear separator, additional demands of the surface are reaction products, carried away from the target that operation can take place at elevated tem- by thermal diffusion through the ohmic heated peratures (in order to minimize the delay-time transfer tube, impinges onto the emitter surface loss in this step for short-lived nuclides), that it with a geometrical probability of —50%. The has a low vapor or decomposition pressure, that ions formed are extracted by the extraction elec- it is chemically inert, and that it has high enough trode, kept at a potential of 20 kV. In order to electrical conductivity to ensure a fast replace- avoid the acceleration of the abundantly emitted ment of the electrons removed from the surface electrons to the 60 kV potential of the separator in the ionization process. they are deflected onto a separate powered elec- trode kept at a potential of ~ 1 kV." The deflection is obtained by means of an 800 G 3. Choice of ionizer permanent magnet field.

A characteristic property of the LaB6 emitter The electron affinity of the halogens is always is its sensitivity to poisoning by adsorbed im- greater than 3eV [14], except for At, (2.8 ± purities [13]. It therefore needs an activation by 0.2)eV [15]. Hence a surface with <£<2.8eV heating for some minutes to 1500-1600°C each will ensure a high degree of ionization for these time it has been exposed to the atmosphere. The elements. Results are reported for different sur- degree of activation can be followed by measur- faces which have been tested for negative ing the electron emission, which reaches about halogen-ion formation. Examples are Cs [16], 10 mA in the present set-up. The negative ion 7 m carburized thoriated tungsten [17] and GdBe beam yield measured with * Br (Tm = 1.35 s) as [18]. But the most promising material is LaB„, of a function of emitter temperature shows a which the thermoionic properties have been pronounced maximum at ~ 1260°C. This extensively investigated [17-24]. A survey of behaviour may be understood as the efficiency measurements of its work-function shows a range variations for a porous emitter described by Pel- of values from 2.36 to 3.3 eV, with the majority letier et al. [27], who showed that the efficiency of values around 2.7 eV. The spread in the values increases as a function of decreasing surface most probably reflects the effect of surface coverage, followed by a decrease when, at high poisoning, which can be quite pronounced for temperature, the diffusion of the halogen in the low work-function surfaces [13,25,26]. After porous system changes from surface diffusion to promising off-line tracer ionization with radioac- volume flow. I31 tive I, LaB6 was chosen as the emitter material Two targets have been used in these experi- for the halogen experiments. This material is ments: a nobium powder target [23] containing refractory and stable up to 1500°C in vacuum, 131.6 g Nb (thickness 85.2 g/cm2) and a UC- B. Vosicki et al. I Production of intense beams of radioactive halogens 309

Fig. 1. The principle of and the technical details for the negative surface-ionization source. 1, permanent magnet; 2, pole-piece; 3, ionizer constriction; 4, product transfer line; 5, electron catcher electrode; 6, insulator; 7, mounting plate. graphite target [5] with 84.5 g ^U (thickness 5. Results 54.9 g/cm2) where the amount of uranium con- stitutes approximately 10% of the atoms in the 5.1. Delay-times target matrix. It is well known from off-line studies that halogens are released efficiently from The delay-time is the time that has elapsed both these targets [2,3]. The target layout is between production and detection of a nuclide. described in ref. [2]. It cannot usually be quantitatively described in

V. ION SOURCES TECHNIQUES 310 B. Vosicki el al I Production of intense beams of radioactive halogens terms of a simple constant or average value for a from the Nb-power target and for chlorine and given system. It must be represented by a delay bromine from UC-graphite target. These are time distribution-function, as discussed in refs. illustrated in figs. 2-5. The experimental points [1], [3] and [5]. Only from a knowledge of the are not corrected for decay according to the shape of this function can conclusions be drawn actual delay time distribution, and represent about the delay-time losses of short-lived nuclei simply the isotopic yields (in atoms/s) at the and the release mechanisms. It is often difficult detection position. The values have been to measure this function accurately and a more measured by a Faraday cup and a Air /3-detector simple readily-measurable quantity, the delay operated in the multiscaling mode, except for the 94 half-time (rd)l/: [28] may be used qualitatively to two nuclidcs "• Br where a 'He neutron counter describe the speed of the system. It expresses was used. A relative yield curve containing the how fast the constant production yield of a given mass numbers 92-94 can be constructed by in- nuclide is halved after the proton beam is swit- tegration of the peaks in the neutron scan shown ched off [29]. The {u)m value indicates whether in fig. 5 corrected with their p„ values. The p„ or not there is an appreciably fast component in the diffusion mechanism. For the presently used Nb powder target (t = 2100°C) for production of Br it was found that at low transfer line temperature (<1000°C) the delay-time distribution as measured with 76mBr showed a line-temperature-dependent single exponential behaviour [28] with (rd)i/: = 169 s. At higher line-temperature (1300°C) distribution changes into a function which is best described by a sum of exponentials with (td),n - 25 s. Off- line measurements with stable Br showed that at low line temperatures the overall delay is 40 42 MASS NUMBER determined by desorption of Br from the line and ionizer surfaces. This partial delay could be Fig. 2. Yields of O isotopes as CT from a 600 MeV proton- reduced to (fd)i/2~3s at a temperature of irradiated UC-graphile larger The points correspond to a uranium target thickness of 54.9 g/cm2 and a proton beam ~1500°C. In combination with the UC-graphite intensity of 1 fiA. The solid curve is drawn to guide the eye. cloth target [5] kept at 2I00°C and a line-tem- perature of > 16WC a (f

('d)i/2 = 25 s for the release of Br from the Nb »7 powder target is caused by the combined effect / of a lower line temperature and a slower release «• from this matrix. No systematic study of the „i delay of Cl has been made but the steep slope of 10 Din the yield curve discussed in section 5.2 indicates w4 a somewhat longer delay. Most likely it ori- w1 ginates on the line and ionizer surface, since \ • off-line measurements [3] of CI release from w1 high-temperature target materials shows a parti- 1 • , . .,...... \ cularly rapid release of Cl. 75 BO «S on •• MASS NUMBER 5.2. Production yields fig. 3. Yields of Br isotopes as Br" from a 600 MeV proton- irradiated UC-graphite target. The points correspond to a uranium target thickness of 54.9g/cm2 and a proton beam Production yields are determined for bromine intensity of 1 jiA. The curve is drawn to guide the eye. fl. Vosicki et al. I Production of intense beams of radioactive halogens 311

maximum probably around mass 131-132 of about the same magnitude as the one presently given for bromine. By comparing the Cl yields in fig. 3 with those of K from the same target [1] it is striking that the slope of the Cl curve is much steeper than that of K. This effect could be due to pronounced delay-time losses of the short-lived nuclei. The elements fluorine and could not be found in the mass spectra at their respective atomic masses although they were expected to be released from the target. The formation of non- n MASS NUMBER ionized chemical compounds or molecular side- Rg. 4. Yields of Br isotopes as Br" from a 690 MeV proton- bands is however possible, but no systematic irradiated Nb-powder target. The points correspond to a search for sidebands has been carried out, to large! thickness of 8S.2 g/cm2 and a proton beam intensity of date. I p. A. The points marked m are only isomehc yields and the curve is drawn to guide the eye. 5.5. lonization efficiencies

The ionization probability of the halogens which impinge on the LaB6 surface is according to eq. (1) close to unity. The expected ionization efficiency of the present source arrangement should therefore be ~ 50%, since the probability that the atoms strike the LaB6 is -50%, as discussed in section 4. By comparing the production rates of Br in the Nb target with the obtained yields shown in fig. 4, an efficiency of (73 ±25)% was found, in agreement with the expectations. However, in combination with the XJC- graphite target the Br ionization efficiency determined by the same method gives only 1%. Furthermore this efficiency continuously decreases but could be restored by reforming the ionizer as described in section 2. This is taken as clear sign of poisoning of the ionizer, presumably by carbon which is being outgassed abundantly Fig. 5. Mass scan on heavy Br" isotopes obtained by means from the target as CO. In fact Avidenko and of an integral 'He neutron counter. Malev [13] have demonstrated the sensitivity of LaB6 to carbon contamination. value for æBr is taken from ref. [25], and the The ionization efficiency for Cl at present /vvalues for '*MBr are estimated on the basis of reaches only a few per mille, most likely owing a linear fit to the values for """Br [25]. This yield to the combined effect of poisoning of the emit- curve is then normalized to the yield of 92Br ter surface and formation of molecular com- measured by the ^-counter. pounds. So far no complete yield curve has been determined for iodine. But the two isotopes 5.4. Identification of new halogen isotopes 137 7 measured, YP( I) = 3.5 x 10 atoms/s and 5 YP('"I) = 3.1 x 10 atoms/s, point to a yield curve The first successful on-line use of the negative

V. ION SOURCES TECHNIQUES 312 B. Vosicki et al. I Production of intense beams of radioactive halogens

Table 1 Nuclear dala on some new and poorly known halogen isotopes

Nuclide Haiflifc Number of Other radiation Literature quotations measured halflivcs followed "a 6.8 ±0.3 6 ncmission not detectable - "a 3.3 ±0.2 5 n-emission not detectable "Br 2.2 ±0.2 10 nol measured »,,j-H0msf291 "Br 21.5 ±0.5 6 nol measured - '»"Br 7.2 ±0.5 5 Delayed p-branch of 6 x 10 "' (in- 1.3min[3O| «Br 73.2 ±0.5 7 *""Br 1.35 ±0.05 6 /„, = (1.49 ±0.02) y emission (keV)*: y emission (keV): 44.5 ±1.0 45.48 56.6 ±1.0 57.11 "Br nol measured n-emission _ wBr nol measured n-emission

•The origins of the proton-emitting levels are not clear at the moment.

''The large error limits are due to the simple energy calibration with Ihc two X-ray lines in lead K*, (74.97keV) and KSl (84.94 keV). surface-ionization source allowed the extension targets are used. Two such high-temperature of the region of halogen isotopes to a number of targets are at present under study; Ta powder for new and previously poorly known nuclei. Results fragmentation production of Cl and ThO2 for from preliminary radioactivity measurements on production of the neutron-rich Cl, Br, and I. The 4l4'CI and ™^™™»5*Br are summarized in table long delay-times for Cl may be due to the higher 1. The halflives are all determined by multiscal- chemical stability of the chlorides, which ing of the /?~ray activity and computer resolution prevents their dissociation and desorption from of the decay curves. At mass 70 a nuclide which the line or ionizer surfaces. Higher line tem- decayed with a halflife of 2.1 s was identified. perature combined with an ionizer like ZrC that This value is different from the halflife of 80 ms operates at high temperature may solve this determined by Alburger et al. [29] for ""Br and problem. Furthermore, this measure could in- indicates that isomerism might exist in this crease the possibilities of making radioactive nucleus. The halflives of the two low-energy beams of fluorine, which is the halogen that •y-rays at mass 76 are both determined to 1.3 s by forms the most stable compounds. multi-spectrum analysis of the y-ray activity. A number of elements have electron affinities They are reported [30] to be coincident and in the range 2.8-1.8 eV [15] such as Au (2.80 eV), interpreted as the isomeric transition. Pt (2.56eV), S (2.07eV), Se (2.12eV), and Ag (2.0 ev) [15, 31], for which promising target materials also exist. Efficient negative ionization 6. Conclusion and outlook may be extended to these elements in case sur- faces with < 1.8 eV can be found. In fact, such This test constitutes the successful adap- surfaces are reported to exist either as elec- tation of a surface-ionization source to an on-lire tropositive layers on a refractory metal like W- mass separator for selective production of nega- Ba (4> = 1.6 eV) [26] or mixed oxides such as tive ions in high yields from a complex nuclear SrOBaO ( = 1 eV) [26]. From the latter surface reaction mixture. This is demonstrated by the an S~ beam of 15 JJ A was obtained in extracted beams of the halogens, chlorine, the preliminary work [32]. The reproducibility of bromine, and iodine. The efficiencies for Br and the low work-functions still remains to be Cl, affected by poisoning when released from a verified, since they are likely to be sensitive to carbide target, may be increased if carbon-free poisoning. B. Vosicki et al I froducrion of intense brams of radioactive halogens 313

Recent observations have demonstrated that [10] I. C Canraz. Nuclear spectroscojiy of fission products. remarkably high molecular electron affinities are ed.. Till von Egidy. Conf. Ser. No. SI (The Institute of observed for several hcxafluorides (33]. The Physics. Bristol. 1980) p. 36. [It] N.I lonov. in: Progress in surface science. Vol. 1 (Per- EA(UF6)=» 5.1 eV allows a 100% efficient surface gamon. New York. 1972) Part 3. ioni/ation. Also in the present experiments a (l.'j M.F. Harrison. Rctcaicli rcpurt AERE GP/R 250S number of negative molecular ions have been (195K). observed. Regardless of whether they have been [13) A A Avidcnko and M.O. Malcv. Vacuum 27 (1977) S83. formed in surface ionisation of secondary effects, j I4| R.S Berry and C W Rcimaiin. J Chem. Phys 38 (I9fi3) 1540 they may offer possibilities for selective ionizalion ||5| R.J /ollweg. j. Chrai Phys. 50 (1969) 42SI. of a number of elements. [16] W Geigcr. Z Iliy* 140 (1955) «18. It seems that the results reported here indicate [17] A. Persky. I".F Greene and A. Kuppcrmann. 1. Chcm. interesting prospects for the future, not only as Ph>s 49 (1968) 2347. regards selective on-line separation but also for a (18] WD. Dong. W.D Kilpartick. J M. Teem and D.E. more general ionization method to be used in Zuccarn. Progr. Astron. Acron 9 (1963) 2t». f 19] J.M Laflcny. 3 App!. Phys. 22 (1951) 299. other fields. [20] E.Ya. Zandhcrg and V.I. Paleev. Sm-. Phys.-Tech. Phys. 10 (I9M>) 1014 (English translation). f21] B.S. Kulvarskaya. A.I. Rekov. V.E Serebrennikova. V.A. Nikolacva and Kh.S. Kan. Sov. Phys.-Tcch. Phys. References 14 (1969) 122 (English translation). [22] D.E. Zuccaro and C.R. Da)geroft. Proc. Second Syrop. [11 HL. Ravn. Phys Rep 54 (1979) 201. on Ion sources and formation of ion beams. Berkeley. [2] Hl. Ravn. S. Sundcll and L. Westgaard. Nucl. Instr. and California. Research Report LBL 3399. (1974) p. viii-8-l. Melh. 123(1975) 131. {23] I. Rachidi. J. Monte. J. Pelletier. C. Pomoi and F. (3| L.C. Carraz. l.R. Haldorsen. ILL. Ravn. M. Skareslad Rinchel. Appl. Phys 1-elt. 28 (1976) 292. and I.. Westgaard. Nucl. Instr. and Meth. 14S (1978) [24] J. Pcllclicr and C. Pom«. Proc. IRth Int. Conf. on 217. Phenomena in ioni/cd Gases 1977. Berlin. 1977. (VEB (4) E. Hagcbd. Proc. Int. Conf. on Electromagnetic isotope Buch-Expnrt-linport. Leipzig. 1977) p. 108. separators and the technique of their applications. [25] J. Pcllcticr and C. Pomot. Appl Phys. Lell 34 (1979) Marburg. 1970. eds.. H. Wagner and W Watcher Run 249. dcsminislcrium fur Bildung und Wissenschaft - For- [26] CJ. Smilhclls. ed.. Metals reference book. 5th cd. (Bul- schungsbcricht K70-28, (1970) p. 146. tcnvorth. Ixmdan and Boston. 1976) p. 1027. [5] L.C. Carraz. S. Sundell. H.L. Ravn. M. Skarestad and L. f27] I. Pelletic:. C. Pomot and J Cocagne. J. Appl. Pl.ys. 50 Westgaard, Nucl. Instr. and Meth. 158 (1979) 69. (1979)4517. [6] A. Venezia and S. Amiel. Nucl. tnstr. and Meth. 87 [28] H.I. Ravn. S. Sundell. E. Roeckl and L. Westgaard. J. (1970) 307. Inorg Nucl. Chem. 37 (1975) ?S3. [7) PL. Reedcr. J.F. Wright and L I. Alquist. Phys Rev [29] D.E. Alburgcr. Phys. Rev. CIS (1978) 1875. CIS (1977) 2108. j.Klj Nuclear Data Sheets II (1974). [8] C. Ekstrom and L. Robertson. Phys. Scrip. 22 (19K0) [3lj L.M. Branscomb. J. Chem. Phys. 25 (1956) 598. 344. [32] B. Vosicki. private communication. [9] A. Huck, G. Klott. A. Knipper. C. Miehc. G. Waller. T. (33] P.F. Dinner and S. Datz. J. Chem. Hiys. 68 (1978) 2451. Bjørnslad. H. Ravn ond C. Richard-Serre. Proc. Int. [34] W.D. Schmidt-Oil, A.J. Haulojarvi and U.J. Schrewe. Conf. on Nuclear physics. Berkeley. California. Z. Phys. A289(197K) 121. Research report LBL-11118 (1980) p. 147.

V. ION SOURCES TECHNIQUES PAPER P.XIII. CLRN SERVICE D'INFORMATION SCIENTlFIQUt

Volume 91B, number 1 PHYSICS LETTERS 24 March 1980

STRUCTURE OF THE LEVELS IN THE DOUBLY MAGIC NUCLEUS 'soSn32

T. BJORNSTAD3, L.-E. De GEER»-1, G.T. EWAN*2, P.G. HANSEN a.3, B. JONSON a-4, K. KAWADE b-5, A. KEREK a-6, W.-D. LAUPPE b, H. LA WIN b, S. MATTSSON a'4 and K. S1STEMICH b a The ISOLDE Collaboration, CERN, Geneva, Switzerland b Institut fiir Kernphysik, Kernforschungsanlage Jiilich, Jiilich, Germany

Received 22 January 1980

Delayed 7—7 coincidence measurements give a half-life of the 4415 keV state in 132Sn of 2.1 ± 0.3 ns. The K conver- sion coefficients of the 299 keV and 374 keV transitions correspond to M1/E2 multipolarity. The levels excited in 132Sn very likely form a 4+, 6+, 8+ sequence above the first excited 2* level.

The Z = 50, N = 82 doubly magic nucleus 132Sn 4415 keV level was determined through delayed 7-7 occupies a unique position between S6Ni and 2O8Pb coincidences. The measured time spectra are shown and thus is of considerable theoretical interest. When in fig. 1. The centroid shifts lead to a half-life of 2.1 its first excited state was identified at 4041 keV [11. ± 0.3 ns for the 4416 keV level and to an upper limit the systematics seemed to indicate a 3~ spin-parity of 0.4 ns for the 4041 keV level. The transition proba-

assignment for this level. Following the observation bilities for the 374 (Jy = 85.4%) and 4416 (Iy = 14.3%) of a 1.7 ps isomeric state at 4847 keV [2] it was keV transitions are given in table 1. pointed out by Dehesa et al. [3] that the first 2+ In a second experiment the beta-radioactivity level also was expected near 4 MeV. The new experi- mental data presented in the present paper do not definitely resolve the puzzle of the spin and parity of Table 1 Comparison between the experimental data and the results of the 4041 keV first excited level, but a 2+ assignment the RPA calculation (set I of the single-particle energies was is strongly indicated. used, cf. ref. |3]). In one experiment performed at the recoil separa- tor JOSEF at KFA Julich, the 1.7 ix% isomer 132mSn Level energy (keV) was produced from thermal fission of 235U. Gamma- exp. RPA calc. rays following the depopulation of the 4847 keV iso- meric state were observed and the lifetime of the 4847 4822 8+ 4714 4704 6+ 4415 4529 4+ 1 Research Institute of National Defence, Stockholm, (4351) 4476 3" Sweden. 4041 4090 2* 2 Queens University, Department of Physics, Kingston, Transition Transition probability (s Ontario, Canada. 3 Institute of Physics, University of Aarhus, Aarhus, Den- exp. RPA calc. mark. 4 On leave from Department of Physics, Chalmers University 132 keV 4.08 x 10s 2.9 x 10s 8* of Technology, Goteborg, Sweden. 8 + s 374 keV 2.82 X 10* 4.3 X 10 4 On leave from Nagoya University, Japan. 7 4416 keV 4.7 X ]0 J.6X 107 4* 6 Research Institute for Physics, Stockholm, Sweden.

35 Volume 91B, number 1 PHYSICS LETTERS 24 March 1980

I32in was produced a! the ISOLDt isotope separator c at CERN by bombarding a UC2 target at 2000 C with 600 MeV protons. After mass separation, the activity was collected on a moving tape that removed the 40s 132Sn daughter activity as well as 6.5d l:52Cs, also present in the beam. A Si(I.i) and a Ge(Li) detector placed at the collection point recorded the e~- and 7-spectr.i simultaneously and allowed the determina- tion of the K conversion coefficients of the 2^9 and 374 kcV transitions. The normalization of the elec- tron- and ganima-inlensitics was based on the 354 keV

Fig 1. Tii.ic distributions ol delayed 7—> coincidences. The curves on Ihc left were obtained when the time-to-amplitude converter was started by the low-energy event and stopped by the higli-cnergy event, and vice versa for the curves on the right. The experiment used two Get Li) detectors (139 cm3 and 148 cm3). The isotopes 13aTcand 136Xe present in our beam provide a simultaneous measurement of the prompt 4* - 2*~(V cascades 297-1279 and 381-1313 keV. respec- tively. The sliifl of the 299-4041 kcV cascade relative to the I32 0 2 1 5 8 10 12 U I ! l 6 I S li « '6 374-4041 kcV cascade (both in Sn) and to ihe references Time channels[2£7ns/chl is easily seen.

r. %\ ~ -1000 » 4041 keV •\\ '••• ^ • 132ln •£ 132Sn ' \ \ \ . M2 1 01

0.05 M1 - \ '--;X299kev : E1 \ *"-J- 374 keV \ I- x T

z •500 °°1 o o 4351 keV 100 200 300 400 keV ••

4416 keV J t

3300 3400 3500 3600 CHANNEL NUMBER

Fig. 2. High-energy 7-rays following the decay of l32In. The 2+- 0* and 4* -• 0* transitions are seen together with a line at 4351 keV, which is not observed in the d -y of the 132mSn isomer. This line very likely represents the expected 3 " — 0* transition. The inset shows measured and theoi^iical K conversion coefficients for the 299 and 374 keV transitions; the results are consistent only with M1-E2 multipolarity. 36 Volume 9IB, number 1 PHYSICS LETTURS 24 March 1980

in the scheme by transitions of 132, 299 and 374 keV.

The measured K conversion coefficients ofaK(299) = 0.028 ± 0.005 and aK(374) = 0.020 + 0.07 are only consistent with an Ml -E2 assignment (see inset in 4 »47 '8*' 17 pi ll3>3'03 <£2' fig. 2). The 132 keV transition (fig. 3) is known to 132 '2«92*O3 have E2 character |2], All known levels in Sn I'M!E2> therefore must have the same and probably positive 4 416 '4*' 2 1 ni parity. The information, however, does no! suffice to 4 351 (3 'I 3 74 3 * O 3 fix the spins; for the first excited state, in particular, 857;2 41% + 'Ml E2. the experiment allows the assignments (1.2. 3) . of + —g 4 041 i2*:<0 4OS which 2 clearly is the more plausible.

43516*05 44157*05 4040 8*0.5 An argument for interpreting the isomeric decay as S .14.3124'% a 8—6-4- 2-0 sequence comes from RPA calcula- tions [3], The comparison given in table 1 shows that not only does the predicted sequence explain all levels observed in this work, and the absence of certain cross-over transitions, but the calculation also quanti- tatively agrees for three measured transition probabili- ties. The detailed agreement with the lifetimes pro-

132- vides strong support for the assignments in fig. 3. 50Sn82 Finally we note that the high-energy singles 7-spec- 132 Fig. 3. Proposed level scheme for Sn. The positive-parity trum, illustrated in fig. 2. shows a strong line at 4351 assignments are almost definite, but several experimental pos- keV which decays in the spectra with a half-life close sibilities exist for the corresponding spins including, in princi- 132 ple, that of the first excited state. The spin assignments given to that of In. This transition, which was not ob- here have strong and detailed support from the RPA calcula- served in the isomer decay, is a strong candidate for tions, see text and table 1. The existence and nature of a the missing 3~ -> 0" transition. 4351 keV state represents conjectures only. The authors wish to thank Professors I. Bergstrom, E2 transition in 124Xe. The high-energy 7-spectrum O.W.B. Schult and J. Speth for fruitful discussions. was also recorded. A summary of the main experi- mental results is given in fig. 2. References The measured half-life of the 4416 keV transition agrees best with an E4 assignment corresponding to a 11 ] A. Kerek, G.B. Holm, L.-E. De Geer and S. Borg. Phys. strength of 15.4 + 3.3 single-particle units. An M3 Lett. 44B (1973) 252. with a strength of 0.20 ± 0.05 spu is also possible [2] W.-D. Lauppc et al., Proc. Intern. Conf. on Nuclear struc- while an E3 assignment, which would correspond to ture (Tokyo, 1977), J. Phys. Soc. Japan 44 (1978) Suppl. 3 p. 335. a strength of less than 10~ spu, is less likely. Positive 13] J.S. Dehesa, W.D. Lauppe, K. Sistcmich and J. Speth, parity and spin 4 or 3 is therefore indicated for the Phys. Lett. 74B (1978) 309. 4416 keV level. This level is linked to the other levels

37 PAPER P.XIV. Zeitschrift A-tytmc / Phvs A - Au.ms and Nudel .'06. 95-97 <1982l fur Physik A MlUI I KJ and Nuclei < Springer-Verlag 19X2

Excited States in the Doubly Closed Shell Nucleus

T. Bjornslad1. J. Blomqvist2, G.T. Kwan3. B. Jonson'. K. Kawade4. A. Kerek2. S. Maltsson1, and K. Sistemich5 The ISOLDF. Collaboration. CERN. Geneva. Switzerland 1 CERN-ISOLDE. CERN. Geneva. Switzerland 2 Research Institute for Physics, Stockholm. Sweden 3 Queens University. Dept. of Physics, Kingston. Ontario. Canada 4 Nagoya University, Nagoya. Japan 5 Institut flir Kernphysik. Kernforschungsanlage Jiilich. Federal Republic of Germany

Received March 18, 1982

New excited states in the nucleus 132Sn have been identified from •/•/ coincidence measurements. Strong beta feeding to a state at 7.210 keV was established. This level is interpreted as a 6" state formed after a KJ;,, 'l-»vj»- \GT(i~ transition from the 7 ground state of 132In. The deexcitation of the 7,210 kcV state passes through a 4,351 keV state, providing support for a 3" assignment of this level.

The first information about the level structure in the suggest that the states observed in the Julich experi- doubly-magic nucleus 112Sn came from experiments ments [2] form an 8' ->6" -»4* -»2* ->()' se- on the ji decay of IJ2ln [I]. A -/ transition with quence. the energy 4.041 keV was interpreted to represent The •/ spectrum from the li2ln /S decay obtained at the deexcitation of the first excited state in 132Sn. ISOLDE showed as a further interesting feature a Later, this state was found to be the lowest of a strong y line at 4.351 keV. which was absent in the y series of four excited states, where the highest state spectrum from the deexcitation of the 1.7 us isomer. at 4.847 keV is an isomer with the half-life 1.7(2) jas This line is a strong candidate for the 3 ->()" [2J. The latter experiment was performed at the ground state transition, which is predicted at about recoil fission product separator JOSEF in Jiilich. this energy from level systematics [lj and RFA The multipolarity of the 132 keV transition deexcit- calculations [4], ing the isomer was deduced to be £2 based on According to the shell model, the first neutron and intensity balance arguments. In a conversion elec- proton-hole orbitals outside the (JV. Z) = (82. 50) core tron experiment at ISOLDE [3], the multipolarity are 2/7 , and lKy2- respectively. With a <5-force [1] £2 was also found for the 299 keV and 375 keV the 7~ member of the (v/, ,.jrg^j) configuration transitions (see Fig. 1). becomes the ground state of 1J2ln. This assignment Recently, the half-lives of the 4.715 keV and also follows from Nordheim's third rule [5]. Further 4.416 keV states have been measured in Julich to be support for the assumption that the state with J =/p 1 20.2(1.6) ns and 4.0(0.3) ns (cf. note below ), respec- +/A — 1 comes lowest in energy derives from a com- tively. These values give good support for the £2 parison with the particle-hole configuration (v/5~,'. 208 assignments based on the conversion electron nh9/2) in Bi, where all members of the multiplet measurement. The results mentioned above strongly are known. The ji~ decay should proceed mainly by an allowed Gamow-Teller transition (\-f. ,. ' Note: From an earlier experiment a somewhat lower value n v n had been deduced [3]. Details about this experiment will be Sv\)-i- ~»(»'/7 2- S-\h • ' 'his paper we report on published separately new experimental data, which show strong evidence

0340-2193 82/0306 0095 SOI.00 96 T. Bjornslad et al.: Doubly Closed Shell Nucleus ''JSnK,

decay scheme shown in Fig. 1 could be construct- ed. The new level at 7,210 keV is most likely the expect-

7210 ed (v/7/2, vg^lK state. The feeding /?"" transition has log//=4.5(1), which is of comparable strength as the elementary Jig^j-^vg,"^ transition in the p~ decay of 13lIn [6]. The reduced GT transition rates l32 to the (v/,/2, vg7,J)j states in Sn should be 23797 226)3 (48) (29) «7/2 7/2) J, I, 7|7/2.(7/2 1)9/2, 7)2 (8/9 for ;=6 ms (5") 1320 [1/9 for J = l 1.912 (7) _S1 4S47 17 tis -u 4791 ^t «715 20 ns in units of the single-hole rate. The predicted small 07) -7 4416 40 ns fractional yield to the J — l state explains why this «351 3X7 (10) P~ decay branch has not been observed. «0*1 The gamma deday of the (v/7|2, vgf.2|h state is expected to proceed by £2 transitions to (v/, 2. vrfj,2)j states, analogous to the simple vg7,2-»vrfj2 gamma decay in 13tSn [6] by a 2.4 MeV £2 tran- sition. The reduced £2 transition rates in 132Sn should be

435U 40406 4415 6 2 (40) (100) (13) ((7/2 3/2) J, 2, 6|7/2,(3/2 2) 7/2, 6) _|3/5 for J = 4 "(2/5 for J = 5 in units of the single-hole rate. The observed in- tensity ratio of the 2,380 keV and 2,268 keV gamma lines /(2,38O)//(2,268)= 1.6(2) makes it highly probable that these transitions pro- 132Sn ceed to the 4~ and 5" states, respectively. Con- I32 Fig. 1. Level scheme of Sn. Energies are given in keV. Relative figuration mixing effects are not likely to change this gamma intensities are given in parenthesis conclusion. The order of and even the spacing be- tween these two members of the (v/72, »'rfJ.J) multi- for the feeding of a 6" state at 7,210 keV, and also plet conform with the analogous case of the (vg9 -,, 208 supports the 3" assignment of the 4,351 keV state v/572') multiplet in Pb, where the 6" (3,919 keV) [3]. and 7" (4,037 keV) states are separated by The experiment was performed at ISOLDE, CERN, 118keV. where the indium activity was produced in an ura- The decay of the 4~ state is expected to go by a nium carbide target bombarded with 600 MeV pro- rather fast Ml transition to the lowest 3" state, tons [3]. Multispectrum analysis measurements were which is a collective mixture of many particle-hole performed with an 18% Ge(Li) detector. The experi- configurations [4], but with (v/7,2, vd^\) as one of ment showed a number of new transitions in ' 32Sn the main components. The only interpretation of the following the p~ decay of 132In (Fig. 1), and gave in observed 479 keV—4,351 keV cascade, which is con- addition the new value T1/2 =0.19(2) s for the half- sistent with firm theoretical predictions, is the life of ' 32In. yy coincidence measurements were per- 4~->3~->0+ alternative. Direct experimental proof formed with two Ge(Li) detectors with standard ef- of the 3~ assignment is, however, lacking. ficiencies 6 and 18%. Parallel production of 6.5d The p~ decay of 132In can also proceed by 1st i32 Cs, in a yield about four orders of magnitude forbidden transitions of the types ngg.[-*vh^l2, higher, severely complicated the experimental con- vf7/2-*nd5/1 and v/7/2->n;g7/2. The first of these ditions. Clear coincidences were, however, found for would mainly lead to the (v/7/2, v/i-,'/2)8, configu- the strongest transitions and from these data the ration, which is the dominant component of the T Bjornstad cl al.: Doubly Closed Shell Nucleus '^S 97

4,847 keV isomeric state. The observed feeding of ing pure (v/, ,. vh,*2) configurations. The good this slate gives log//= 5.2(1), which is a rather low agreement indicates that these states are indeed ,,ui nol unreasonable value for an unhindered 1st mainly of (v/,,2, v/ij",'2) character. forbidden transition. The log/r value of the elemen- The emerging picture of li2Sn is that of a nucleus l tary ng^'2->vh-, 2 transition in the /?" decay of with exceptionally strong shell closures. The first few '•"In [6] is unfortunately not known. The tran- negative-parity particle-hole excitations, collective as sitions to the other (v/7,2, v/if,'2)j states with J = 6,7 well as non-collective, occur at considerable higher should be considerably retarded by geometrical fac- energies in l32Sn than in 2O8Pb, even allowing for tors an A'1 ' scaling. A detailed comparison of the pro- perties of these two closely related nuclei should be «7/2 11/2) J. I, 7|7/2,(ll/2 1)9/2, 7)2 instructive. The results presented here are not com- liki for J = 6 plete. We hope that further experimental studies may be able to confirm the present interpretation = ,1 for 1 = 1 l32 and provide data on new levels in Sn. I mi for •/ = «• The other two types of 1st forbidden fi~ transitions 132 from the (v/,2, itgq\)-,- state of In are expected to go mainly to (Jtg7,2, ng^l),- and {nd5l2, ng^^t- References 132 states in Sn. The excitation energies of these 1. Kerek, A., Holm. G.B.. De Geer. L.-E., Borg. S.: Phys. Lett. + states are estimated to be about 5.5MeV(7 ) and 44B. 252(197.1/ 6.5 MeV(6*). Based on the known /J decay rate of 2. Lauppe. W.-D.. Sistemich. K.. Khan. T.A.. Lawin. H.. Sadler. O.. I33Sn [7] the branch into these two states should Selic, HA., Schult. O.W.B.: Proc. Int. Conf. Nuclear Structure. Tokyo 1977. J. Phys. Soc. Jpn 44. suppl. 335 (1978) together amount to about 10% of the total decay. 3. Bjornstad. T., De Geer. L.-E.. Hwan. G.T.. Hansen. P.G.. Jon- No evidence for this has been found in the present son. H., Kawade. K.. Kerek, A.. Lauppe. W.-D., Lawin, H.. experiment. Mattsson. S.. Sistemich. K.: Phys. Lett. 91 B, 35 (1980) The positive parity states in Fig. 1 are connected by 4. Dehesa, J.S.. Lauppe. W.-D.. Sislemich. K.. Speth. J.: Phys. £2 transitions. The experimental reduced rates of Lett. 74 B. 309(1978) + + + + 5. Brennan. M.H.. Bernstein. A.M.: Phys. Rev. 120. 927(1960) the 8*->6 , 6 ->4 and 4 ->2 transitions are 6. Dc Geer. L.-E.. Holm, G.B.: Phys. Rev. C22. 2163 (19801 compared in Table 1 with calculated values, assum- 7. Borg, S., Holm. GB, Rydberg. B.: Nucl. Phys. A212. 197 (19731 Table 1. Comparison or experimental and theoretical B(£2) rates, expressed in single-particle units B .(£2) = 40e2 fm4. The theoreti- T. Bjornstad B J. Blomqvist cal values are calculated for pure (v/1/2, v/if,'2) configurations, 2 2 G.T. Ewan assuming ecll = e and the radial matrix elements / = t = 33 fm2 B. Jonson K. Kawade Transition B{E2) B(£2) A. Kerek exp th S. Matlsson K. Sistemich 0.13 0.11 EP Division 0.28 0.24 CERN 4*-2* 0.41 0.19 CH-1211 Geneva 23 Switzerland PAPER P.XV. UNIVERSriYOF OSLO

THE CONSTRUCTION OF A FACILITY FOR 14 MeV NEUTRON ACTIVATION ANALYSIS OF OXYGEN IN ALUMINIUM

T. Bjørnstad , J. Alstad , T. Johannesen

1) Institute of Physics, University of Oslo 2) Department of Chemistry, University of Oslo 3) State Pollution Control Authority, Division Lower Telemark, Skien

Report 84-05 Received 6/2-84

INSTITUTEOFPHYSICS REPORTSERIES

ISSN 0332-5571 1.

ABSTRACT

The article describes the construction and performance of a facility for 14 HeV neutron activation analysis, n>ainly for oxygen in aluniniun, at the University of Oslo. It consists of a sealed tube neutron generator, a biaxial rotation system for samples in irradiation position, a nechanisn for autonatic separation of sample and container, and appropriate detectors and counting electronics. Especially considered are the proce- dures for sample preparation and oackin?, standardization, optimization of the energy range for integration and optimiza- tion of the irradiation, decay and counting tines. The applica- bility of the facility is exemplified by analytical results from comparative investigations in different laboratories using different methods. 2.

n. INTRODUCTION

In the metallurgical Industry the control of oxygen contents

In raw materials is important. In aluminium the oxygen content may take different forms: It may be distributed relatively homogeneously as small particles of aluminiun oxide (or hydroxide), so-called plankton, or it may exist as larger conglomerates of oxide particles. The disadvantages of a high oxygen content are of different kinds . For one thing, the hard oxide particles may cause damage to the tools during the machining of the metal. Then there are observed several quality- lowering effects. Larger conglomerates may give rise to brittle areas on sheets and foils and may even lead to streaks and tears.

The high elasticity needed in form pressing and extruding processes of profiles is reduced by an elevated oxygen content, and damage due to material exhaustion is more frequently experi- enced. Accordingly, one needs an extensive control at each step in the production process, in order to keep the oxygen content as lov/ as possible.

This has led to a call for precise and effective analytical procedures.

Several analytical methods have been developed for analysis of oxygen in aluminium, both of chemical ' and physical nature. Intercomparison of these methods often shows large differences in analytical results, and these differences enlarge as the oxygen content diminishes. In aluminium the oxygen level is often as low as 1-10 pom (and even below). In addition, some of the methods are rather time consuming. Improvements are 7—ft 1 offered by the various existing nuclear activation methods , where the ones using fast neutrons (mainly 14 MeV) generally are 3.

the cheapest and most versatile. Here, the lower Unit for quantitative determination of oxyaer. is presently about 1 ppm.

Review article* for 14 MeV instrumental neutron activation analysis (INNA) of oxygen in various materials (including aluminium) are found in ref.s , and for aluniniun matrices in particular in ref.s 17"20)•

The present report describes the 14 MeV INNA-technique developed and the facility built up for analysis of oxygen in aluminium in the authors' laboratory at the University of Oslo.

B. THE METHOD IN BRIEF

The analytical procedure is based on the nuclear reaction given in Table 1, which also lists the only two important disturbing reactions and gives some basic physical data.

Table 1. Essential nuclear reactions and sone physical data for the determination of the oxygen content by 14 MeV INNA

Reaction Isotopic Half-life Cross-section Radiation abundance (s) (mb) energy (MeV)

16O(n,p)16N 0.9976 7.13 39 + 4 a) Y: 6 .128, 7 .117 6 : 4.3, 10 .42

19 16 a) F(n,o) N 1.0000 22 t 5 »

11 11 b) B(n,p) Be 0.8102 13.6 3. 3 Tr: 2 .125, 4 .67, 7 .97 B~: 9.4, 11.5

b) At En = 14,5 MeV from ref. 21). From ref. 22).

The direct interferences may be corrected for. The possible fluorine content may be determined non-destructively by the 4.

19 19 reaction F(n,p) O (T, » 26.8 s, Ey: 0.198, 1.357 MeV) or by the reaction 19F(n,2n)I8F (T, - 109.8 min, B*, no v) by detecting the Sll keV annihilation radiation after chemical isolation of the fluorine. The content may be determined either by the reaction 11B(n,t«) 8Li (T, = 844o. E _ = 13 MeV, no Y) * 8" in fast runs with high-energy fl -measurements, or in neutron absorption measurements. Then the interfering contribution nay be subtracted from the gross counting number and the oxygen content derived. Experimentally it is found that the fluorine content needs to be i 2.6 tines the oxygen content (by weight) in order to yield the sane number of counts ' ', which is in agreement with the listed cross sections, while the corresponding number for the boron interference varies between 11 and 40 depending on the experimental conditions.

Hence the oxygen analysis experiences a minimum of direct spectral interferences, while indirect interferences like pile-up effects from the matrix element are to be considered more care- fully in the present work.

The analysing process comprises sample preparation, transport to the irradiation position, irradiation, transport back to the counting position with contingent post-treatment, counting and data handling. The next section describes the 14 MeV INNA techniques developed and applied in the authors' laboratory.

C. EXPERIMENTAL TECHNIQUES

C.I. The irradiation facility

The main apparatus in the analysis facility is a 14 MeV neutron generator with associated control electronics. The generator is 5.

at present of the Philips PK 5320 nodel and lias a ceramc accele- ration tube of the type 16604 with oil cooling of the damping syston. Our experience with this oem.-r.iior typo is "

The 18604 tube type produces a rather stable neutron flux with a total neutron output of 3-10 ns which gives a maximum 9 _i _2 flux of -. 2-10 ns en- . The generator tube is nounted in the irradiation cell in an upright position and shielded with successive layers of paraffin, paraf fir./boric acid mixture and concrete. An overview of the nenerator room is given in fig. 1.

The irradiation and subsequent counting is operated either manually or autonatically (see below) , ar.d the sanple is trans- ferred in a pneumatic rabbit systen designed to fit the chosen sample configuration.

C.2. Sample configuration.

Due to the neutron flux distribution fron the disc-shaped generator target the optimal sample shape is a half-sphere

(or close to it) for a given sample volume. However, the preparation of a large number of samples in this shape on a routine basis, and the automatic handling of them, is cunber- sone. Disc or cylinder shaped sanples are mere practical. For a given sample volume the disc would receive the hinher neutron flux. Still, after a total consideration, it was found appro- priate to standardize the analysing system on the cylindrical samples with the following dimensions: dianeter 12.5 mm, length

(height) 50.8 mm. The main reason for this choice of dimensions is the commercially available polyethylene capsules of that .•'.'••:>.V':"..'. NEUTRON DETECTOR •.'••v NEUTRON > SLIDING DOOR GENERATOR CONTROL PARAFFIN

SAFETY CONTROL WvS PARAFFIN / BORIC \ V:££ ACID \ 1

ENTRANCE

gpNCRETESHIELDINGV^va>^.^!:a:a;

O 0.5 1.0 1.5 20m

Fig. 1. Layout of the neutron generator room. The room is divi- ded by heavy shielding into an irradiation chamber where the neutron tube is installed and a section where the controll electronics are situated. The shielding on the walls and ceiling of the irradiation chamber con- sists of a 20 cm thick layer of paraffin followed by 20 cm of paraffin/boric acid mixture (44 % by weight of belie acid) and finally 100 cm of high-density concrete.

The radiation dose of neutrons at the highest exposed spot in the control room (close to the water inspection window) is ~ 0.5 mrem/h.

Various interlocks are connected to the generator ion- source for safe operation of the generator: warning lamps at different locations in the building (must be turned on), the sliding door at the entrance to the labyrinth leading into the irradiation cell (must be locked) and a microwave detector safety system for moving objects. particular size , to be used as cent a ir.crs. ' i tf. I i n i v.f'E

packings of tho containers , '-y i i ndr ici 1 s.'iT;'-1 c- r f sr^i

dimensions nay also be hand] c-ci.

C.3. Preparation of the sample surface

Aluniniun has a affinity for nxyoi--:, ^nd an o>: ide surface _ > layer is read i 1 y f o rncd a ven at air : ire ssuros rlovn to 1C torr . When the layer has arown to a certain thickness ,

further oxidation is prevented. The thickness o* this layer in air

at STP-condi t ions is exper inenta lly fount! te r anno between 2 "> 0 . b ..a/en and 2.0 . n/cm" of cxyner«. The 1 at tor va 1 ue

corresponds to a surface oxycen ancunt of 4r:.5 . ci for the

described sample, constitutinn 2.7 pnn. Th is is cf the sane

order as th« expected matrix content. The re•! a*_ ivo innortance

of this oxide layer can be dinin i shod sorewhaL bjt not drast i-

cally by the choice of other sanple cenf M-i.r.iMir.« .'fin. 2).

1.0 1.5 2.0 Sample radius (cm)

Fig. 2. Concentration of oxyqen ( in j-i-ro» i

th lck sii r f uco ox ide c f a:-. .:::.-: i :\ i I S il of the samp le d iir.cr. s U-:i for .i I'v.-r. t- 8.

Hence, the surface oxide nay introduce a considerable error if not recognized and corrected for.

Two procedures may be used. The first is to remove the surface

layer after irradiation, for instance by chenical etchinc. in this case a careful preparation and packinn of the samnle prior to irradiation is unnecessary. But such a procedure is tine consuming (1-2 half-lives are needed), and the detection limit will increase.

For routine analysis one may prefer a technique which combines a suitable surface treatment with an adequate packing procedure prior to irradiation.

Several methods of surface treatment have been reported in the

literature ' , including polishing with hl^O-^ or diamond

paste, machining with steel or diamond tools, and several types

of chemical pickling. The rethod whic.'i gives the best results,

and the one adopted by us, is compose.-j of the following steps:

a) Careful machining of the surface with a steel tool (depth of cutting 0.4 mm, feeding 0.008-0.04 nar./rotation, speed 1500 rpn) .

b) Chemical etching in a bath conposed of one-to-one (by volume) of HF (40 %) and HN03 (14.4 N) for 2 min at 20 °c. c) Six successive rinsings, three in destilled water and three

in methanol. d) The sample is dried in a varn nitrogen stream, and subsequently transferred to the packing.

After this treatment the surface oxide layer is reported to be reduced tc 0.22 - 0.28 „g en of o.xycen.

C.4. Sample packing

The surface oxide layer is not the only disturbing source,

during irratiation N from oxynen in the container or air 9.

adjacent to the sample surface nay recoil into the sample, 2 R ? 9 1 thereby raising the apparent oxygen content ' . This effect has been denonstrated in this laboratory '. A connarison between three series of irradiations has been performed: a) by packing the sample in air, b) in purified nitrogen, and c) by covering the internal walls of the capsule by a

25 in thick foil of stainless steel and packing in purified nitrogen. The results in table 4 show the importance of the nitrocren atmosphere in the packing process and also indicate the positive effect of the steel foil to stop N-recoils fror. the capsule wall itself.

Table 4. Apparent oxygen content of a particular sample after di f Cerent- packing procedures.

Packing nethod In air In N,-atm In N_,-atn with steel foil

Measured oxygen 23.8 i 2.6 13.5 * 1.9 9.4 • 2.4 cone, (ppm)

Therefore, for routine analysis, the samples are packed in polyethylene capsules which are heat sealed with a thin poly- ethylene foil, using a specially constructed sealing oven

(fig. 3 shows the principle). The whole operation is performed in a glovebox flushed with purified nitrogen.

The use of a recoil collecting foil may also be recommended for routine work. (The contribution from the recoiled N may be eliminated completely by renoval of ' 10 nr\ of the surface before the counting, for instance by chemical etchina. The necessity of such a procedure has to be considered in each actual case). 10.

- Polyethylene capsula with Al -sample

Polyethylene film

=• o- Brass ring L J~ " Brass casing

Thermocoax. cable Ventilation channel

Fig. 3. Packing and sealing procedure for an aluminium sample in a polyethylene capsule.

C. 5. The rabbit systen and the detection equipment.

Fig. 4 gives an illustration of parts of the analysing systen,

while fig. 5 shows in more detail the various components in the

detection equipment.

The sample is loaded into the system via the loading gate

(fig. 6) , and conveyed by compressed air (*^ 5.4 kp/cm ) into

the irradiation oosition. Due to the anisotropic flux distribution

of neutrons from the generator, spinning the sanple during

irradiation is advantageous.The rotation system developed is a

double axial device. A photocell detects the sample in its

irradiation position, allowing the rotation to start and the

generator ion source to be switched on (irradiation starts)

simultaneously. The sample is spinning both around its long

axis (15 rps) and around the short one (2.5 rps). This ensures 11.

Fig. 4. Parts of the experimental system for 14 MeV neutron acti- vation analysis: 1. rabbit tube to the irradiation posi- tion, 2. sample loading/unloading gate and separation system for sample and container, 3. sample guiding tube to the detection position, 4. high purity nitrogen gas reservoir, 5. glove box for packing of the sample, 6. lead shield, 10 cm thick, 7. two 3x3" Nal(Tl)-detec- tors, 8. and 12. magnetic valves for the compressed air, 9. valve to adjust the air cushion, 10. suction/ventila- tion, 11. HV-supplies, 13. compressed air, 14. multi- channel analyzer, 15. NIM-electronics, 16. printer, 17. plotter. 12.

BIAXIAL ROTATION SYSTEM MULT i-CABLE I i j COMPRESSED AIR RABBIT TUBE

M3 COMPRESSED AIR SN - SCAUR FOR THE NEUTRON DETECTOR RH = RATEMETER

ML, M2, M3 » MAGNETIC VALVES ND = NEUTRON DETECTOR

NGC = NEUTRON GENERATOR CONTROL DHV = BISTRIEUTIDN FOR THE HIGH TENSION RSC - ROTATION SYSTEM CONTROL HV = HIGH TENSION

TI, T2, T3 ••TIMERS FOR IRRADIATION-, DELAY- LAI, LA2 = LINEAR AMPLIFIERS AND COUNTING TIMES RESPECTIVELY SCA1, SCA2 = SINGLE CHANNEL ALALYSERS S * SCALER P = PLOTTER fiDG = GATE AND BELAY GENERATOR TT = TELETYPE

Fig. 5. Schematic diagram of the 14 MeV neutron activation system.

the best possible homogeneity in activating samples. The system

is shown schematically in fin. 7. At the end of the preset

irradiation time, the ion-source is switched off, and the

rotating bridge is stopped in its initial position before the

compressed air is again applied tc brina the sample back to the

loading gate. The latter device is simultaneously a separation

mechanism for the capsule and the sample. The capsule is stopped

by a steel knife cutting the sealing foil along the inner wall

of the capsule, such that the sample is ejected out of the 13.

CUTTING EDGE To DETECTION

\ CHANNEL FOR To EXCESS AIR IRRADIATION

Fig. 6. Close-up look of the loading/unloading gate and the separation mechanism for the sample and the container.

capsule into the transfer channel and is brought to the

counting position, while the foil renains attached to the

capsule. An adjustable air cushion slows down the speed, and

causes a vertical soft landing in front of the two 3"x3" NallTU-

detectors placed face to face inside a 10 en thick lead shield.

The brass tube that surrounds the sanple in the counting posi-

tion has a wall thickness of 2 nn in order to stop the high

energy ^-particles of N and fron Li ant? B, if present.

After a preselected decay tine (transport tine) the counting

starts. The transport tine fron irradiation to counting position

varies from -v. 1-2.5 s depending upon the sarmle weight. The

whole sequence nay be nanually controlled or entirely automa-

tically controlled by specially designed electronic circuits.

After counting, an air-pulse transports the sample back to the

loading gate for renoval. Built-in error nessages are important:

if the sample should rove out of position during irradiation,

or if the photocell lamp should extinguish, the analysis 14.

Fig. 7. The biaxial rotation mechanism in irradiation position: 1. electric motor, 2. clutch, 3. driving cog-wheel, 4. large rotating cog-wheel, 5. rotation indicator (microswitch), 6. rabbit tube, 7. electric contacts, 8. rotating bridge, 9. rotation start/stop push magnet, 10. microswitch sencors, 11. lamp emitting a focused light beam, 12. compressed air on/off relay, 13. large non-rotating cog-wheel, 14. rotating bridge cog-wheel, 15. compressed air supply, 16. light collector (sencor), 17. neutron generator, 18. ceiling-mounted support for the rotation system. 15.

sequence will be interrupted and the sample returned automati- cally.

C. 6. Standardization

The ideal standard should be as similar to the sample as possible in all respects. Differences in natrix composition, shape and size, will in principle make it necessary to correct for neutron attenuation during irradiation, for v-ray attenua- tion during counting, and for the difference in the average neutron flux and counting geometry.

A number of different standard designs have been reported:

1) pure perspex (CH- • C(CH,)CO2CH3), or drill-holes in supra- pure aluminium filled with air or a mixture of behenylalcohol

in docosan ((C22H46)! 17>' 2) carefully prepared

- mixtures of oxalic acid (COOH)2 2H2O) and graphite , 3) a nixture of aluminium powder and aluminium oxide (Al,0 ) encapsulated in aluminium , 4) pure sintered aluminium

' ,5) drill-holes in aluminium samples filled with threads 4) of nylon or stearic acid , 6) stearic acid or lead oxide (PbO) 4) in polyethylen capsules , 7) thin aluminium discs which are oxidized at high temperature, stacked together to a full sample thickness (disc-shaped) and encapsulated in aluminium, or where a material like aluminium oxide or phenacetin (C.-H. ,1JO,) is homogenously distributed between the discs , 8) a mixture of iron oxide (Fe2O3> and graphite pressed to pellets and loaded into a steel capsula ' , or 9) pure beryllium oxide (BeO) or silica (SiO2> . Some of the standards are designed in order to avoid neutron and y-attenuation, while others are not.

In our laboratory several standardization orocedures have been 16.

Table 2. Oxyqmn standards evaluated and applied In the author«' laboratory.

Standard Compos lt ion/chai Design Application Oxygen content formula

Pur« Perspex, Massive cylinder General oxygen 320 000 ppB CH2-C(CH3ICO2CH3 standardization

Mixture of oxalic acid and Powdar mixture encapsulated Varying« a feu

graphite,

Mixture of iron oxide and Pallets of — 1 ma thickness and Varying, a few graphite, Fe-O, /C diameter 10 MI stacked to cy- hundred to a few linder lengths of — 48 mm and thousand ppn encapsulated in high purity Al-cylinders

Steels« National Bureau of Massive rods of length - 50 mm Standards Material (NBS), and the following diameters: British Standards Material (BS) 6.35 M (NBS) 492 ppM 7.94 MM (NBS) 132 ppM 12.70 a« (BSI 103 ppM The former two were encapsulated in high-purity Al-cylinder*.

Alusiniua »atal and various Massive Al-cylinder with consen- For Al samples Various. oxygen-containing additives trie, equally spaced snail-dia- with a high to Example-. With meter holes in the axia: direc- moderate oxygen 12 holes of dumetcr- tion filled with oxygon-contai- content 0.2 m, fUling Ma- ning naterials c terial: Perspcx: 362 pjw Aix: 279 ppn

Oxidized aluminium metal Hollow high-purity Al-cylinders length 50.0 mm, wall thickness 0.4 MM and variable diameter oxidation: to fit exactly into each other - 100 ppm to a normal full-size sample. All surfaces oxidized in air at 550 °C for variable time.c'

Aluminium Metal Massive cylinder of normal Mainly for high- 138 ± 10 ppM sample size, oxygen content de- purity Al-samples termined by other primary standards.

*' P.S. - primary standard« S.S. - secondary standard

For samples with a high oxygen content (i.e. the contribution of the counting statistics to the analysis error is negligible): Correction factors may have to be evaluated taking into account differences in shape and composition of the standard and the sample. See the text. For samples with a low oxygen content (i.e. the counting statistics dominates the analysis error): Special corrections may be simplified or omitted.

clFurther description in the text.

'standard nearly identical with the sample both with respect to shape and macro composition which makes further corrections superfluoures. 17.

considered and evaluated, - different designs for different areas of application. Those standards are described in table 2.

The standards 1-4 differ f ror. the normal aluminium samples in macro conposition and partly in shape and size. For samples with a high oxygen content it will therefore be necessary, as mentio- ned above, to evaluate a nunber of correction factors described below.

Due to the generator stability (no short tine variations) , it is not essential to irradiate the sample and the standard simultaneously. A neutron detector installed in a low-geonetry position and calibrated against the standard, is used as a flux-nonitor for each irradiation. The oxygen content nay then be found fron the fornula

w R ; I C D

R C C D BF3 x *s x x where

w = the nass of the oxvoen in the samnle x w = the mass of the oxygen in the standard,

K = the ratio R /(R _ ) where R and (R ) are Ut ~ s DC - S S Ot ~ 5 the number of N-counts recorded for the standard

and the nunber of neutron counts in BF,-detector

during the irradiation of the standard, respectively,

Rx = the backaround-corrected nunber of N-counts for the sanple,

R__ = the recorded number of neutron counts in the BF,- BF3 3 detector during irradiation of the sanple,

- = the average neutron flux in the void space occupied

by the oxygen-containing part of the standard Inure

geonetrical effect), 18-

4 « the average neutron flux in the void space occupied

by the sample.

I /I - the relation between the number of neutron counts X S

in the BF3-detector with the sanple and with the standard, respectively, v/hile in the irradiation

position,

C /C = (C /C ) • (C /C ) where C is a "transmission 5 X S X TI S X Y S factor" which takes into account the attenuation

of neutrons in the standard during irradiation

(index n) and the ^-rays during the counting (index i),

and C is the sane for the sanple,

D /D = the ratio of detection efficiencies for the tl

•y-rays from the standard and the satfple, respectively

(pure geometrical effect). The correction factors C , C , I and I nust be experimentally S X 5 X determined for each matrix type. In order to essentially avoid any of the corrections mentioned

in eqn. 1 for aluminium samples, standard no. 5 has been evaluated.

It consists of a high purity aluminium sample where thin axial drill-holes placed equally spaced and concentric in the cylinder

are filled with a chosen standard material. The drill-holes are

positioned to receive the average flux of the whole sample. The

positioning of the drill-holes was evaluated from an experiment

in which the radial and axial flux distribution of the sample

was determined. A full-size sample was built up of nine equally

long sections and each section was composed of eight concentric

rings of a wall thickness 0.6 - 0.7 nn and a centerpin of

tfianeter 1.9 ran. This compact samples was irradiated under

"normal" conditions to produce Na (t. = 15 h, Y-2754 keV) 19.

in the reaction 7A1 (n,'i) Na. Fror. the specific Na-activity

in each of the sinrlo pieces after irradiation cne may construct

relative flux-distribution curves like those in fio. 8.

f1 2 3456789

"0 1 2 3 4 5 6 RAW«. DISTANCE FROM CYLWDER CENTER ( mm )

Fig. 8. Example of relative flux-distribution curves along the radial and axial direction across a cylindrical sample. 20.

Let A. be the specific activity in cylinder section j. The corresponding radius r. nay then be found Graphically as illustrated in fig. 8.

A full size standard sample nay now be built up of a number of sections when each section j has individually drilled holes corresponding to r..

However, in order to reduce the error introduced due to surface air oxidation, one single nassive cylinder nay be preferred.

Let A be the average specific activity for the whole cylinder.

The corresponding average radius r for positioning of drill- holes may then be approximated by a weirhcd value of the indivi- dual r.-values

r s= -^- • Z r . • A. (2) n-A j=l 3 3 where n is the number of identical sections of the sarnie.

In our case n=9 and r=3.8 nm.

The hole diameter will in practice be a few tenths of a ran.

Standards with different oxygen content may then be obtained by varying the number of holes and the oxygen content in the filling standard naterial.

Another standard (no. 6) intended for use in order to make all attenuation corrections unnecessary even for a high oxygen content, of the aluminium sample, is briefly described. It 30) is in principle similar to the one nentioned in point 7) above , modified to fit cylindrical sanples. Hollow cylinders with wall thickness 0.4 mr. and of different dianeters to fit exactly into each other, was built up to a sample of normal size where the outer cylinder functions as a capsule for the others. After surface preparation, the individual cylinders were oyiilized at 21.

high temperatures (^ 555 °C) for some hours. The increase in weight is supposed to be due exclusively to a uniform formation of aluminium oxide on the surfaces. This standard is not yet in practical operation because the first attenpt to produce a standard curve from a series of three samples failed , but the method is still considered accomplishable.

For routine analysis of high-purity alumiuin samples we use a secondary standard consisting of an aluminium sample of normal size (no. 7) whose oxygen content has been determined by primary standards of type 1, 2 and 4 (table 2). The procedure seems to serve the purpose well. A good knowledge of the background allows us rather safely to apply as a calibration curve the line between the measured point and zero (origo).

C. 7. Optimization of the method.

The optimization criterion should express generally accepted quality probing quantities like the relative standard deviation, o , or the lower limit of detection (L -value as defined by

Currie 33)).

Here we have chosen as optimization criterion the expression for oR:

= (ST where

S = the total number of counts within a gated energy region,

S = the background counts within the sane enercty renion,

S = the net nunber of counts within the sane region.

This expression nay be minimized by giving special consideration to experimental parameters like the sanple shape and volune, 22.

the energy range for integration and the irradiation, decay and

counting times.

Taczanowski 34) has treated the first point for cylindrical

samples. Since sample shape and irradiation qeonetry is fixed

for practical reasons in our laboratory, it will e\ot be further

discussed here.

The other naraneters are briefly discussed below.

C. 7.1. Eneray range for integration.

A pure N ,-ray spectrum from Nal(Tl)-detector has the form

indicated in f ia. 9. The escape peaks dominate the spectrum.

16 % LU uj 2 a z (/) <2 •' ro' . <£•> in CD A a> i z II ^ ^

1 2 3 4 5 6 7 Energy (MeV)

Fig. 9. A pure Y-ray spectrum of N accumulated with a 3x3" Nal(Tl)-detector. Hatched area of interest for activity measurement.

Due to the high -,-energies, the direct spectral interferences

are negligible (except those mentioned in table 1). Hence, a

relatively broad energy region around the escape peaks nay be

integrated safely. There are, however, indirect interferences 23.

f ror1 the general back" ro une. and f rer. ri ie-ur: o f feet E of tr.e tv/o 2 4 ^7 mat r i;: act i vi ties Na ', \ r> h) an ti *" Mf '^.46 ri in t s i rul taneously produced in (n, t J - and ' n,pi-react ions, respect i ve I1/; the tv/c 24 coinc ident -. -rays of Na at 1366 koV ane 2704 keV r-ivc r ist.- to a sun-peak at 4122 ko V. Acc ido p. tal coincidences within the detect ion sys ten resolving tine betv.-oor. th is

For N and the neneral backorour.d the probaLi 1 if curve? will essentially have a constant shape for al] r:\ict ical ac c. i vit y

levels. for the variable back c re und the s Lape nay vari- sonewi-.at "'A "* 7 with the relative natrix activity of ~ TJa and *" ,Vn, Their activities depend upon the number of irradiation-decav-countinrj cycles (in this text called accumulation cycles) and the various tines involved. The shape of the probabi1 i ty curve of the various contributors to the actual enerny rosier, is shown nual itat ively in fin. 10.

The problem i s now to find the opt inur nfitmrs of the upper and lov/er discriminators of the single channel analyzers.

The answer depends en several factors. For a Tiven analytical sys ten and procedure , the ert i nun sett- in f s are obtai rier in .» three-dimensional ratrix. The axis are defined Ly i) the oxyqen content (net nunber of '^-counts) , M the ac:.:vit_- of "4Na an^i c) the act i v i ty of *" MT . If <-.' i f foront •:;<"• unt i ^-i i i no s are used

for the ti iff ere r, t oxyncn cori tent p , \'r,:*-ror.nci wc; 11! de fine a f o ur t r. axis. 24.

A. General back- ground. B. Variable (pile-up) background C. 16N-spectrum

Energy (MeV)

Fig. 10. Qualitative picture of the N v-ray spectrum around the photopeaks and their escape peaks, indicating the back- ground interferences.

A system dependent natrix is determined in the following way:

a) Establish the probability curves for each of the contributors

by recording with good statistics a pure N-spectrum, a

number of pile-up spectra with varying but well-known 24 27 activities of Na and Mg and a general backnround spectrum.

Smooth the curves with a suitable procedure 35)

b) As an optimization criterion use the expression for the

relative standard deviation:

S rel -V ' 16N

where S is the net number of counts of the various contributors

between the discriminator levels. Choose a set of S-values,

fold them with their respective probability curves and

calculate ~re^ for a nunber cf corbinations of upper and 25.

lower discrininator in the resultina "spectra". The setting

corresponding to the mininun value of ••• is taken to be

the optinun combination for this set of £>-values.

The sane calculation nay be repeated for any set of S-values

and, thus, the optinization natrix nay be deterrined.

A general qualitative picture of the variation of the optirnun

discrininator settings as a function of the net N-counts and

the natrix activity is shown in fig. 11. The lower discriminator

180

Lower discriminator level 160 Upper •• •• — ui 0 140

1 120

Irradiation (accumulation cyclus) z o number o z 80 hi o 60

20

4.0 4.6 5.0 5.2 6.3 6.5 7.5 ENERGY (MeV)

Fig. 11. Qualitative picture of the variation in the optimum setting of the lower and higher discriminator as a func- tion of the net N-counts and the matrix activity. 26.

level increases with decreasing oxygen amount and increasing matrix activity. The bunching together of the curves betv;een

5.0 and 5.1 MeV is due to the start of the double escape peak fron the 6.13 MeV .-ray. The higher discriminator level is practically insensitive to the matrix activity, but increases with increasing oxygen content. The steep raise of the curve around 7.2 MeV is due to the decline of the 7.12 MeV ,-neak.

The continued increase of the discriminator level above 7.12

MeV is due to pile-ups on the N-spectrum.

Since our analysing system electronics are tuned by hand, it is not practical to readjust the discriminator levels for each different sanple and number of replicate irradiation-counting sequences. In routine analysis we normally choose the lower level at 4.3-4.5 MeV and the upper level at 7.0 MeV.

C. 7.2 The irradiation, decay and counting tines

The problem of optimizing the experimental times in activation analysis is sparingly studied, but a few literature reports exist on the subject . None of them can be generally applied. They are partly based on artificially constructed criteria ' , partly designed to calculate the optimum value of the irradiation, decay and counting times separately

' or the optimum combination of the irradiation and decay times ' only, or is constructed for experimental conditions different from ours '

Accordingly, for our specific problem we had to develop the optimization procedure ourselves.

In the following let the subscript numbers 1, 2 and 3 symbolize the nuclides N, Na and Mg, respectively. 27.

By irradiating aluminium samples we have the following relation-

ship between counting rate R within the gated energy region u.3 - 7,0 MeV and the desintegration rate D:

R2 = k2 • D2

R3 = k3 • D2 • D3

Here, k and k, are system dependent constants with dinension s.

Their numerical values (table 3) were determined in a series of

irradiations of aluniniun sanples. Introduction of the detection

probability e = R/D yields

C2 = k.

k3 • D2

Minimization calculations have been carried out by a DEC-10 computer using prograns written in BASIC. Results on optimum combinations of t. and t are niven in ficr. 12 for oxyoen con- l c " •* - tents fron 10 to 100 ppn and for 1 to 10 successive accumulation cycles, with a total cycle time of r = 600 s. One observes, for

instance, that the optimum irradiation time, t. (opt), at a

10 ppm oxygen level decreases from 23 s for 1 cycle to 11 s for

10 cycles. This is due to the building up of the variable back-

ground activity, and the fact that c2 and e3 increases linearly 24 with the desintegration rate of Na.

Fig. 13 shows the relative standard deviation at the optimum

combination of t. and t for oxygen levels from 10 to 100 ppn

and for 1 to 10 accumulation cycles. Fron this figure one nay

find the number of accumulation cycles reasonable to perform

at each oxygen level. If one demands, for instance, that

(orel(q)-orel(q+l))/orel(q)^0.1, the reasonable value of q

would be 3 to 4. 28.

Table 3. Scheme of necessary information for the optimization of the irradiation/ decay and counting times: Physical com tants, system dependant constants and fixed vari- ables« explicit minimization formula.

Physical constants of importance for the variable background activity: 24Na 27MC Nuclear reactions 27Al(n,o)24Na 27Al(n,p)27Mg Cross section (sib) 116 ±3 75 * 8a) Half-life (s) 54000 567.6 Dominant Y-rays (keV) 1369, 2754 844, 1014

Constants and variables with fixed values: Number of Al target nuclides per sample, N: 3.68-1023 Average neutron flux across the sample volume, 4: 4.5-108 s"1 cm"2 Decay time, t,: 3 s

General background cound rate (4.3-7.0 MeV), RQB: 0.21 cps Counting efficiency of 16N (4.3-7.0 MeV), c^ 0.045 Value of k- (2*Na, 4.3-7.0 MeV): 7.98-10"10 s 27 11 Value of k3 ( Mg, 4.3-7.0 MeV): 8.58-10" s

Explicit minimization formula:

x=0 2 2 "3 2 -Pi2-Pc2- l ( l Pd2» j=1 x=0 ? «V x=0

where

c d Pcn - d-e » , , P(Jn . e "

r = the total duration of one accumulation cycle. (r = tA + t^ + tc + extra handling and waitin q = number of succeeding accumulation cycles.

a) For 14.5 MeV neutrons from ref. 46). b) From ref. 47) 29.

1 1 . . | . • •« 1 < 1 | •

L i CYCLE y 2 CYCLES 3 CYCLES / ] ^ \

— —— • l . . 1 i .ii l 1 l . . i. ...1 - CYCLES S 5 CYCLES > 6 CYCLES S \

/^ 20 -"" -. \ - "TtopO "

to • • — • , i . . i .... I ...1 :

7 CYCLES S 8 CYCLESy> 10 CYCLES

^ \

^^ ' i .—'^

-" — —. •

1 . . 1 .... 1 10 20 50 100 20 50 100 20 50 100 OXYGEN CONTENT ( ppm )

Fig. 12. Average optimum combinations of the irradiation time,

t^lopt), and the counting time, tc(opt) as a function of the oxygen content with the number of accumulation cycles as parameter. The "reasonable" value of the irra- diation time, t.(reas.k,and the corresponding counting time, t (reas.)as defined in the text, are also given. The total cycle time is 600 s. 30.

OPTIMUM COMBINATION OF 22 - IRRADIATION TIME.t, . ANO COUNTING TIME tc DECAY TIME. t„.i3s -• 20 TOTAL CYCLE TIME = 600s

ae 18 i I- 2 16 i u 12 i — lOppm j£ 10 i. I f -20—-

2

0 2 i 6 8 10 NUMBER OF CYCLES

Fig. 13. Relative standard deviation corresponding to t±(opt) and t (opt) as a function of the number of cycles with the oxygen content as parameter. The total cycle time is 600 s. 31.

For routine analysis it may be more important to keep the ana- lysis low in cost at a controlled, reasonable standard devia- tion level than to obtain the real optimum conditions. This may be achieved by reducing the irradiation time to a "rea- sonable" value. This value, t Ireas.), may be defined, for in- stance, as the lowest possible value of ti in order to obtain a standard deviation J , (reas.l, which lsio % higher than the minimum value r ^(mini. The t (reas.) nas a corresponding value of the counting time, t (reas.l.

Such reasonable combinations of t and t are also calculated and

plotted in fig. 12. In this way the irradiation time is reduced

by nearly 50 i.

For experimental set-ups where the applied electronics make it

unpractical to readjust the times for each sample, one may, in

average, recommend an irradiation time of ~ 10 s and a corres-

ponding counting time of ~ 25 s. 32.

D. APPLICATIONS

A few examples of the applicability are given below. For the

reason of comparison the analytical facility described here

has been applied to analysa oxygen in a set of 16 aluminium

samples previously analysed commercially by 14 Hev neutron ac-

tivation (NA) at Gulf General Atomic (GGA)12' in USA. The sample

shaoe and size have been kept the same, but have in the present

investigation been treated according to the procedure described

previously. The results are given in table 4.

Table 4. Results from 14 MeV neutron activation analysis of oxygen in aluminium samples by two different laboratories.

Sample Present GGA- Sample Present GGA- no values values no. values values (ppm) (ppm) (ppm) (pom)

1 6.0*2.0(4)a) 10.H2 .4 9 3.0H .0(41 8.74*2.2 2 7.0*2.516) 7.83*2 .0 10 5.012.0(11 5.2 11.7 3 10.0±3.0(6l 22.1*2 .9 11 6.0i1.5<5) 3.6611.4 4 111.0i7.0(1) 139 18 .9 12 5.512.5(4) 0.28b) 5 6.5*2.5(5) 11.6i2 .6 13 5.011.5(4) 2.1011.1 6 9.5i2.0(5) 30.0*4 .3 14 3.0i2.0(1) 2.2 11.3 7 9.512.0(51 18.513 .2 15 3614.5(1) 52 *6.5 8 4.0*2.0(1) 36.7i3 .7 16 138*10 138 19.1

The numbers have been rounded upwards to the nearest 0.5 ppm. The parantheses indicate the number of parallel determinations.

Upper limit.

Except for a few agreeing numbers, the GGA-values are in general

higher than ours. The explanation may partly be the different sur-

face treatment and packing procedure, but mainly that the GGA-pro-

cedure did not include separation of the sample and the container

before counting, i.e. the sample was irradiated and counted in 33.

the same polyethylene capsule . Since the oxygen content in the

capsules (up to several mgl may vary, a proper blank correction

may be difficult. Accordingly, we have taken our values to repre-

sent the actual bulk oxygen content more accurately.

The facilities described here is now routinely and successfully

used to perform fast and reliable analysis of the oxygen content

in aluminium samples for the Norwegian Light Metals Industry, but

also for research institutions. The present technique has,for

instance, been used as a standardization method to examine the

applicability of a new chemical method for characterization of

inclusions in aluminium and aluminium-based alloyo. The che-

mical method includes dissolution of the bulk material in acids.

The dissolution residue is assumed to consist mainly of A1,O,-

and Al.rtgO -particles, and the amount of aluminium and

was determined by X-ray fluorescence spectroscopy (XF). Results

from the comparative study is shown in table 5.

Table 5. Oxygen analysis of aluminium-based samples;- comparison of 14 MeV NA and a chemical method based on acid disso- lution and X-ray fluorescence analysis of undissolved oxide particles.

Sample NA XF no.a) ppm 0, ppm oxides , measured, ppm. estitn. values , ppm, measured calculated Hg Al Si A12M

3A-1 4.5*2.0 8.5*3.8 3A 2 4.4*1.2 8.3*2.3 - 2.4 0.9 4.5 - 3A 3 4.3*1.2 8.1*2.3 L1 3.0*1.1 5.7*2.1 0. 15 1.4 0.3 2.1 0.9 L2 3.3*1.5 6.2*2.8 0.25 1.0 0.2 0.9 1 .5 L3 4.5*1.5 9.1*9.8 0.17 6.4 0.1 11.5 1 .0 L4 3.6*1.4 6.8*2.6 0. 10 0.7 0.5 0.9 0.6

3A-samples contain 0.08 wt % Si and 0.5 wt % Fe L-samples contain 1.0 wt % Mg, 1.0 wt % Mn and 0.15 wt % Cu. b) The oxygen is supposed to be bound as AI.O . 34.

Except for the value of sample L3, the chemical values are gene- rally lower than those found by the NA-method. The explanation is

and A1 M are partly that some ^2°2 2 9°4 dissolved during the acid

treatment, thus escaping detection by XF-analysis. The Si-content

is argued to be mainly 3-S1, but any SiO, present will increase

the chemical values. Finally, any surface oxide and recoiled N-

particles will tend to increase the apparent bulk oxygen content.

The surface effect has not been quantisized in these experiments,

but is probably not decisive.

The equipment is basically constructed to be applicable for ana-

lysis of a number of elements in various matrix systems. Hence,

samples of magnesium, steels and various alloys have been analysed

for oxygen with only small adjustments in the analytical procedure

and minor mechanical modifications (see example of results in

table 6).

Table 6. Analysis of oxides in magnesium by 14 MeV NA and a che- mical method based on dissolution in methanol and X-ray fluorescence spectroscopy (XF) on the residue.

Sample NA XFa) no. ppm O ppm MgO ppm Hg ppm MgO measured calculated measured calculated

1 9.4i1.,3 23..7±3.,3 15.2 25.2

3-1 15.3±1,.7 38..5*4.,3

3-2 11 .4*3 .2 28,,7*8., 1 18.2 30.2

3-3 11 ,0i3.4 27,,7*8.,6

Values taken from ref. 49.

Likewise, metal powders, salts, oils and various liquid solutions

have been analysed for oxygen and other components. The determina-

tion of the protein content (proportional to the nitrogen content)

m various food ingredients is also a mam area of application. 35.

E. CONCLUSION AND OUTLOOK

The described facility has served its purpose well. Still there is room for improvements, and it is at present under considera- tion for upgrading and partly reconstruction. The future system will include possibilities to handle also discshaped samples. A commercial programmable microprocessorbased electronic control unit will replace the home-made control unit now used and all timers in the system. This change will bring about a higher flexi- bility in designing particular analytical procedures for instance by allowing the optimum combination of the experimental times t., t, and t to be automatically adjusted for each individual accumu- lation cycle. In addition, it will be possible to include a com- puterbased multichannel analyzer (with CAMAC link) directly into the automatically operated analytical sequence. Recording and ana- lysis of the accumulated spectra with sequential calculations may be initiated by the control unit.

The described future improvements will lead tc a smoother ana- lytical sequence as well as to qualitative improvements in the analytical results.

The analytical service for the metallurgical industry and research institutions is foreseen to continue, and other applications de- manding either high resolution .-detectors or equipment for detec- tion of --particles or delayed neutrons more extensively investi- gated. 36,

Acknowledgement.

The authors gratefully acknowledge valuable help and advices received during the construction and build-up of the irradiation/

eountir.g facility froir. the late Dr. E. Kvåle, the late

IT. h. Fergerce-, Dr. E. Fjc-lls*ac, Mr. E. BarTh and

Dr. C.J. Simensen has kindly permitted to refer his data on

X-ray fluorescence measurements on some aluminium and magnesium samples.

Financial support has been obtained from the Elkem Research

Foundation and the Astrup's Foundation. 37.

References:

1. G. Kraft, EURISOTOP - 59, (1971).

2. G. Kraft, EURISOTOP - 71, (1972).

3. H. Aschehoug, Tidsskr. Kjemi, Bergv., Metallurgi, 3J),26 (1970).

4. R. Sulzberger, Erzmetall, 2_4,123 11971).

5. J. Fischer, W. Kastner and H. Aschehoug, Giesseroi, 58, 680 (1971)

6. E. Grallath, Fresenius 2. Anal. Chem., 300, 97 (1980).

7. C. Vandecasteele, A. Speecke and J. Hoste, EURISOTOP - 68,(1972).

8. C. Engelraann, EURISOTOP - 74, (1973).

9. E.L. Stcele and W.W. Meinke, Anal. Chen., 34, 185 (1962).

10. D.J. Voal and C.F. Cook, Anal. Cher-.., 34, 178 (1962).

11. O.U. Anders and D.W. Briden, Anal. Che»., 36, 287 (1964).

12. V.P. Guinn and D.E. Bryan, Report GA-9733, (19b

13. E. Karttunen and D.G. Gardner, Int. J. App) . Rad. Isotopes, 20, 801 (1969).

14. R.G. Clark and W.A. Stensland, J. Radior.al. Chem. , 4, 365 (1970).

15. B.F. Schnitt, Erzmetall, 26, 78 (1973).

16. H. Forster, Isotopenpraxis, 1_5, 184 (19791.

17. D. Gibbons, G. Olive, D. Swicr and J.E. Deutschman, J. Inst. Metals, 95, 280 (1967).

18. K. Cottling and H. Vogg, Report KFK-1294, (1970).

19. C.E. Espanol and A.M. Marafuschi, Talanta, 17, 653 (1970).

20. K. Takada, I. Fujii, M. Ishihara and H. Muto, Anal. Chem. Acta, 5j>, 19 (19 711.

21. V. Krivan and K. Krivan, J. Radioanal. Chom., 29, 145 11976).

22. A. Volborth and H.E. Banta, Anal, chem., 3_5 , 2203 (19631.

23. G. Beurtsn and R. Pillon, J. Radioanal. Chem., 40, 189 (1977).

24. J. Hoste, D.Dc Soete and A. Speecke, EUR. 3565e, (1967). 38.

25. L.D. Klu'wer, Thesis, University of Oslo 11974).

26. P. Boisot, M. Cuypers, G. Kraft, J. Pauwels, L. Quaglia and G. Weber, Report ITE - 76 - E, (1972).

27. L. Quaglia, G. Weber, D. Davids, J. van Audenhove and J. Pauwels, Report ITE - 90, (1976) .

28. F. Dugain, M. André and A. Speecke, Radiochem. Radioanal. Letters, 4, 35 (1970).

29. F. Dugain and C. Michaut, Radiochem. Radioanal. Letters, 9, 119 (1972).

30) T. Johannessen, Thesis, University of Oslo, (1978).

31. D. Brune and K. Jirlow, J. Radioanal. Chem., 2, 49 (1969).

32. R. Gijbels, A. Speecke and J. Hoste, Anal. Chim. Acta, 4_3, 183 (1968).

33. L.A. Currie: Anal. Chem.,4_q, 586 (1568).

34. S. Taczanowski, J. Radioanai. Chem., 12, 535 (1972).

35. A. Savitzky and M.S.E. Golay, Anal. Chem., 9J>, 1627 (19641.

36. M. Okada; Anal. Chin. Acta,24, 410 (1961).

17. T.L. Isenhour and G.H. Morrison; hn&l. Chem.,36, 1089 (1964).

38. T.L. Isenhour, C.A. Evans jr. and H.H. Morrison_, Proc. Conf. Modern Trends in Activation Analysis (College Station, Texas, 1965), p. 123.

39. Yu.A. Tsirlin, A.Ya. Gelfman and E.P. Mokhir, 2. Analit. Khim.j 2_1, 1012 (1966) .

40. P. Quittner and A. .Montvai, Acta Chim. Acad. Scient. Hung. 51, 371 (1967).

41. H. Fedoroff, Nucl. Instr. Meth..,9J, 173 (1971).

42. M. Fedoroff, J. Radioanal. Chem.j 1_5_, 435 (1973).

43. L. Zikovsky, J. Radioanal. Chem., 22^, 165 (1974).

44. J.I. Watterson, J. Sadioanal. Chem., 26, 135 (1975). 39.

45. Yu.N. Bourmistenko, I.N. Ivanov, V.V. Sviridova and Yu.V. Feoktistof, J. Radioanal. Chen., 40, 147 (1977).

46. M. Bormann, H. Neuert and W. Scobel, in "Handbook on Nuclear Avtivation Cross-Sections", IAEA Technical Report Series no. 156, 1974, p. 87.

47. G. Friedlander, J.W. Kennedy ana J.M. Miller: "Nuclear and Radiochemistry", John Wiley & Sons, 2nd edition, p. 175 (1964) .

4 3. C.J. Simensen and G. Strand, Fresenius Z. Anal. Chem., 308, 11 (1981) .

49. C.J. Simensen and A.I. Spjelkavik, Fresenius Z. Anal. Chem., 300, 177 (1980). PAPER P.XVI. UNIVERSITYOF OSLO

PRODUCTION OF 7Ga AT THE OSLO CYCLOTRON

by

T. Bjørnstad and T. Holtebekk Institute of Physics, University of Oslo, Blindern, Oslo 3, Norway

Report 83-I Received 5/1-1983 k.

INSTITUTE OFPHYSICS REPOKTSERIES

ISSN 0332-5571 Abstract

A method for production of Ga at the Oslo Cyclotron is described. The method is based on the nuclear reaction Zn(p,2n) Ga. The target is natural zinc metal of thickness 1.3 mm fixed by a thin alloy layer to a copper disc for efficient cooling during irradiation. By applying a beam of 29 MeV protons, a maximum production yield of ~1.8 mCi/pAh was ob- tained. By demanding a contamination level of Ga i 1%, the "useful" yield after a decaytime of 88 h is ~ 0.8 mCi/pAh.

Gallium has been separated carrierfree from the zinc matrix by cation exchange from 7.5 M hydrocloric acid solutions and prepared as a citrate complex at pH 5.5. After sterile filtering, autoclavation, pyrogene testing and analysis for iron and zinc, the Ga-radiopharmaceutical has been applied in human investigations at the Ullevål hospital in Oslo. INTRODUCTION In 1949 reactor produced radioisotopes of gallium were for the first time considered and evaluated for possible medical applications . When 72 gallium-lactate containing Ga as a tracer was injected in animals, a certain deposition in bones was found to take place.

72 In order to overcome the disadvantages of Ga such as low specific activity and short half-life, the idea of using carrierfree Ga was introduced. This was obtained in 1953. Gallium produced by proton bom- bardment of natural zinc was prepared as a citrate complex and used in 2) animal experiments . However, these experiments showed less bone uptake than the previous ones. Not until 16 years later it was realized that carrierfree gallium is sequestered by the plasma, and enhanced bone deposition will not be seen until the plasma is saturated with gallium 3).

During these experiments the same group accidentally observed that Ga-citrate concentrates in soft tissue tumors . This discovery marked the start of a new development of radiogallium in nuclear medicine, and Ga-citrate has later been recognized and gained reputation as a soft tissue tumor localizing agent (for more detailed information see ref. 5) and ref.s therein). Its physical half-life of 78.26 h and its decay by pure EC followed by the main y-ray energies of 93.3, 184.5 and 300.2 keV makes it suitable for in-vivo use, and it is readily detected by com- mercial gamma-cameras.

Today, Ga is produced regularly by different methods at several accele- rator facilities throughout the world. The present article describes the production technique developed at the Oslo cyclotron.

PRODUCTION METHODS

At the particle energies available at small cyclotrons carrierfree Ga may be obtained by: a) proton bombardment of natural or enriched zinc, b) deuteron bombardment of natural or enriched zinc, c) alpha-particle bombardment of copper, d) He bombardment of copper and e) indirectly by alpha-particle bombardment of natural or enriched zinc. The main nuclear reactions involved are listed in table 1 together with some results reported in the litterature. TABLE 1. REPORTED PRODUCTION METHODS FOR 67Ga WITH A SMALL CYCLOTRON

NUCLEAR TARGET TARGET BEAM BEAM PRODUCTION REFERENCES REACTION TYPE THICKNESS ENERGY CURRENT YIELD mg/cm2(ran) MeV MA MBq/pAh (pC

ft 67 Zn(p,2n) Ga Nat. Zn, 99.99%, 140-240 21 235 12.6 (340) 6) electrolysed on Ag (0.2-0.33)

Zn(p,2n) Ga Flattened Zn-tube 140(0.2) 21 350 15.9 (430) 6)

Zn(p,2n) 7Ga Enriched 68Zn, 98.46%, 135(0.19) 22 350 46.3 (1250) 7) electroplated on Ni(Cu)

Zn(p,2n) Ga Nat. Zn electroplated 22 20.4 (550) 8) on Cu

6ft A 7 200(0.3) 21 30-40 20.4-22.2 9) Zn(p,2n) Ga Nat. Zn or enriched (550-660) Zn electrolysed on Au(Cu)

66Zn(d,n)67Ga Nat. Zn 16 12.8 (350) 10) 67Zn(d,2n)67Ga Table 1 continues:

66Zn(d,n)67Ga Nat. Zn electroplated 50(0.075) 8 200 1.1 (30) 11) 67Zn(d,2n)67Ga on Ag(Cu)

66Zn(d,n)67Ga Nat. Zn melted onto 70(0.1) 16 200 10.2 (275) 12) 67Zn(d,2n)67Ga Cu

66Zn(d,n)67Ga Nat. Zn foil folded, 30(0.04) 8 2 1.2 (33) 13) Zn(d,2n) Ga ZnO powder

66Zn(d,n)67Ga Nat. Zn soldered onto 70-140 14 150 7.6 (205) 14) 67Zn(d,2n)6?Ga Cu (0.1-0.2)

66Zn(d,n)67Ga Nat. Zn electroplated 93(0.13)? 16 600 12.6 (340) 15) 67Zn(d,2n)67Ga onto Cu

66Zn(d,n)67Ga Enriched 66Zn, 90%, 93(0.13)? 16 600 35.0 (950) 15) electroplated onto Cu

63Cu(or,Y)67Ga Nat. Cu plate 2670(3) 30 500 5.9 (160) 16) 65Cu(a,2n)67Ga

65Cu(3He,n)67Ga Nat. Cu - 23 - 0.04 (1.1) 11) Table I continues:

64Zn(ot,p)67Ga Nat. Zn, folded foil, 30(0.04) 19 20 2.9 (77.5) 13) Zn(«,n) Ga -* ZnO powder >* «o.

Zn(ot,p)67Ga Nat. Zn, stacked foils 195(0.27) 25 0.5 6.1 (165) 17) 64Zn(a,n)67Ga -» The accelerator installed in our laboratory is a Scanditronix MC-35 cyclotron which can deliver variable particle energies with maximum values of 35 MeV protons and alpha-particles, 18 MeV deuterons and 3 47 MeV He-ions. Since the irradiations have to be performed in the extracted beam, the maximum obtainable beam currents are generally below 50 uA.

On the basis of the litterature results listed in table 1, the capabili- ties of our cyclotron and some initial irradiations of alpha-particles on natural copper, we finally decided to develop a Ga production technique based on proton irradiation of natural zinc.

OPTIMIZATION OF EXPERIMENTAL CONDITIONS

When irradiating natural zinc with protons in order to produce Ga, an unavoidable contaminant is Ga(t, = 9.5 h). It decays by p -emission followed by high-energy y-ray emmission which may expose the patient to higher radiation dose and obscure the gamma-camera pictures.

In order to optimize the experimental conditions with respect to high Ga content and low Ga contamination, the production yield for these two isotopes were determined as a function of the proton bombarding energy by irradiating a stack of 15 zinc foils. The foil thickness was o 0.1 mm (71.A mg/cm ) and the purity was 99.95%. The energy of the extracted protons was 29 MeV, the mean proton current 0.2 (jA and the irradiation time 5 min. In order to determine the absolute amount of Ga and Ga the foils were allowed to cool for about 24 hours, and then counted individually on a Y"ray spectroscopy system composed of a Ge(Li)-detector and a computerized multichannel analyzer.

The yields from the foil experiments were normalized to 1 h irradiation time and 1 uA beam current. The normalized yield for the individual foils are given in fig. 1, and the integrated yields in foil stacks composed of 1,2,3,...,15 foils are illustrated in fig. 3.

In order to reduce the contamination level of Ga below a preselected value before administration to the patients, a decay time of the sample is necessary. Based on the results from the foil experiments this decay time was calculated as a function of the applied target thickness. Irradiation times of 1 h and 3 h were supposed, and the contamination levels selected were 1% and 0.1% (defined as the ratio of the activity of Ga to Ga). The results are illustrated in fig. 2. The data from the foil experiments allow similar curves to be constructed for other bombarding energies between 15 and 29 MeV.

It may be useful to know the maximum obtainable yield when one of these contamination levels has to be kept. The curves in fig. 2 have been applied to calculate these yields as a function of the target thickness for 29 MeV bombarding energy. The results are illustrated in fig. 3. The upper curve, corresponding to the yield at the end of irradiation, levels off towards higher target thickness. The lower two reaches a maximum with a subsequent decrease in the yield and the target thickness of 1.3 mm with a bombarding energy of 29 MeV.

The average bombarding energy for each single foil in the experiment may be calculated by the procedure for stopping power calculations described in ref. 18. Then the information inherent in each of the curves of fig. 3 may be expressed in illustrative curves like those given in fig. 4. Having a particular maximum proton energy available, one may read off the optimum target thickness. For a particular target thickness the optimum proton energy may be found.

TARGETRY

The zinc targets which have been used in experiments described in the litterature, may be divided into three groups: those composed of zinc oxide powder, those of metallic zinc without a substrate and those of metallic zinc on a substrate. Columns 2 and 3 in table 1 contain some concentrated information on various target types, and column 7 gives the corresponding references. The listed target thicknesses range from 0.04 to 0.33 mm. Obviously, a target with a desired homogeneous thickness of 1.3 mm is not difficult to produce if an oxide target or a metallic target without a substrate can be applied. However, the use of zinc oxide powder was not desirable due to the contamination danger inherent in the handling process. Moreover, the planned design of the irradiation chamber supposed the target to be placed as a window to the beam tube vacuum system with direct water cooling (pressure 3 bar) applied to the window backside. Uncertainties to whether a zinc plate of thickness i.3 mm would sustain this pressure during bombardmeiit, made us decide that the zinc target should have a mechanically strong substrate with good thermal conductivity. A copper substrate was chosen because it meets these two requirements and because it, due to its J ow cost, may be dispensed after use.

In order to attain an efficient cooling, the mechanical contact between the two metals should be good, and the zinc should not flake otf or be distorted during irradiation. Various ways of producing the target have been tested including electrolysis, evaporation, gluing and melting.

The electrolysis experiments were performed both from sulphuric acid 19) solutions by the procedure described by Koehler and from hydrochloric 20) acid solutions according to Hampel . The bath temperature could be kept constant at various values, and electrolysis series with different current fluxes were accomplished. None of the procedures were fully successful. It appeared difficult to avoid grain or thread growing with increasing zinc thickness. This effect leads to metal porosity with possible liquid inclusions. A zinc thickness of 0.6 mm showed a porosity of 8%. This target was subject to a test irradiation with 10 (jA 27 MeV protons. The zinc layer was deformed by blisters. Besides, the zinc adherence to the copper substrate was relatively poor. The best results were obtained if, during the initial phase of the electrolysis, a rather weak current density was applied. The time for preparing a 0.6 mm target thickness (after machining the zinc layer to a plain surface) were several hours, and the bath had to be frequently watched in order to remove the zinc "beards" whenover necessary. Accordingly, electrolysis were judged to be impractical for our purpose.

The evaporation experiments of zinc onto a copper surface werf carried out using a commercial vacuum evaporating apparatus. The zinc was con- tained in a tantalum vessel with an evaporating orifice which could be varied in size. The vessel was heated ohmi rally and the copper plate placed above the orifice at variable distances. The desired thickness of 2 1.3 mm covering the required 4-5 cm target .i rea could easily be ob- tained. However, the zinc layer adhered poorly to the copper, and had a tendency to flake off during the following machining of the surface. Besides, it was practically impossible to avoid a zinc layer onto other surfaces in the vacuum clock, resulting in current-leading bridges on electrical insulators, break-down of vacuum gauges and a thorough and lengthy cleaning process after each target preparation.

A single test was also made to glue a zinc disc onto the copper plate with an epoxy-based glue (Araldit). After a short irradiation period the zinc layer was etched away with hydrochloric acid, and, as might be expected, a black layer of carbonized material appeared close to the center of the beam. Hence, during extended irradiations the cooling would probably be insufficient, and the method was discarded.

Melting of zinc onto a copper surface appeared to be the most successful method. In order to achieve a thin alloying layer at the contact surfaces the copper plate must be heated, and the temperature carefully control- led. The final designs of the target and the apparatus made to perform the controlled melting of the zinc, are illustrated in fig. 5. A 2 mm thick zinc disc of diameter 25 mm is placed into the 1.3 mm deep circular cavity of a 40 mm squared 2 mm thick copper plate. The zinc surface has been prepared by mechanical polishing, and the copper surface by etching in warm dilute nitric acid with subsequent rinsing in distilled water and ethanol. The copper plate is mounted into a brass chamber and the zinc covered with a thin stainless steel foil. Mechanical pressure is applied to the zinc disc by means of the screw piston. This ensures good mechanical contact between the zinc and the copper. A thermocouple mounted in mechanical contact with the copper plate indicates the tempe- rature. The chamber is then slowly heated in a small ceramic laboratory oven until 420 °C is reached (The melting point of zinc is 419.4 C). The heat is then immediately turned off. The chamber is removed from the oven and allowed to reach room temperature. During the entire heating and cooling period the chamber is flushed continuously with nitrogen. After cooling the steel foil is torn off the zinc layer, and the surface is machined to the appropriate target thickness. The zinc is thus fixed firmly to the copper through a thin alloy layer at the contact surfaces. 10

THE PRODUCTION APPARATUS

Although the maximum proton energy is about 35 MeV, a practical limit is approximately 30 MeV when high beam current is desired. The beam is extracted through a 10 m linear beam tube into the irradiation area. The end of the beam tube with the target chamber is schematically illustrated in fig. 6, and the target mechanism is shown with more details in fig. 7. The square formed target plate is loaded manually by dropping it into position onto the loading spring which ensures a proper positioning. The target is fixed firmly by turning the backplate into a semilocked posi- tion. The vacuum and water systems are thereby sealed by "o"-ring tighte- nings, and the water cooling is applied directly on the backside of the target plate.

The target mechanism is surrounded by a 20 cm thick neutron shield made of a solid suspension of boric acid in paraffin (44% by weight of boric acid, melting point of the paraffin 54-56 C). This shield reduces the neutron flux by a factor of 5-10. The shield is mounted on a movable support illustrated in fig. 8.

After irradiation the entire handling procedure is executed remotely. The cooling water is turned off and remaining water removed from the target backside by compressed air. The target chamber is ventilated. The target mechanism is opened by a pneumatic cylinder (fig. 7). The target is then automatically released and guided by the transfer channel directly into a stainless steel box situated in the middle of a cylindrical lead container (fig. 8 and 9). The transfer channel is then slid away by means of an electrical motor, and the lid is placed on the container by means of an electrically driven tackle. The lid for the inner steel box is fixed to the outer lid by a small permanent magnet, and is automati- cally snapped onto the box. An "o"-ring tightening on the lid prevents humidity from penetrating into the box and activity from leaking out. After the top lid is fixed by screws, the irradiated target is ready for the 25 km transport to The Isotope Laboratories, IFE, Kjeller, for chemical processing. 11

CHEMICAL SEPARATION OF GALLIUM AND PREPARATION OF A RADIOPHARMACEUTICAL

The final product, the radiopharmaceutical, is wanted in the form of a Ga-citrate complex in an isotonic saline solution. The solution must contain minimum amount of zinc. The iron amount must also be kept at a lowest possible value because there are recent indications that iron- citrate inhibits the uptake of Ga in tumors . Hence, we are faced with the problem of separating traces: amounts of Ga from the target material zinc and from microamounts of iron which is present as a con- taminant in the target material.

Chemical separation procedures reported in the litterature are mainly based on solvent extraction and ion exchange. The main procedures are a), solvent extraction with isopropyl ether from strong hydrochloric 22) acid solutions , b) . solvent extraction with methyl isobuthyl ketone from a 2 M hydrochloric acid solution containing 1 M sulphate 14), c). solvent extraction with 4-methyl-2-pentatone from a 5 M hydrochloric acid solution, a stripping with water and a subsequent anion exchange step , d). cation exchange from strong hydrochloric acid solutions , and e). adsorbtion on alumina from a hydrochlorice acid solution con- 23) taining nitrate at pH 2.5

All the five procedures contain steps to decontaminate the solution for possible iron content, and for all of them on acceptable removal of the target material zinc is reported. When considering their applicability two factors were of importance: the processing laboratory wanted to avoid the involvement of volatile organic materials which when mixed with air ultimately might produce explosive conditions, and the procedure should be as simple as possible and keep the involvement of man to a minimum. Procedure d. was judged to meet these requirements best, and was chosen as a starting point for further studies.

Each step in the procedure has been tested several times under realistic conditions, and parameters like acidity, the dimensions of the ion exchange column, elution rate, elution volume etc. have been varied. Detailed results from these investigations will not be given here. This discussion will be restricted to a few specific comments on selected points followed by a description of the final procedure. 12

24 25) It has been found ' by chromatographic methods that hydrolysis of gallium occurs to a significant degree when increasing the pH above 6. Although the Ga biodistribution has been reported to be largely unaffected in the pH-range 6-10, it is probably advantageous for storing purposes to reduce the pH of the final citrate solution below 6 in order to avoid a strong hydrolysis. Commercial suppliers deliver their product at a pH ~ 5.5-7.0.

The buffered isotonic saline solution may be prepared at different pH-values by varying the amount of citric acid and sodiumcitrate, keeping the total citrate concentration at 1.53 mg/ml. This is equivalent to the recommendation of the company New England Nuclear NEN, which is one of the great suppliers of Ga. The relation between the three parameters has been determined by measuring the pH-value in solutions of NaCl, Na-citrate and citric acid demanding an osmotic pressure corresponding to normal isotonic salt water (7 mg/ml of NaCl), and the constant citrate concentration of 1.53 mg/ml. The results are illustrated in fig. 10.

The main separation principle, sorbtion of radiogallium on a cation exchanger from strong hydrochloric acid solutions, constitute in fact a chemical problem. When increasing the HC1 concentration one would ulti- mately expect the formation of GaCl, which should not be fixed on a cation exchange column. The opposite is indeed the case. The overall distribution coefficient has been found to increase above 5-6 M HC1. 27) This behaviour has previously been noticed by other investigators , but no explanation of this unexpected deviation has, to the authors' knowledge, so far advanced. This problem may be the subject for a future study.

The final procedure slightly modified from the one described in ref. 10 consists of the following steps:

1. Prepare a column of the cation exchange resin DOWEX 50x8, 200-400 mesh (column length 10 cm, diameter 0.5 cm), preequilibrated in cone. HC1. Prepare at room temperature a saturated solution of Nal in cone. HC1. 13

2. Dissolve the zinc target in 25 ml of this solution and add another 5 ml after the end of reaction. The I reduces any Fe(III) present to Fe(II). The final solution has an acidity of [H+] = 7.5 M.

3. Percolate the solution through the column at a speed of 0.5 ml/min. Ga(III) is fixed to the column while Zn(II) and Fe(II) run quanti- tatively through.

4. Wash the column with five separate portions of 5 ml of the Nal-solu- tion to remove any traces of Zn and Fe, and subsequently with two portions of pure cone. HC1 to remove traces of I .

5. Elute Ga(III) with 20 ml 3MHC1 at the same speed as mentioned previously.

6. Evaporate the solution containing Ga(III) to dryness by letting a sterile air stream blow through the liquid under gentle and control- led heating (50-60 °C).

7. Cool down to room temperature and dissolve the residue in an isotonic saline solution containing citrate as a complexing agent for gallium. Use the curves of fig. 10.

8. Perform sterile filtering and autoclavation of the citrate solution and analyze the product for traces of Fe and Zn by atomic absorbtion.

A process equipment based on this procedure has been built inside a leadshielded fume hood at The Isotope Laboratories, IFE, Kjeller. The main principles are illustrated in fig. 11. The apparatus has been used in several full-scale test productions with good results. The procedure gives routinely > 80% chemical yield, and the contamination levels of Fe and Zn is below 2 ppm. It requires at present four to five hours of working time and human attention, but parts of it may easily be auto- matized.

Our product at pH 5.5 has been pyrogentested and injected on mice with no negative reactions. This part was performed by personnel at IFE. The radiopharmaceutical has subsequently been used in human investigations at the Department of Nuclear Medicine, Ullevll Hospital, in Oslo. 14

Acknowledgements

The authors would like to express their gratitude to the steering comit- tee of the project which, in addition to one of us (T.H.), has been composed of the following members: Dr. K.F. Nakken (Ullevål hospital), Dr. P.O. Bremer (IFE), Dr. F. Devik (Rikshospitalet) and Dr. A. Bull (Univ. of Oslo). Our project engineer Mrs. K. Isberg Straumsheim has skillfully participated in the investigations.

We are also indebted to the personnel at The Isotope Laboratories, The Institute for Energy Technique, Kjeller, especially to Mr. B. Andreassen and Mr. R. Beck for having directly participated in carrying out some of the chemical investigations and for having built the prototype of the process equipment. Dr. Nakken has kindly performed the final human investigations with the product.

Without the financial support from The Royal Ministry of Education and the Norwegian Research Council for Science and the Humanities, NAVF, the present work would not have been possible. 15

References

1) H.C. Dudley, G.E. Maddox and H.C. LaRue, J. Pharmacol. Exp. Ther., 96, 135 (1949).

2) D.H. Bruner, R.L. Hayes and J.D. Parkinson, Radiology, 6J, 602 (1953).

3) R.E. Hartmann and R.L. Hayes, J. Pharm. Exp. Ther., 168, 193 (1969).

4) C.L. Edward and R.L. Hayes, J. Nucl. Med., 10, 103 (1969).

5) M.L. Thakur, Int. J. Appl. Rad. Isotopes, 28, 183 (1977).

6) H.B. Hupf and J.E. Beaver, Int. J. Appl. Rad. Isotopes, 21, 75 (1970).

7) L.C. Brown, A.P. Callahan, M.R. Skidmore and T.B. Wilson, Int. J. Appl. Rad. Isotopes, 24, 651 (1973).

8) G. Herbillon, J.P. Massin, P. Collee, M. Guillaume and R. Collee, Univ. Liege, Fac. Sci. Appl., Coll, Publ., 74, 137 (1978).

9) N.N. Krasnov, P.P. Dmitriyev, 1.0. Konstantinov, N.A. Konjakin. A.A. Ponomarev, A.A. Ognev, A.E. Romanchenko and V.M. Tuyev, in proc. conf. on "The Uses of Cyclotrons in Chemistry, Metallurgy and Biology", St. Catherine's College, Oxford, 22-23 Sept. 1969, C.B. Amphlett (ed.), London Butterworths, p. 159, 1970.

10) R.D. Neirincks and M.J. Merve, J. Radiochem. Radioanal. Letters, 2» 31 (1970).

11) J.R. Dahl and R.S. Tilbury, Int. J. Appl. Rad. Isotopes, 23, 431 (1972).

12) J. Steyn and B.R. Meyer, Int. J. Appl. Rad. Isotopes, 24, 369 (1973).

13) F. Helus and W. Maier-Borst, in proc. symp. on "Radiopharmaceuticals and Labelled Compounds", Copenhagen, 26-30 March, 1973, IAEA, Vienna, Vol. 1, p. 317, 1973.

14) M. Vlatkovic, G. Paic, S. Kaucic and B. Vekic, Int. J. Appl. Rad. Isotopes, 26, 377 (1975).

15) R.D. Neirincks, Int. J. Appl. Rad. Isotopes, 27, 1 (1976.)

16) D.J. Silvester and M.L. Thakur, Int. J. Appl. Rad. Isotopes, 21., 630 (1970).

17) Y. Nagame, M. Unno, H. Nakahara and Y. Murakami, Int. J. Appl. Rad. Isotopes, 29, 615 (1978). 16

18) H.H. Andersen and J.F. Ziegler, "Hydrogen. Stopping Powers and Ranges in All Elements", Pergamon Press, 1977.

19) W.A. Koehler, "Principles and Applications of Electrochemistry", Vol. Ill, p. 151, John Wiley & Sons, Inc., London, 1944.

20) C.A. Hampel, "The Encyclopedia of Electrochemistry", Reinhold Publishing Corporation, New York, p. 466, 1964.

21) L.J. Anghileri, P. Thouvenot, F. Brunotte, C. Marchal and J. Robert, Eur. J. Nucl. Med., 7, 266 (1982).

22) L.C. Brown, Int. J. Appl. Rad. Isotopes, 22, 710 (1971).

23) P. Kopecky and B. Mudrovå, Int. J. Appl. Rad. Isotopes, 26, 323 (1975).

24) S. Kulprathipanja and D.J. Hnatowich, Int. J. Appl. Rad. Isotopes, 28, 229 (1977).

25) D.J. Hnatowich, S. Kulprathipanja and R. Beh, Int. J. Appl. Rad. Isotopes, 28, 925 (1977).

26) "Gallium Citrate Ga ", professional education service pamphlet from New England Nuclear, Radiopharmaceutical Division, Atomlight Place, North Billerica, Mass. 01862, printed in New York, 1976.

27) K.A. Kraus, D.C. Michelson and F. Nelson, J. Am. Chem. Soc, 8_1, 3204 (1959). 17

Figure Captions

Fig. 1. The figure shows the depth distribution of Ga and Ga in a 1.5 mm thick target of natural zinc irradiated with 29 MeV protons. The depth resolution is 0.1 mm (one foil thickness). The results have been normalized to an irradiation time of one hour, and are expressed in mCi/pAh.

Fig. 2. The required decay time of the irradiated sample in order to reduce the contamination of Ga below 1% (left axis) and 0.1% (right axis), respectively, as a function of the applied target thickness. The proton energy is 29 MeV. The results are illustrated for the two different and realistic irradiation times of 1 h and 3 h.

Fig. 3. The upper curve shows the integrated production yield of Ga after an irradiation time of one hour as a function of the target thickness, expressed in mCi/pAh. This curve has in the two lower curves been converted to curves expressing the "useful" production yield taking into account the decaytime required to reduce the Ga-contamination below 1% and 0.1%, respectively. The proton bombarding energy is 29 MeV.

Fig. 4. The integral "useful" production yield of Ga at 1% contami- nation level of Ga as a function of the applied target thickness. The proton bombarding energy is parameter.

Fig. 5. Cross section of the heating arrangement used for production of the zinc targets.

Fig. 6. Principle layout of the proton beamline at the irradiation position.

Fig. 7. The target-holding mechanism shown in the semilocked (a.) and the open (b.) position. 18

Fig. 8. Sketch of the apparatus for production of Ga.

Fig. 9. The container used for transport of the irradiated zinc target between the Cyclotron laboratory and the Isotope laboratory, Kjeller. The wall thickness of the lead container is at minimum 10 cm.

Fig. 10. The pH-value a.3 a function of the concentrations of citric acid and Na -citrate, keeping the total citrate concentration at 1.53 mg/ml.

Fig. 11. Principle layout of the chemical processing apparatus built for the production of the Ga-citrate radiopharmaceutical.

119A/TB/TH/HA i r i i i r i r t I I i i

Thickness of Zn foil: 0.1 mm

sz

o

__J o LL. ca UJ a. a UJ o ID a o cc 0.1 a UJ o

1 2 3 A 5 6 7 8 9 10 11 12 13 14 15 FOIL NUMBER

Fig. 1 100 o m o

Proton bombarding energy 29 MeV 135 v/ _l m LJ LJ 95 Irradiation time 1 h o Irradiation time 3 h o 130 I I 21 o O u z: r~ 90 m LJ m

125/A 5 U LJ O 0.1 0.5 1.0 1.5 TARGET THICKNESS (mm)

Fig.2 Proton bombarding o energy 29 MeV 1.0 o

CD li. O

O LU > At the end of irradiation O 0.1 o 'After the decay time corresponding o to a contamination level « 1°/o o rr o. "After the decay time corresponding o to a contamination level «s 0.170 LU cc o LU «— S 0.01 i i 0.1 0.5 1.0 1.5 TARGET THICKNESS (mm)

Fig. 3 1.0 I I I I 'I I

o o I to U- O z o

o o

Bombarding energy

a: : 29.0 MeV 1-0.1 b : 28.1 <3 c : 27.2 m E d : 26.3 o e • 25.3 f : 24.3 EJ g . 23.3 h 22.3 1 21.2 o j 20.1 o k 18.9 o er 1 17.7 o. m 16.4 n 15.1 < O UJ

0.02 i i i i ill i i 0.1 1.0 2.0 TARGET THICKNESS (mm)

Fig. A Thermocouple N2-inlet

Laboratory \\Steel foil oven \ Zn-disc Brass chamber xCu-plate

Fig. 5 Isolating Vacuum Target flanges valves mechanism \ \ CoUimator 2 \ \ CoUimator 1 Proton beam

Roughing pumps

21 a. Target with Backplate the copper backing

Pneumatic cylinder Proton beam

Graphite collimator

Fixed turning Loading spring points

b.

I

Fig. 7 Neutron shield of paraffin/boric acid

Cover for lead con- tainer \ El.-motor Lid for sample con Microswitch tainer Sample trans fer channel

Lead con- tainer

Sample container

Fig. 8 Permanent magnet

Lid for Sample sample container container with "o"- ring seal — Lead shield

Fig. 9 i r i i i i i i i 8

pH

(O I I I i i i i i i i o 0.1 0.5 1.0 1.5 2.0 CONCENTRATION (mg/ml) 1. Cone. HCl with Nal Target 2. Cone. HCl Cone. HCl 3. 3N HCl with Nal

Solution of ^-citrate, citric acid, NaCl

Vacuum

Waste ZnCI?"

Waste water (•HCl)

Fig.11 PAPER P.XVII. UNIVERSITYOF OSLO

v

THE PRODUCTION OF 81m Kr GENERATORS AT THE OSLO CYCLOTRON

T„ Bjørnstad, T. Holtebekk and A. Ruud Institute of Physics, University of Oslo , Blindern, Oslo 3, Norway

Report 84-02 Received 19.01.84

INSrrrUTEOFPHYSICS REPORTSERIES ABSTRACT

The techniques developed for the production of Kr(13s) ventilation generators at the Oslo cyclotron are described. The procedure is based on the rereactioa n Kr(p,2n) Rb using krypton gas of natural composition as target.

The target chamber is designed for a gas pressure of 10 bar. The produc- tion unit is dimensioned for a parallel production of five generators, but this number may easily be increased.

Attention is drawn to the fact that Rb(31 min) is produced in about 81 the same amount as the Rb (4.57 h) ground state. The consequences for the calibration procedure is discussed.

Routine production has been exerted without major difficulties for nearly 1^ years. The procedure has proven to be inexpensive and reliable. 1. INTRODUCTION

In the present article the techniques developed at the Oslo cyclotron for production of Kr-generators for lung ventilation studies are described. As an introduction a review of the role of radioactive noble gases in nuclear medicine is summarized.

Radioactive noble gases have been used in diagnostic nuclear medicine 133 1) since about 1955 when some investigations with Xe were reported This nuclide decays mainly (99 %) by emission of soft f}-rays accompanied by 81 keV -y-rays. The half-life is 5.29 days. The nuclide was used with limited success, and its use was temporarily renounced. It has, however, 2) later on come into extensive use, and is today the main noble gas radionuclide used in routine investigations to measure physiological functions, to vizualize organs and detect lesions and disorders.

In 1960 tests with 85mKr (t, = 4.48 h) were initiated 3"). This nuclide decays by emission of harder p-rays (77 %) or by internal transition to the ground state (23 %) with accompanying 'y-radiation of 151 keV and 305 keV. Also the long-lived ground state is, however, a 0-emitter. The general use of this nuclide in human examinations is therefore not recommended.

Application of 127Xe to ventilation studies was examined by Hoffer et al. in 1973 4). This nuclide has a half-life of 36.4 days and decays by followed by v-rays of 172, 203 and 375 keV. Apart from a few disadvantages with respect to collimation and handling and disposal, 127 5) 133 Xe has several documented advantages over Xe, such as greater photon yield, better photon chest penetration, lower radiation dose and longer shelf-life, and it is applicable together with Te perfusion imaging. The lack of access to substantial quanta has, however, limited the general use. Now the availability seems to improve.

In 1968 attention was drawn to the advantages offered by generator systems where very short-lived noble gas radioisotopes are continuously produced from longer-lived mothers . For instance, Kr (t, = 13 s) 81 ^ may be produced by decay of Rb (t, = 4.57 h). The first report on the 8lm "* use of Kr for lung ventilation and perfusion studies appeared in 1969 . The nuclide is now in common use around the world, and the literature on the medical applications is voluminous. 1.1 Applications of ""Kr

Some applications of Kr are mentioned briefly below. The corresponding references may be regarded as entrance keys to the literature.

Lung_examinations: Regional lung ventilation imaging either by single 81m 2 & 13 breath or during tidal breathing of a Kr/air(O2) mixture > ~ \ often in combination with mTc perfusion for detection of pulmonary embolism.

Heart_examinations: 1. Selective scintigraphic angiography of the right heart after infusion of a i(r-containing 5 % dextrose solution, in order to measure size, configuration and location of infarcts and other physiologic or pathologic sites of absent, or increased or decreased circulatory exchange ' . 2. Quantification of the blood flow in the 81 myocardial tissue by incorporation of Rb (potassium analogue) and measurement of the ^r clearance . 3. Determination of the right ventricular ejection fraction during intraveneous infusion of 81mKr in 5 % glucose solution 19).

Brain examinations: Evaluation of regional cerebral blood flow by continu- "••"*———------20} ous infusion of a Kr-containing 5 % glucose solution where also three-dimensional pictures may be obtained by means of single photon 21 22) emission computed tomography '

Vein examinations: Venography by constant infusion of Kr-containing 5 % glucose solution, to detect, for instance, malformations and deep 23) vein thrombosis

A non-diagnostic application is monitoring of the inhalation of, for 24) instance, cigarette smoke in man

1.2 Advantages

Some of the advantages of mKr as compared with Xe in medical exami- nations are: y- ^he emission of a single photon of 190 keV makes the nuclide well suited for use with conventional gamma-cameras without modification of the collimator. It gives less low-energy scatter than 133 the 81 keV y~ray from Xe. This implies higher spatial resolution. In addition, Kr is less soluble in fat tissue than Xe, thus causes less unsharpness of depicted boundaries.

Smaller radiation dose. The use of Kr instead of Xe implies less radiation hazard. Examples are given for lung ventilation studies and brain examinations in table 1.

TABLE 1. Comparative estimates of radiation doses from Kr and Xe examinations

Lung examinationsa>b), mGy (mrad)

1 year old 15 years old

Lungs Gonads Whole body Lungs Gonads Whole body

133Xe 7.3 0.9 - 3 0.2 (730) (90) (300) (20)

81mKr 0.75 0.04 0.0096 0.15 0.0078 0.0091 (75) (4) (0.96) (15) (0.78) (0.91)

Brain examinations , pGymin -MBq (mrad*min *mCi )

Brain Lungs Whole body 133 Xe 800 (2950) 1050 (3900) 90 (330)

0.55 (2.03) 6 .3"10~4(2.35-10~3) 1.4- 10'3 (5.15-10"2) a^From ref. 12

Estimated doses for a good-quality picture of 300 000 counts for one projection c^From ref. 25

Renders possible combined examinations. A lung ventilation study with Kr may be performed simultaneously with or immediately after a Tc (tj = 6.02 h, E = 140 keV) perfusion examination. This is feasible because of the difference in y-ray energies and the possibilities for single channel energy-window selection of the gamma-cameras: A strong activity of Tc does not hinder the recording of the ventilation scintigram.

§l!2Et_rfBfiiii23_EfEi2^§- Due t0 tne short physical half-life the admini- stration may be repeated several times during one examination period, in order to obtain pictures in multiple projections.

Studies of dynamic functions. The constant production and continuous administration cause physical equilibrium-conditions, and short-time changes may be recorded.

Minor waste disgosal_Droblems. Due to its short half-life the use of Kr causes no waste disposal problems. The exhaust gas from ventilation 133 studies may be vented out the nearest window, whereas for Xe a re- cycling system is indispensable. This also makes the Kr simpler in use, especially on children, because only minor patient cooperation is needed during the examination.

In spite of all the positive characteristics, the use of Kr is still 81 rather limited. The relatively short half-life of Rb (A.57 h) makes difficult a worldwide distribution of the generators from larger produc- tion centra. Techniques therefore have been developed at several smaller cyclotron laboratories to meet the needs locally. Informations available in 1981 showed that 113 generators from 12 cyclotrons were produced and distributed each week in Europe. Plans existed to upgrade the pro- duction to 192 generators the following year.

At the Oslo cyclotron a program for medical nuclide production develop- ment started in 1980. The present article describes the technique developed to produce mKr ventilation generators, mainly for hospitals in the Oslo area, but also for possible distribution to other Norwegian and Swedish hospitals. 2. DECAY CHARACTERISTICS OF 81m+8Rb/81raKr

For calibration purpose it is necessary to know the decay scheme for Rb. Also radioactive decay of the 31 min mRb must be taken into account.

The desintegration properties and level structure has been studied by several authors. Some of the more recent results are found in refs. 27-32. The overall results from these investigations are in general agreement. An extract of the emerging decay scheme is shown in fig. 1.

81m Rb{31.0 min) 86.2 819f?b(«.57 h) Q =2260 keV

3/2-

B1mKr(13s) 1/2- 9/2. 819Kr(2J-105y) 7/2.

3/2-

Fig. 1. Simplified decay scheme of *R

81 The metastable state of Rb decays with 97 % through the ground state. 81 If Rb partly is formed in its metastable state, there will be a radio- active build-up of the ground state until transient equilibrium is obtained (within 1.8 h). The 81gRb decays with 96 % 30'31) to the 190 keV mKr state, either directly or through higher excited states of 8lKr. The decay of 81mRb through the 190 keV state is negligible. The spin-parity change 1/2 •* 7/2 for the ground-state transition from this level ensures an almost pure E3 transition. Using the theoretical value 33) the for the total internal conversion coefficient a_ = 0.49 , 190 keV y-ray transition rate will be 64 % of the desintegration rate 81 for the Rb ground state. If the 446 keV line should be used for calibration, its absolute inten- sity must be known. This may be determined by comparison with the 190 keV line. However, recent reported values of measured ratios: WeV = 2"74 * °°8 ^ '3 "5 ± °'2 >^ ± °-007 ^ > 2.76 ± 0.07 are in disagreement. Until more reliable values are obtained, the 446 keV line is therefore unsuitable for calibration purpose.

3. PRODUCTION METHODS FOR 81Rb

81 In general Rb may be produced along several muclear reaction routes. At small cyclotrons the projectile energy limitation makes only few- nucleon transfer or exchange reactions possible. Nuclear reactions of current interest are given in table 2, together with the projectile energies required to exceed the reaction Q-value and the Coulomb barrier.

TABLE 2. Possible production methods for Rb at moderate energies.

Nuclear Q-value Minimum a E reaction (MeV) projectile 3X energy (MeV) (mb) (MeV)

79Br(0,2n) - 14.4 15.1 300 28 d) 81Br(c,4n) - 32.4 34.0 79Br(3He,n) + 6.2 12.1 c> d) 81Br(3He,3n) - 11.9 12.4 320 29 80Kr(a,3n) b) - 28.6 30.0 80Kr(o,p2n) - 23.5 24.7 80Kr(3He,2n) b) - 8.1 12.5 80Kr(3He, d) - 0.7 12.5 C) 80Kr(d,n) + 2.6 6.4 r) 82Kr(d,3n) - 16.3 16.7 82Kr(p,2n) - 14.0 14.2 83Kr(p,3n) - 21.5 21 .8 8ilKr(p,4n) - 32.0 32.4 85Rb(p,5n) b) - 44.2 44.7 29 68 <0 85Rb(p,p4n) - 39.0 39.5 200 69 e)

a) From ref. 34 b) Indirect reaction route

c) Based on a Coulomb barrier height E = ^-7-5 f7T~ 1.5(A.' + Aj ) d) From ref. 35 e) From ref. 36 The maximum energies available at the Oslo Cyclotron (Scanditronix MC 35) are for protons and a-particles 35 MeV, for deuterons 18 MeV and for He-particles 47 MeV. Thus most of the reactions listed with bromine and krypton as target materials may be iniated, but not those with rubidium target.

A survey of previously reported production methods including targetry and production yields is given in table 3. On the basis of these data the best production method for a capasity of 5 generators simultaneously has been evaluated, taking into account the energy limitation of the cyclotron, investment cost, the ease of routine operation and the radia- tion dose to the production staff.

Deuteron induced reactions are excluded on the basis of production cost, as they require highly enriched target gases of either Kr for the 80 (d,3n)-reaction or Kr for the (d,n)-reaction.

3 Reactions with He-particles on solid targets have little advantages over a-particle reactions. The production yield may be of the same order or somewhat higher at the energy available with our cyclotron, but the 3 running of a high-intensity He-beam is more expensive than an a-particle beam.

The production yield for 30 MeV a-particles on a thick target of a salt of bromine of (50.69 % 79Br) is 75 - 100 MBq/pAh . If 79 pure Br is used, the yield may be doubled. A salt target requires extensive post-irradiation manipulation: disconnection of the target or the target holder, dissolution of the target material and separation of tracer amounts of rubidium. Enriched material, being relatively expen- sive, has to be regenerated and reused. The manipulation should be partly manually controlled and would require a hot-cell for safe opera- tion because of the very strong radiation from the target unit shortly after irradiation (~150 Gy/h at about 2.5 cm distance ). With a beam intensity of 10 pA and a natural abundance target, the irradiation time for production of one generator, 1 GBq, is about one hour. TABLE 3. Reported production data for Rb.

a) Nuclear reaction Target type Target Particle Beam Prod, yield Institution Ref. thickness energy intensity (MBq/pAh) (mg/cm2) (HeV) (pA) (mCi/MAh)

79Br(u,2n)81Rb NaBr, natural thick 30 p: 74 - 92 MRC Cyclotron 8,37 abundance, melted (2.5 g total) (2 - 2.5) Unit, HaMersmilh onto Cu-backing Hospital, London, UK

Cu-Br,, natural 100 s: 74 (2.0) Argonne Cyclotron, 38 abundance, melted Argonne National onto Cu-backing Laboratory, USA and covered with anodized Al-foil

Cu-.Br, natural ~ 15 30 p: 74 (2) Rudjer Boskovic 39 abundance, melted Institute, Zagreb, onto a Cu-hacHng Yugoslavia

NaBr, natural thick 40 p: 90 (2.5) Birmingham Uni- 40 abundance, melted versity Cyclotron, onto a Mo-backing Birmingham, UK

KBr, natural thick 40 p: 74 (2.0) IPCR Cyclotron, abundance, netted Tokyo, Japan onto iron foils

NaBr, natural thick p: 52 (1.4) Washington 13 abundance, melted (1 g total) s: 40 (1.07) University Cyclo- onto a Cu-backing tron, USA and covered with Al-foil

CHBr-, natural 30 - 14 Brookhaven 41 abundance, continu- National Labora- ously flowing in a tory, New York, Ti-target holder USA

81Br(0,4n)81Rb NaBr < 200 50-31 - 15 p: 107 (2.9) Lawrence Radia- tion Laboratory 88" Cyclotron, Berkeley, USA

79Br(3He,n)8IRb NaBr, powder of -200 22 p: 1.5 (004) Sloan-Kettering 42 and natural abundance Institute Cyclo- 81Br(3He,3n)81Rb in Al-compartment tron, New York, covered with vari- USA ous foils

Conl. next page Table 3 com.

Nuclear reaction Target type Target Particle Beam Prod, yield Inst i tut ion Ref. thickness energy intensity (MBq/pAh) (mg/cm2) (MeV) (pA) (mCi/yAh)

81Br(3He,3n)81Rb KBr, natural thick 30 1 p: 74 (2.0) IPCR Cyclotron, abundance, «elted 40 p: 100 (2.7) Tokyo, Japan onto iron foils

82Kr(p,2n)81Rb Kr-gas, 70 X en- thick 22 - 14 p: 550 (15) fledi-Physirs, riched in 82Kr Emeryville, Cali- fornia, USA

Kr-gas, natural -450 21.5 - 14 p: 61.4 (1.66) German Cancer 44 abundance Research Center Cyclotron, Heidel- berg, West-Germany

Kr-gas, natural 26 15 p: 163 (4.4) Mount Sinai 45 abundance Cyclotron Facility, Miami Beach, USA

82Kr(p,2n)81Bb Kr-gas, natural - 1100 32 - 16 10 - 15 p:430 ± 30 Brookhaven Natio- 10,41 and abundance (11.6 t 0.7) nal Laboratory, 83Kr(p,3n)81Rb New York, USA

83 81 Kr(p,3n) Rb Kr-gas, natural 40-30 p: 590 (16) University of 46 abundance Milan Cyclotron, Milan, Italy

82 81 Kr(p,2n) Rb Kr-gas, 705 en- -134 26 - 10 s: 74 (2) Kernforschungs- + isotope sepa- riched in Kr zentrum Karlsruhe Q 1 ration of Rb Cyclotron, West- Geraany

85 81 Rb(p,5n) Sr RbCl, natural 1000 68 p: 37 (1.0) Harvard Cyclotron, 48 81 •* $*(2S min) Rb abundance Harvard University Cambridge, USA

8S 81 Rb(p,5n) Sr Rb2S04 powder, 720 70 1.5 p: 196 (5.3) Osaka University 49,50 • B+(25 min)81Rb natural 1000 p: 590 (16) Cyclotron, Japan and abundance 85Rb(p,p4n)8IRb

85Rb(p,p4n)81Rb 85RbCl, 100X «n- thick 70 p: 1150 (31.2) Osaka University 36 riched on a mylar Cyclotron, Japan foil

82Kr(d,3n)81Rb Kr-gas, enriched 90 22 40 p: 37 (I) Argonne Cyclotron, 51 82„ in Kr Argonne National Laboratory, USA a) p = production yield at the end of irradiation s = yield after the chemical separation 10

The thick target production yield for protons on krypton of natural abundance (11.6 % Kr, 11.5 % Kr) is of the order of 400 MBq/fJAh. If 82 enriched Kr is used the yield may be multiplied. A production process with a krypton target may readily be operated remotely, thus preventing high radiation loads to the personnel. With 10 (JA beam intensity and a thick natural abundance krypton target, the production of one generator, 1 GBq, would require about 15 min irradiation time.

After considering these facts a system for proton irradiation of natural abundance krypton gas was chosen. The construction and operation of the system is described in the following.

4. THE PRODUCTION EQUIPMENT

The production of a Rb source as generator for Kr from a target of krypton gas may, in principle, be carried out as follows: Rubidium is produced in the Kr(p,xn)-reactions in a target gas chamber filled with natural (or Kr enriched) krypton gas. The rubidium deposits on the inner surfaces of the target chamber from where it may be leached by dissolution in water and then fixed on a proper column. The production apparatus designed to take care of this procedure is shown in fig. 2. 11

COMPRESSED AIR WATER RESERVOIR TARGET

DISPENSER PROTON VOLUME BEAM

CRYOGENIC KRVPTON PUMPING RESERVOIR (liquid N2>

VACUUM — PUMP COLLECTION VOLUME

WASTE WATER LIQUID DISTRIBUTOR

GENERATORS IN LEAD SHIELD

Fig. 2. Sketch of the production apparatus. The main rig carries the target, the gas and water systems, magnetic valves, pumps, storage vessels and the waste reservoir. It is mounted on wheels for easy disconnection and removal from the beam line. The small trolley carries the lead-shielded generators and the distributor for the radioactive liquid.

4.1 The target

81 The Rb production cross-section for protons on natural krypton gas has its treshold, for the (p,2n)-reaction, at 14.2 MeV proton energy. It has a broad peak around 25 MeV and increases again when the (p,3n)- and 47) (p,4n)-reactions in the more abundant appear . In our laboratory 30 MeV is considered as maximum proton energy for stable runs. From curves for energy loss, shown in fig. 3, it is seen that degrading the proton energy from 30 MeV to 14 MeV corresponds to a path length of 300 cm of gas at atmospheric pressure. To minimize the dimen- sions of the target chamber, it was designed for a path length of 30 cm. For full production power the gas pressure must then be 10 atm at room temperature. During irradiation the pressure will increase with about 12

300 -

0.4 20 25 30 PROTON ENERGY (MeV)

Fig. 3. Differential energy loss for protons in Kr-gas at STP-condi- tions and the corresponding curve for the useful target length as a function of the proton bombarding energy.

50 % due to heating. In order to keep the target volume at a minimum and simultaniously avoid loss of beam intensity and activation of the chamber walls due to particles scattered into the wall, the chamber was made conically shaped, the diameter at the entrance being 25 mm, and at the beam stop 45 mm. Before entering the chamber the beam passes through an 3 aperture 20 mm wide. The total volume of the chamber is 300 cm .

Details of the chamber are shown in fig. 4. The target chamber material is stainless steel. The inner chamber wall has been carefully polished both mechanically and chemically in order to obtain a smooth surface which could be uniformly wetted.

The back-plate contains a beam-dump consisting of a 1.5 mm thick tantalum disc. A steel capillary entering through the back-plate is located in the bottom of the target container in its entire length. On the capillary TARGET GAS MANOMETES INIPT AND OUTLET

OUTLET FOR •'Bb-CONTAINING I WATER

Fig- 4. Cross section of the target chamber. The conical shape reduces the totsl volume while taking advantage of the diverging beam, minimizes the activation of the inner walls by scattered protons and facilitates the removal of the rinsing water from the target. 14

fifteen tiny nozzles with hole diameter of 0.2 nun are mounted with 20 mm mterdistances. Through these nozzles water is sprayed on the inner walls of the chamber. The same capillary is also used as a gate for loading and withdrawing of the target gas. The chamber wall and the tantalum beam stop are cooled by flowing water.

The high-pressure target gas is separated from the vacuum in the beam- tube by a metal foil. The foil is firmly fixed by pinching it between a steel ring and an indium tightening on the foil-flange. A number of pressure crack-tests were carried out to find the best foil material and thickness. The selected stainless steel foil, 25 pm thick (Goodfellow Metals Ltd, AISI 302, 18 % Cr, 8 % Ni), has a crack pressure of 24 - 27 bar at the selected diameter of 20 mm.

For safety precautions a second foil of the same quality is mounted upstream in the beam-tube with a large expansion volume (collimater and fluorescent screen chamber) in between.

Since the actual beam-line position is used for several other purposes, it is imperative that the production apparatus can be easily discon- nected and removed. Therefore, the target is coupled to the beam-line via a bellow by snap-on vacuum flanges. The target chamber itself is mounted on the production rig in a suspension which may be adjusted both horizontally and vertically. Target alignment is aided by a laboratory laser.

4.2 The target gas system

The target gas main reservoir consists of a high-pressure cylinder with reduction valve. Before the target chamber is loaded with krypton gas, it is evacuated to a pressure well below one mbar. The gas is then fed into the chamber until the desired pressure is obtained. After irradia- tion the target gas is transferred by cryo-pumping into a small storage pressure vessel placed in liquid nitrogen. The stored gas is reused at the next production run. In principle there is no macro-consumption of target material. In practice, however, small leaks may occationally develop in tightenings and connections. Foil failure may occur causing loss of the entire load of krypton gas. The risk of such events hampers the use of expensive enriched Kr-gas. 15

4.3 The water handling system

Demineralized and destilled water is contained in a 1000 ml glass reser- voir. After the irradiated target gas has been stored in the small pressure vessel, air is let into the container. In air the rubidium converts to rubidium-oxide which is readily dissolvable in water. Por- tions of water of preselected volume, normally 50 ml, is measured in the dispenser volume by a conductivity probe. Compressed air at 5-6 bar is applied to squeeze the water through the capillary nozzles, and the rubidium-containing water is subsequently provisionally stored in a glass storage vessel. Froiti this vessel the liquid may be directed to maximum five ion exchanger columns. The water is percolated through the columns at a speed of ~10 ml/min by means of suction supplied by a vacuum pump. It ends up in a lead-shielded waste vessel. During the production process there is no leak of krypton or rubidium activity to the air since the entire liquid system is closed and well vented.

4.4 The electronics and beam monitoring

Three video cameras watch over the production process. The various operations are executed remotely by means of an electronic operative system. This consist of a power supply and relay box for the magnetic valves and pumps fixed at the production rig, and a plug-in push-button box to activate the various relays. The entire rinsing process may also be executed automatically.

The proton beam is collimated by a variable-size aperture consisting of four isolated and watercooled graphite pieces for individual measurement of the beam current. A fluorescent screen may be applied to check the beam profile. The beam intensity onto the electrically isolated target is measured by current integration.

4.5 The generators

Various generator principles and constructions are reported in the literature (see table 4). The most common designs are based either on an organic cation exchanger of the sulfonated styrene divinylbenzene polymer 16

T&BIE&. Ventilation generator types for elution with humidified air or oxygen

Generator type Loading criteria Elution yield References f'ipnir cation exchangers, For tracer solution: 70 - 90 % 9,41,44,53 Dowex 50, » 4 or « 8, Loading speed 5 35 ml/mir., Bio Rad AG 50 « 4, > 97 %, extraction. (10 % ior Bio Rad 100-200 mesh or 200-400 mesh, For dissolved NaBr diluted AG50 « 4) contained in lucite, pyrex or to < 0.5 %: Lower speed stainless steel cylinder, and lower extraction inner dimensions 22 mm * 2.5 mn to 60 mm < 11 mm.

Inorganic ion exchanger, From directly flowing CHBr 35 - 80% 8,13,37,41 zirconium phosphate, target: Loading spee-d at grain sizes from 12-25 mesh 5-30 ml/min gives 91-94 % (low repro- to 100-200 mesh, extraction efficiency. ilucibilily container material and dimen- High extraction efficiency may occur). sions as above. for Uh from water solutions.

Water solution of the target Rapid dissolution of 70 - 80 % 8.49,50 material (NaBr or Rb SO,), target material in a glass tube.

Water solution absorbed in Rapid dissolution and ab- 70 - 80 % 40 cellulose chromatography paper, sorbtion of NaBr-target spiraled around the gas inlet material. pipe, contained in a brass housing, inner dimensions 100 mm x 500 mm.

type or on an inorganic ion exchanger of zirconium phosphate. We have chosen to design a generator of the former type. It is simple to prepare, has excellent retention of water-dissolved rubidium (> 99 % at 10 ml/min), good and reproducible elution yield (70 - 90 %) and it is safe and easy in transport and operation.

The resin DOWEX 50X8, 100 - 200 mesh, was found to be a suitable choise. The final generator design is shown in fig. 5. It consists of a perspex cylinder with outer dimensions 50 mm x 10 mm and inner dimensions 35 x 5 mm. Sintered polyethylene discs (Vyon, Porvair Limited) are used as resin support. A perspex spacer is introduced to prevent function damage if the gas elution by accident is reversed. It consists further- more of steel supporting rings and teflon screw-on fittings with o-rings in order to fix the teflon capillary tubes (1.6 mm o.d., 0.8 mm i.d.) to the cylinder. The capillary ends are supplied with connectors (Omnifit 1001) for small tubings (0.5 - 4 mm o.d.). The generator is shielded with a 5 mm thick lead hose and a lead tower of 40 mm thickness and placed in a tin packing with a handle (fig. 6) for easy transport. This 17

CAPILLARY FITTING

SEALED POLYETHYLENE TUBING

0-RINGS

STEEL DISC

SPACER

PERSPEX CYLINDER

ION EXCHANGE RESIN

SINTERED POLYETHYLENE FILTER; RESIN SUPPORT

TEFLON CAPILLARY

TEFLON SCREW CAP

W 81m,, Fig. 5. Design of the generator with teflon tube fitting on the """Kr- carrying capillary. An identical fitting is also supplied on the capillary for the incoming gas.

LEAD SHIELD

GENERATOR

i. o SCALE (cm)

Fig. 6. The generator in the lead-shield surrounded by the tin can. The lid is machine-sealed onto the can. 18

unit is during production situated on a small trolley directly beneeth the target (fig. 2), and coupled by quick-connectors (Swagelock) to the radioactive liquid storage vessel and to the waste vessel.

5. OPERATION AND PERFORMANCE 5.1 Preparation and production

The ion exchange resin is swelled for several hours before use. Each 3 generator is packed with approximately 0.3 cm resin, and the liquid resistance is adjusted with microvalves for each generator for a flow rate of ~ 10 ml/s. Thus, generators produced simultaneously will have approximately the same strength.

The loading of the target gas is performed with the plug-in push-button box at the production rig, while the rest of the production process is executed remotely from the cyclotron control desk.

The irradiations are performed with a beam intensity of 10 - 15 |JA.

In order to design a reasonable rinsing sequence, the effect of surface treatment on the dissolution properties of rubidium has been examined. Stainless steel foil strips were treated in various liquid media before inserted into the target chamber where rubidium was deposited during a test irradiation. The activities on the foils were measured before and after ten consecutive single dips in demineralized water. The highest dissolution, > 99 %, was found for a foil which had been treated with a mixture of 64 vol% HF(39 %), 35 vol% HNO3(65 %) and 1 vol% H^ (30 %) for 3 s. The lowest dissolution yield, 96 %, was found for an ethanol treated foil. In the target chamber itself, which was treated with the HF - HNO„ mixture, the total dissolution yield after six consecutive rinsings was larger than 90 %. The relative yield in each individual rinsing sequence is illustrated in fig. 7. Consequentially to these results, the rinsing sequence is repeated four times with 50 ml portions in routine production. More than 98.5 % of the practically possible yield is obtained.

After the suction of water through the ion exchangers, only traces of rubidium activity are found in the waste vessel. Hence, the absorbtion yield of rubidium is nearly 100 %. 19

TOO 90

C? 80

3 70 Ui >60 o I 50 S 40

20 LLI 10 i 12 3 4 5 6 RINSING BATCH NO.

Fig. 7. The effect of rinsing the target container by sequential portions of 50 ml demineralized and destilled water, measured radiochemically.

->.2 Calibration, control and packing

After completion of the rubidium separation, the trolley with the gene- rators is quickly disconnected from the rest of the production apparatus and transported to the chemical laboratory for calibration, control and packing of the generators.

The calibration is routinely performed 30 min after end of irradiation by measuring the dose rate from the lead-shielded generator at a fixed position. The dose rate has previously been calibrated against the saturation activity of Kr on the ion exhanger.

As is seen from the decay scheme (fig. 1) any produced amount of mRb may influence the calibration. This effect has not been mentioned in the 54) topical literature until recently . It is unclear whether it has been taken into account or has any influence on values previously given for production yield (table 2).

In order to quantify the effect, the saturation activity of Kr on a weak generator has been followed as a function of time. The rubidium activity was produced with the following irradiation parameters: Irradia- 20

tion time t. = 30 s, beam intensity = 500 nA, energy interval 26 - 17 MeV (which is normally used in the routine production of one single genera- tor). The resulting growth and decay curve (fig. 8) is resolved into two components corresponding to the half-lives of Rb(31 min) and Rb(4.57 h).

103

T,,2 = 4.57 h.

T,,2 r 31 min

102

z

L__L i i i i 01 23456789 DECAY TIME(h)

Fig. 8. Growth and decay curve of mKr saturation activity on the generator.

The growth and decay curve may be expressed by the formula

-AtAt. --AA t 81)>e 8 = R (i e g • -k t. -At. -At. "At , m 2 . „ 8 1- 8 d m

R -f-A -At. -At, -At, JD i m i s / m d . eds 8 " « 21

where R and R are the production rates and A and A the des- 8 ro g m integration constants of the ground state and the metastable state respectively, f is the branching ratio for the isometric transition (97.6 %), t^ the irradiation time and t, the decay time measured from end of bombardment.

The three terms express the activity of the ground-state nuclide formed directly, formed by decay of the metastable state during bombardment and formed by decay of the metastable state after end of bombardment, respec- tively. For short irradiation times decay during irradiation is negli- gible and the formula for the decay curve simiplifies to

-A td -Amtd where A.. + A„ and A„ are the values at t, = 0 on the resolved decay curves.

The production ratio for the two states is then

R. Am Aj

m " g 2

From the decay curve we find k^^ = 0,84 ± 0,10 or R /Rm = 0,94 + 0,10. Thus the production rates of the two isomers are about equal for the energy interval 26-17 MeV.

If the radioactiv build-up during and after end of irradiation is dis- regarded, the calibration may be erroneous. For instance, if the irradia- tion time for production of one generator is 15 min and the calibration is performed within the first hour after end of irradiation, the genera- tor strength at the user time (3 h after end of irradiation) will be underestimated by 17 % to more than 90 %. Calibration of the generator after 30 min implies an underestimate of 35 %. The correct generator strength at any time may be derived from the formula above. 22

For control each generator is, in sequence, connected to a gas elution system where humified high-purity nitrogen is flushed at controlled rate through the generator. The eluted gas is led through a capillary which passes a Ge(Li)-detector. The recorded y-ray spectrum is used to check the radiochemical purity of the eluted gas. A typical y-ray spectrum is shown in fig. 9. Only Kr activity can be seen.

,13.

81m Kr

£ 212 z

10 o

100 200 300 400 500 ENERGV(keV)

Fig. 9. Gamma-ray spectrum of Kr in the gas eluate demonstrating the radiochemical purity.

For elution control the counting rate from the eluted gas in the capilla- ry section seen by the detector, is recorded. For a selected gas flow rate, v, the counting rate, R , is a function of the generator strength, D, the elution yield £(v), and a transport factor k(v):

R = k(v)-e(v)-D ,

The elution yield has been determined directly on a weak generator by measurement of the mKr activity as a function of the gas flow. The elution yield, given as the percentage of the activity carried away to 23

the saturation activity, is shown in fig. 10. It is seen that the elution is nearly constant, 80 - 85 %, for a gas flow between 2 and 20 ml/s.

1 ' '

0.90 -

§-- —§ 9 V o • / / UJ ; / / Activity at equilibrium - Activity during flow y o.7o e(v) = Activity at equilibrium u. / u. 1 UJ §0.60 H

_j 1 UJ 0.5a * 1 1 • i 1 t 1 ' 1 lii • I 1 I i ' " 10 15 20 GAS FLOW RATE (ml/s)

Fig. 10. Elution efficiency of mKr from the generator as a function of the gas flow rate.

By elution tests with a gas flow of 10 ml/s on several generators it has been confirmed that the ratio R /D normally takes the same value. This ratio is used as a check for complete elution. If the counting rate is lower than expected as compared with the measured generator strength, the gas is allowed to flow for some minutes. The elution efficiency then normally reaches the expected value. The explaination is probably that residual water in the column, which caused the gas to channel through the resin, has been blown away.

After calibration and control the generator tin can is sealed and packed according to Norwegian regulations. The control process is overseen by a radiopharmacist from the Isotope Laboratories, Institute for Energy Technique, Kjeller, who also is responsible for the marketing and tran- sport of the product. 24

6. PERFORMANCE AND OUTLOOK

The apparatus and the techniques described here has by now been reliably functioning in commercial production of mKr-generators for nearly \\ years. The only severe mis functions which have occurred, are a few foil failures. We have experienced that pin holes may develop in the foil or it may crack when a combination of high pressure, heavy radiation load and heat deposited by the proton beam is imposed on the foil for a long time. Also distinct signs of corrosion may appear, probably due to reminicent humidity in the target chamber. Therefore, the foil is renewed for every third production run.

In the continuous use of the generators one has from time to time noticed that the Kr-activity in the eluate decreases faster than expected from the physical half-life of 8Rb. This is due to dry-out of the resin, resulting in longer diffusion times of Kr atoms out of the grains. The generator may be reconditioned by injection of demineralized water into the generator to allow the resin to swell. If a well humified elution gas is applied, the dry out does not appear.

The expected production quantity of 4 - 5 generators per week has not yet been realized, mainly due to lack of dedicated interest in the Norwegian hospitals. Only the Ullevål hospital in Oslo is regularly supplied with one generator per week. Occasionally generators are produced for some of the other neighbouring hospitals. A total of approximately 70 generators have been sold to a prize of Nkr. 1 600.- per generator.

From the applications referred in chapter 1.2 it is clear that the evaluation of an infusion generator would be of value. From local nuclear medicine departments interest in combined generators has been expressed. This may be a subject for future development, and preliminary prepara- tions have been started. 25

ACKNOWLEDGEMENT

The authors would like to express their gratitude to the steering board of the project which, in addition to one of us (T.H.), has been composed of the following members; Dr. K.F. Nakken (Ullevål Hospital), Dr. P.O. Bremer (IFE), Dr. F. Devik (Rikshospitalet) and Dr. A. Bull (Univ. i Oslo).

We are indebted to the personnel at the Oslo Cyclotron, especially to Mr. E.A. Olsen, who has the responsibility for keeping the cyclotron in a running condition at any timt.'.

Financial support has been granted from The Royal Ministry of Education and The Norwegian Research Council for Science and the Humanities (NAVF).

157I/TH/HA/8.12.83 26

REFERENCES

1) H.W. Knipping, W. Bott, H. Venrath, H. Valentin, H. Ludes and P. Endler, "Eine neue Methode zur Priifung der Herz und Lungenfunktion", Dtsch. Med. Wochenschr., 80, 1146 (1955)

2) H.P. Gurtner, W.A. Briscoe and A. Cournand, "Studies of the Ventila- tion-Perfusion Relationship in the Lungs of Subjects with Chronic Pulmonary Emphysema Following a Single Intravenous Injection of Radioactive Krypton", J. Clin. Invest, 39, 1080 (1960)

3) R. Pommet and F. Therain, "The Use of Kr-81m in Ventilation Imaging", Clin.Nucl. Med., 7, 122 (1982)

4) P.B. Hoffer, P.V. Harper, R.N. Beck, V. Stark, H. Krizek, L. Heck 127 and N. Lembares, "Improved Xenon Images with Xe", J. Nucl. Med., 14, 172 (1973)

5) H. Susskind, H.L. Atkins, J.F. Klopper, A.N. Ansari, P. Richards and R.G. Fairchild, "Xenon-127 Ventilation Studies", in "Progress in Nuclear Medicine" by A. Donath, H. Rosier and A. N. Serafini (eds.), Vol. 5, p. 144, (Karger, Basel 1978)

6. Y. Yano and H.O. Anger, "Ultrashort-lived Radioisotopes for Visuali- zing Blood Vessels and Organs", J. Nucl. Med., 9, 2 (1968)

7. T. Jones and J.C. Clark, "A Cyclotron Produced 81Rb-81mKr Generator and its Uses in Gamma-Camera Studies", Br. J. Rapid, 42, 237 (1969)

8) T. Jones, J.C. Clark, J.M. Hughes and D.Y. Rosenzweig, "81raKr Generator and its Uses in Cardiopulmonary Studies with the Scintil- lation Camera", J. Nucl. Med., 11., 118 (1970)

9) Y. Yano, J. McRae and H.O. Anger, "Lung Function Studies Using Short-lived Kr and the Scintillation Camera", J. Nucl. Med., 1J, 674 (1970) 27

10) E. Kaplan, L.W. Mayron, W.E. Barnes, L.G. Colombetti, A.M. Friedman and J.E. Gindler, "Continuous Radionuclide Generation. II. Scinti- graphic Definition of Capillary Exchange by Rapid Decay of mKr and its Applications", J. Nucl. Med., 15, 874 (1974)

11) F. Fazio and T. Jones, "Assessment of Regional Ventilation by Continuous Inhalation of Radioactive Krypton-81m", Br. Med. J., 265, 673 (1975)

12) D.K. Li, S. Trevers, S. Heyman, J.A. Kirkpatrick, Jr, R.M. Lambrecht, T.J. Ruth and A.P. Wolf, "Krypton-81m: A Better Radiopharmaceutical for Assessment of Regional Lung Function in Children", Radiology, 13o, 741 (1979)

13) T.R. Miller, D.R. Biello, J.I. Lee, H.H. Davis, A.G. Mattar, G.J. Ehrhardt and B..A. Siegel, "Ventilation Imaging with Kr-81m. A comparison with Xe-133". Eur. J. Nucl. Med. 6, 11 - 16 (1981)

14) E. Kaplan, L.W. Mayron, A.M. Friedman, J.E. Gindler, L. Frazin, J.M. Moran, H. Loeb and R.M. Genmar, "Definition of Myocardial Perfusion by Continuous Infusion of Kr-81m", Am. J. Cardiol., 37, 878 (1976)

15) P.H.T. Kleynhans, M.G. Lotter, A. van Aswegen, C.P. Herbst, J.D. Marx and P.C. Minnaar, "The Imaging of Myocardial Perfusion with mKr During Coronary Arteriography", Eur. J. Nucl. Med., 7, 405 (1982)

16) T. Jones and C.M.E. Matthews, "Tissue Perfusion Measured Using the j5 1 ft 1 m Ratio of 81RRbb tto Kr Incorporated in the Tissue", Nature, Lond., 230, 119 (1971)

17) R.P. Spencer and F. Hosain, "Theoretical Basis for Regional Blood Flow Measurement by Use of the Ratio of Rb: Kr", Int. J. Appl. Radiot. Isotopes, 3, 55 (1976)

18) J.P. Idoine, B.L. Holman, A.G. Jones, R.J. Schneider, K.L. Schroeder and R.E. Zimmerman, "Quantification of Flow in a Dynamic Phantom Using 81Rb-8lmKr, and Nal Detector", J. Nucl. Med., 18, 570 (1977) 28

19) W.H. Knapp, F. Helus, R.M. Lambrecht, R. Elfner, H. Gasper and H.H. Vollhaber, "Kr-81m for Determination of Right Ventricular Ejection Fraction (RVEF)", Eur. J. Nucl. Med., 5, 487 (1980)

20) F. Fazio, M. Nardini, F. Fieschi and C. Forli, "Assessment of Regional Blood Flow by Continuous Carotid Indusion of Krypton-8lm", J. Nucl. Med., 18, 962 (1977)

21) F. Fazio, C. Fieschi, M. Nardini .... "Assessment of Regional Cerebral Blood Flow by Continuous Carotid Infusion of Krypton-81m and Emission Computerized Tomography", Acta Neurol. Scand. Scand., 60: Suppl. 72, 192 (1979)

22) T. Maeda, H. Matsuda, K. Hisada, N. Tonami, H. Mori, H. Fujii, M. Hayashi and S. Yamamoto, "Three-Dimensional Regional Cerebral Blood Perfusion Images with Single-Photon Emission Computed Tomography", Radiology, 140, 817 (1981)

23) H.R. Ham, J. Vandeviven, M. Guillaume, T. Niethammer and R. Sergey- sels, "Radionuclide Venography using Continuous mKr Infusion: Preliminary Note", Clin. Nucl. Med., 6, 461 (1981)

24) N.F. Sheahan, D. Pavia, J.R.M. Bateman, J.E. Agnew and S.W. Clarke, "A Technique for Monitoring the Inhalation of Cigarette Smoke in Man, using Krypton-81m", Int. J. Appl. Radiat. Isotopes, 2!> ^38 (1980)

25) E. Kaplan and L.W. Mayron, "Internal Conversion Electrons and Krypton-81m Dosimetry", Int. J. Nucl. Med. Biol., 5, 199 (1978)

26) J. Clark and I. Watson, "Krypton-81m, the Radionuclide for Pulmo- nary Ventilation Studies. Availability in Europe", leaflet, produced at the "Medical Research Council Cyclotron Unit, Hammersmith Hospi- tal, London, UK, 1981

27) A. Li-Scholz and H. Bakhru, "Studies in the Decay of 4.7 h Rb ", Phys. Rev., HJ8, 1244 (1968) 29

28) S.L. Waters, D.J. Silvester and I.W. Goodier, "Decay of 81Rb", Phys. Rev., C2, 2441 (1970)

29) R. Broda, A.Z. Hrynkiewicz, J. Styczen and W. Walus, "The Excited States of Kr and Rb Isotopes in the 78å A å 85 mass Region", Nucl. Phys., A216, 493 (1973)

30) S. Vaisala, T. Raunemaa, A. Fontell, G. Graeffe and A. Siivola, "The decay of Rb 3/2 Ground State", Phys. Fenn., K), 133 (1975)

31) J. Liptak, K. Kristiakova and J. Kristiak, "Properties of 81Kr 81 Levels Populated in the Decay of Rb Isomers", Nucl. Phys., A286, 263 (1977)

81 82m 83 84m+8 83 85m 87m 32) A. Grutter, "Decay Data of ' ' ' Rb and > > Sr"; Int. J. Appl. Radiat. Isotopes, 33, 456 (1982)

33) R.S. Hager and E.C. Seltzer, "Internal Conversion Tables. Part I: K-, L-, H-Shell Conversion Coefficients for Z = 30 to Z = 103", Nuclear Data, A4, 1 (1968)

34) K.A. Keller, J. Lange and H. Miinzel, "Q-Werte und Anregungsfunk- tionen von Xernreaktionen", in Landolt-Bornstein: "Zahlenwerte und Funktionen aus Naturwissenschaften und Technik, vol. 5, Springer- Verlag Berlin, 1973

35) Y. Homma and K. Kurata, "Excitation Functions for the Production of 8lRb-81mKr via the 79Br(a,2n)81Rb and the 8lBr(3He,3n)81Rb Reactions" Int. J. Appl. Radiat. Isotopes, 30, 345 (1979)

36) T. Horiguchi, H. Noma, Y. Yoshizawa, H. Takemi, H. Hasai and Y. Kiso, 85 "Excitation Functions of Proton Induced Nuclear Reactions on Rb", Int. J. Appl. Radiat. Isotopes, 31, 141 (1980)

37) J.C. Clark, P.L. Horlock and I.A. Watson, "Krypton-81 m Generators", Radiochem. Radioanal. Letters, 25, 245 (1976)

38) L.W. Mayon, E. Kaplan, A.M. Friedman and J.E. Gindler, "The Prepa- 81 ration of Rb in High Specific Activity Quantities and its Use in a 81Rb-81mKr Generator", Int. J. Appl. Radiat. Isotopes, 25, 237 (1974) 30

39) J. Nosil, S. Spaventi and I. Slaus, "82mKr: Production, Application, and Use of Computer for Ventilation Studies", Eur. J. Nucl. Med., 2, 1 (1977)

40) J.H. Fremlin, K. Stammer and F.R. Stewart, "A new Generator for Krypton-81m", Nucl. Instr. Meth., 156, 369 (1978)

41) T.J. Ruth, R.M. Lambrecht, A.P. Wolf and M.L. Thakur, "Cyclotron Isotopes and Radiopharmaceuticals XXX. Aspects of Production, Al Q -I Elution and Automation of Rb- Kr Generators", Int. J. Appl. Radiat. Isotopes, 31., 51 (1980)

42) I.A. Watson, "The Production of 81Rb and 82mRb for Medical Use", Radiochem. Radioanal. Letters, 4, 7 (1970)

43) J.F. Lamb, G.A. Baker, A. Khentigan, H.A. Moore, W.C. Neesan and H.S. Winchell, "Manufacture of the Krypton-8lm Gas Generator for Lung Ventilation Studies", J. Nucl. Med., 28, 609 (1977)

44) F. Helus, H. Gasper, U. Sahm and W. Maier-Borst, "Remotely Control- 81 led Target System for the Routine Production of Rb", Radiochem. Radioanal. Letters, 44, 187 (1980)

45) R. Finn, M. Vora, T. Booth, J. Campbell, S. Carroll and J. Clark, "Radiation-Induced Defects as illustrated by the 81Rb-81mKr Target System", Int. J. Appl. Radiat. Isotopes, 33, 349 (1982)

46) E. Acerbi, C. Birattari, M. Bonardi, C. de Martinis and A. Salomone, "Kr(p,xn) Excitation Functions and 81Rb-81mKr Generator Studies", Int. J. Appl. Radiat. Isotopes, 32, 465 (1981)

81 47) A. Hauser and B. Feurer, "Pure Rb for Medical Use Obtained by Electromagnetic Isotope Separation", Int. J. Appl. Radiat. Isotopes, 32, 775 (1981)

48) R.J. Schneider and C.J. Goldberg, "Production of Rubidium-81 by the Reaction 8585Rb(p,5n)8181Sr and Decay of 81Sr", Int. J. Appl. Radiat. Isotopes, 2_7, 189 (1976) 31

49) T. Koyama, T. Koyama, Y. Hirokawa, Y. Yoshizawa, H. Noma, T. 81m Horiguchi, Y. Kiso, H. Hasai and H. Takemi, " Kr Gas Generator for Lung Ventilation Study", Eur. J. Nucl. Med., 5, 481 (1980)

50) Y. Yoshizawa, T. Horiguchi, H. Hasai and T. Koyama, "Production of Q 1 Rb for the Nuclear Medicine", Annual report 1980, Research Center for Nuclear Physics, Osaka Univ., Japan, p. 129

51) J.E. Gindler, M.C. Oselka, A.M. Friedman, L.W. Mayron and E. Kaplan, "A Gas Target Assembly for the Production of High Purity, High SpecifSpecifijc Activity 81Rb", Int. J. Appl. Radiat. Isotopes, 27, 330 (1976)

52) H.H. Andersen and J.F. Ziegler, "Hydrogen. Stopping Powers and Ranges in All Elements", Vol. 3 of "The Stopping and Ranges of Ions in Matter" by J.F. Ziegler, Pergamon Press 1977.

53) L.G. Colombetti, L.W. Mayron, E. Kaplan, W.E. Barnes, A.M. Friedman and J.E. Gindler, "Continuous Radionuclide Generation. I. Production and EEvaluatiov n of a Kr Minigenerator", J. Nucl. Med., lj>, 868 (1975)

54) J.J.L. Mulders, A.G.M. Janssen, A.J. Witsenboer and J.J.M. De Goeij, "Possible Calibration Errors for Rb- Kr Generators due to a 81 Dual Production Route of Rb", Int. J. Appl. Radiat. Isotopes, 34, 717 (1983)