Preparation of Targets by Ion Implantation D.C

AECL-5503

ATOMIC ENERGY WS& L'ENERGIE ATGMIQUE OF CANADA LIMITED •Ajf DU CANADA LIM1TEE

PROCEEDINGS OF THE 1S74 ANNUAL CONFERENCE OF THE NUCLEAR TARGET DEVELOPMENT SOCIETY held at CHALK RIVER NUCLEAR LABORATORIES CHALK RIVER, ONTARIO 1-3 OCTOBER, 1974

Printed by Atomic Energy of Canada Limited

Chalk River, Ontario January 1975

Reprinted April 1976 NUCLEAR TARGET DEVELOPMENT SOCIETY

PROCEEDINGS OF THE 1974 ANNUAL CONFERENCE

K.L. Perry - Conference Chairman J.L. Gallant - Co-Chairinan

Texts of the 2; papers that were presented at the 1974 N'TDS Annual Conference

held at

CHALK RIVER NUCLEAR LABORATORIES CHALK RIVER, ONTARIO, CANADA

1-3 OCTOBER,

Printed by Atomic Energy of Canada Limited Chalk River, Ontario January, 1975

Reprinted April 1976 AECL-5503 NUCLEAR TARGET DEVELOPMENT SOCIETY

Compte rendu des travaux du Congres annuel de 1974

Resume

Cette publication comprend le compte rendu d' UP. Congres international dont les travaux se sont deroules dans le doir.aine general de la preparation des cibles pour les accelerateurs de particules. Le compte rendu ccncerne les 21 communications presentees au Congres. Les questions a 1'ordre du jour etaient les suivantes: 1. preparation de films minces (solides) par diverses methodes: evaporation sous vide; roulage, precipitation chimique, implantation d'ions, electrodeposition, electropulverisation; 2. cibles liquides; 3. cibles gazeuses; 4. substrats et agents de separation; 5. preparation de matieres destinees a former des cibles; 6. mesure de l'epaisseur, de 1'uniformite et de la purete. Tenu du ler au 3 octobre 1974 dans les Laboratoires Nucleaires de Chalk River en Ontario, ce Congres etait parraine par Nuclear Target Development Society en cooperation avec l'Energie Atomique du Canada, Limitee.

imprime par L'Energie Atomique du Canada, Limitee Laboratoires Nucleaires de Chalk River Chalk River, Ontario Janvier 1975

Reimprime en avril 197b

AECL-55 03 NUCLEAR TARGET DEVELOPMENT SOCIETY

PROCEECINGS OF THE 1974 ANNUAL CONFERENCE

ABSTRACT

This publication Incorporates the proceedings of an international conference concentrating on the general field of target preparation for use with particle accelerators. The proceedings include the 21 papers that were presented at the meeting dealing with the following topics:

1. preparation of thin (solid) films by vacuum evaporation, rolling, chemical deposition, ion implantation, electrodeposition and electrospraying,

2. liquid targets,

3. gas targets,

4. substrates and parting agents,

5. preparation of target materials,

6. measurement of thickness, uniformity and purity.

The conference was held at the Chalk River Nuclear Laboratories, Chalk River, Ontario, from 1-3 October, 1974. It was sponsored by the Nuclear Target Development Society in co-operation with Atomic Energy of Canada Limited.

Printed by Atomic Energy of Canada Limited Chalk River, Ontario January 1975

Reprinted April 1976

AECL-55C3 (i)

CONTENTS pAGE TUESDAY, 1 OCTOBER SESSION A

D.C. SANTRY

H.I.. ADAIR à E.H. KOBTSK

L.0. LOVE 29

A. LOUGHEiiD & E.K. HLH.FT 39

SESSION B

G. SLETTEN

W.V. CONNER 59 VuLti'.-lc :.'ir,k

ll..n\ ;;• T-iP.jci Material P. miER-KOMOE 70 '/upep i'l'esciu'a on VI'Ï'J .:>•. B'takiu ;a by

':i>:-j 1 F.o 1.1 ;;•: Cia\ C-A. BOUCHARD 74

K.W. ALLEN et al. 76 i:.ïdi-?.ti>Ju Caviure

WEDNESDAY, 2 OCTOBER SESSION C I.V. MITCHELL

.--. L'<:Ï'.V'ITi.O:: !-!ethod for DeteP'r.ininc A.H. CHUNG, W.T. DIAMOND 84 & A.E. LITHERLAND :.•: ïualc'M' jar jets

i.':;";t .'A;.,v Measurements Problems J.S. MERRITT 86 (ii)

J. KW1NTA & F. AMOL'URY 95

J.D. S1INSCN 100

N. BUR.N & L.B. BENDER 105

SESSION D

i're: 'rali :••' ait.i ,'.. J. VAN A'JDCNHOVE, 119 V. VERDINGH, H. ESCHBACH & P. DE BIEVRE

:'.-. of Substrates I, D. RAMSAY 151

E.W. McDANIEL, L.O. LOVE 159 W.K. PRATER & R.L. BAILEY Palladium, and Platinw ]'.'!'• iiescarak Use

W.D. RIEL 167 Enriched Iron from Ike I>i-Situ Ru da avion of Fe .•• 0.,

Sy.Z'-ti.ai Tocjet Preparation 7cciv:~'.. J.L. GA!.LANT 169 for Chalk River Unclear Pays'C-L

SESSION E

: 'ike lsc of Lasers for the o<:;;ar>al-: R.D. McALPINE 179 PREPARATION OF TARGETS BY ION IMPLANTATION D.C. Santry Chalk River Nuclear Laboratories Atomic Energy of Canada Limited Chalk River, Ontario

[INTRODUCTION The technique of using an energetic irir. beam to introduce ions into a substance is called ion iniplantatxr.n. .Although the method is usod extensively for the doping of cemiroiic.ao-or materials for the preparation of commercial semiconductor devices, ion implantation has not gained wide acceptance in target preparation for nuclear measurements.

In this paper I propose to describe various factors which are involved in target preparation by direct ion implantation and will emphasize the limitations and pitfalls of the method. Examples will be given of experiments for which ion implanted targets are well suited.

Since the central theme of the conference is target preparation for use with accelerators, I intend to say very little about implan- ter machines other than to mention that they too are accelerators, usually laboratory size electromagnetic isotope separators. The topic of target preparation therefore becomes one of the chemistry and physics of accelerated beams stopping in solids.

ION RANGES One of the main differences between targets prepared by ion implantation and the various deposition methods is that the former introduces a specific material into a substrate rather than on top of it. When an ion beam of low energy hits the surface of a target material, most of the ions are stopped and trapped in a thin layer just below the surface. The thickness of the layer depends on the ion. energy, mass of the ion, and the nature of the stopping material. The distribution of the implanted ions is expected to be approximately Gaussian in shape and the thickness penetrated by 50% of the ions is referred to as the median rango.

For many experiments ion implantation can be regarded as producing very thin targets. However, for high precision work such as particle spectrometry, the energy spreac introduced by an effective thickness of any .source can be important. It therefore becomes desirable to know the range of implanted ions in solids. Unfortuna- tely, this information is still sadly lacking. There is available the semi-empirical tabulation of heavy ion ranges by Northcliffe and Schilling (1) which covers the energy range 10 keV/AMU to 12 Mev/AMU, energies too high for most ion implantations, Johnson and Gibbons (2) have provided calculated range data for various projectiles into selected targets, covering a more useful energy range of 10 to 1000 keV. The calculations are based on the L.S.S. (3) theoretical treatment of ion ranges.

I would now like to show that the only certain way of knowing the range of a specific ion is to actually measure it for the material to be used. Figure 1A shows an integral range distribution

l0 °\ l34 40 keV Cs i\ 40 keV IONS IN TUNGSTEN

",'0r

Q ; \ \- 'N TUNGSTEN \

IN GOLC;

50 100 150 200 250 300 25 50 75 100 125 150 175 DEPTH (ug/cmZ) DEPTH (ug/cm2)

Figure 1 measured by ion implanting at 40 keV a radioactive tracer into polycrystalline gold and tungsten. The residual activity was measured as thin layers were removed from ti.e surface (4). As was expected from theory, the median range was identical in Au and W. Figure IB confirms that different elements with similar masses when implanted at a given energy will have the same median range. Range measurements were made using many radioactive nuclides and the median range values obtEiined are plotted in Figure 2. The relationship

MEDIAN RANGE OF 40 keV ions O IN WCJ

D fiO 80 I0D !?D 140 1 ED !8D ilJD ;20

Figure 2 between ion range and ion energy is shown in Figure 3. Note that for high Z stopping media, measured ranges are greater than that calculated by the L.S.S. theory. In Figure 4 are shown a series of

50-

10 60 K e V Figure 3 Figure 4 range measurements made with a W single crystal. The median range for implanting along the 111 direction was 247 ...g/cm , the enhanced penetration being due to the ion channe1ing effect. Implant- ing at an angle of 8° from the 111 direction decreased the range distribution only slightly. An analysis of the crystal structure of W indicated that a tilt of 24° would produce the most random arrangement of atoms for this crystal. Implants made under these conditions gave a median range of 27 -cj/cm which is close to the value of 24 ..g/cm obtained with polycrystalline material. However, the range distribution for the random direction indicated chat many ions are still able to travel great distances into the crystal.

The polycrystalline W metal used for range measurements shown in Figure 2 was prepared by pressing and sintering very fine grained W powder. Range measurements made using rolled W foil are shown in 2 Figure 5. The median range value was 40 u.g/cm and the range

t'\ H

- - \ \ -

1 1 1

Figure 5 distribution showed large penetration depths for a small fraction of the implanted ions. The sintered W when annealed at 2250°C gave a 2 median range of 33 ug/cm and also exhibited the enhanced penetration of ions.

Thus it has been observed that the depth of an implant can also be influenced by crystal structure and the metallurgical process used to prepare the stopping media. Perhaps it should be pointed out that such effects may also be important for other measurements involving ions or particles moving through matter, i.e. positron lifetimes in solids and nuclear lifetimes determined by Doppler-shift attenuation measurements.

ION RETENTION Another difference between targets prepared by direct ion implantation and other techniques is in the amount of material which can be deposited. Studd.es by Qmen and Bruce (5) have shown that at keV energies, targets reach implanted saturation amounts of only a 2 few ~g/cm . For most accelerator experiments this constitutes too few atoms for meaningful measurements. The cause of this saturation effect can be shown to be a combination of at least three factors, (a) Ion reflection. r.r. Figure 6 the dashed curve represents the

Figure 6 calculated differential range profile of 40 keV P in Au. The distribution indicates that ions exist outside the front face of the taraet and is due to ions being scattered backwards out of the target. For "~P in Au at 40 keV the calculated loss is 16%. Since the measured range distribution is greater than that calculated, the fraction of ions actually refected is probably less than -his. (b! Radiation damage induced diffusion. The buildup of radioactivity from ail accelerated 40 keV beam of Xe ions is shown as a function of time in Figure 7. it h "as been observed (G) that inert gases have near zero retention in plastics and paper. .p-'ylar retained only 0.5>£ of the implanted xe dose. A retention vaJue of 4sputtering. A similar type of activity buildup measurement is shown in Figure 8. This time a larger intensity beam of stable

-' Al BEM . » 'Hi BE«U - _

Figure 7 Figure 8 40Ar was allowed to implant into the area previously implanted with 41 Ar activity. The incoming stable ion beam was effective in sputtering off the surface of a target and removing implanted radioactivity. Further retention studies are shown in Figure 9 in which

100—r.~" ~

•10 kv-V Kr IN TARGET MATERIAL

10 DOSE ilONS/c Figure 9 low doses (• 10 ."• jns/cm ) of Kr were implanted into various target materials. The residual radioactivity was then measured as a function 84 of stable Kr dose. The loss of implanted activity from Cu, Zn, W 15 2 and Au at doses ••_ 10 ions/cm indicates that these materials will have range distributions and ion retention values which are dose dependent. Materials such as C, Al and Si are observed to be less sensitive to implanted doses and are expected to have higher retention values.

Since inert gsses were not retained by plastics whereas metal ions were, the question then arises as to whether damage diffusion favours the release of inert gases from metal targets. Figure 10

DOSE (lONS/cm2)

Figure 10 indicates this is not so. Retention measurements for Au and w gave identical release curves for Kr anrl Rb. Included in Figure 10 are weight loss values as measured for several ion doses. It is evident that more activity was lost for each target than can be accounted for by direct sputtering alone. Retention values of 50% 2 would require the removal of 23 ag/cm of Au or W, i.e. the median range value. The observed weight losses for 50% retention were 2 2 13 ug/cm of Au and ==• 6 »g/cra for W.

It would appear then that the main mechanism operating which restricts the amount of material which can be deposited by ion implantation, is a diffusion of the implanted ions to the surface where they are removed as the collector surface is sputtered away.

TARGET PREPARATION It is therefore generally acknowledged that ion implantation is not a suitable method for preparing targets of amounts greater than a few ug/cm . Let us then examine the production of targets which are within the limits possible by ion implantation. To begin with there are several potential advantages over more conventional deposition methods. 1. The ability to introduce into a variety of substrates not only any element desired, but any available isotope of that element. 2. Precise control over the depth of the implant, by changing the ion acceleration energy, and the amount by monitoring the integrated beam current on the target. 3. Ions enter the target as a directed beam, consequently the position and area of the implant can be accurately defined. 4. Control over the purity of the implanted atoms. Although mass assignment is certain, mass separation does not ensure chemical separation. Fortunately there is a very limited number of elements which can appear at one mass position. By virtue of differences in volatility and chemical behaviour of materials in an ion source, some additional chemical separation is obtained. 5. Uniformity of deposit is achieved by sweeping the beam over the collector. This can be done by varying the magnetic field or by electrostatic deflection. Machines with line focusing have an added advantage in that beam hot spots can be reduced by implanting only a portion of the beam which has been elongated to a line « 2mm x -- 5cm.

I should perhaps expand on the subject of purity. Although separators give a high isotopic enhancement in a single stage, the separation is never perfect. Neighbouring beams contribute to a given mass intensity due to such factors as the separator's finite resolution, aberrations in beam optics, misalignments, high voltage arcs, instability of power supplies, operator mistakes, and chemical effects in the beam. Figure 11 is a photograph of the line images

Figure 11 of Xe isotopes implanted into an Al foil. The foil has been treated to show the implanted position of each separated mass. The lower portion of the photograph is an autoradiograph of the radioactive Xe isotopes in the Al foil. For the CRNL 70 KV separator, the distance between mass 131 and 132 on the foil is 13.4mm. Table 1

Cross Contamination of Mass Separated Beams

(Collected Through a 0.5cm wide y. 1.0cm high aoerturel

Mass 131 collected

Mass position fine focus beam swept beam (2 mm x 2 cm) ( 1 cm x 2 cm)

130 0.48 1.25

131 9B.86 97.56

132 0.66 1.19

Table 1 10

Lists the amount of cross contamination of Xe measured. Although it is possible to optimize the system and obtain smaller values for cross contamination, I feel that the data in TaMe 1 represent actual day to day operating conditions for Xe. The use of swept beams to produce homogeneous implants increases the cross contamination.

Fortunately there are solutions to many of these problems. For example, homogeneity with higher isotopic purity can be obtained by a mechanical motion of the collector placed behind a narrow entrance slit. Although Xe does not occur in nature, a beam is often observed at mass 133 which can be shown to be XeH . Hydride formation in beams is particularly important in separations of inert gases and electronegative elements. Hydride ions are the result of water vapour and hydrocarbons in the ion source. Cleanliness of trie ion source is often neglected since it is assumed that mass separation will eliminate impurities from beams. If elemental implants are required, contamination by adjacent isotopes of the same element is not important. In most cases isotopic purity is required for targets of stable and radioactive materials. Ideally, an ion source should be operated with enriched isotopes for stable nuclides and carrier free activity for radionuclides. Ion source materials prepared by (n,p) reactions should first be chemically processed since the parent material may produce a beam at the same mass position as the product. Activities produced by (n,v) reactions are also troublesome since a pre-chemical processing usually will not help. Target purity must then depend on the enhancement factor obtained through mass separation. Finally, there are certain masses which are troublesome due 28 + to memory effects in the ion source; Si is always contaminated

14 32 + 16 + 3 35 + with N+, S with O 2, and V with Cl .

IO(aN) IMPLANTERadioactivitD TARGEy T APPLICATIONS Except for tritium, thorium and uranium, radioactive materials are not usually used as accelerator targets. However in related work radioactive sources for nuclear measurements can be produced by mass separation - ion implantation. Thin sources with positive II

mass assignment are ideal for photon or particle spectrometry. A detailed description of ion implanted sources for precision 3 spectroscopy was given in a CRNL paper by E«rgstrora et al (7). As we have already seen, it is possible to prepare targets of highly volatile elements such as tritium and the inert gases, keeping in mind that there is a saturation activity effect. A 1 mm source of Kr implanted at 40 kev saturates at ~ 10 ^Ci. Heavy element a sources implanted at 40 keV will have a median range in Al of — 5 jig/cm . This corresponds to an energy shift of 3 kev on the 5476 keV i from 241 Am.

For studies involving hyperfine interaction techniques, ex- treme care must be taken to produce very dilute samples. Ion implantation has the ability to penetrate thin oxide layers on the surface of the host material and deposit the radioactivity as dilute "alloys". Radiation damage accompanying the implantation may have to be annealed out. A comparison of .internal fields measured in samples prepared by ion implantation and melting or diffusion has been made by Reid et al (8).

Nuclear reaction measurements for short lived radioactive targets are limited to those with large cross sections. Recent measurements at chalk River (9) of the thermal neutron capture cross sesectioc n of Xe were made possible by using ion implanted targets.

The largest use of radioactive ion implanted targets is associated with isotope separators on-line with accelerators, which is a topic in itself.

(b) Stable Ion implanted targets of stable nuclides can be used for nuclear studies whenever very thin targets are required, or whenever it is not possible or advantageous to use evaporated targets. We are therefore speaking of reactions which have sufficiently large cross section values. For example, Selin et al (10) used ion implanted targets of Ne and Ar to study radiative resonance capture 2 2 40 reactions, Ne(p,v) and Ar(p,v). Bister and Anttila {11) per- 22 formed Doppler-shift attenuation measurements on ion implanted tie to determine the nuclear lifetime of a state in Na. These are special in that inert gas targets are involved. In Figure 12 is shown a partial excitation curve for the Li(p,n) reaction. The 2 target was prepared by implanting a beam equivalence of 12 ,-g/cm of Li. The actual measured saturation amount was 9.3 ,-g/cm . This thin target has a non fragile surface and can withstand beam currents of 10 i-iamp of 2 MeV protons without any target cooling. These targets are used for precise energy calibration of accelerators. Figure 13 is a spectrum of 2 MeV a's elastically scattered off In

THRESHOLD ENERGt

18 8 0 H e V

2 "b "? I 2 2 2 3 PSUTDN ENERGY (MeV) Figure 12

Figure 13 implanted at 40 keV and 1 MeV into Si. The position and width of the 40 kev peak indicates a thin implant at the surface while the shifted position and increased width of the 1 MeV peak is a measure of the depth and straggling of 1 MeV In ions implanted into Si. Such measurements can provide information on ion ranges and stopping powers. These examples serve to indicate the usefulness of ion implanted targets. Other applications include thin target coulomb excitation and charge particle induced X-ray studies. 13

In summary, ion implantation is most useful for the preparation of thin targets of either radioactive or stable nuclides. For deposits of more than a few ug/cm , other deposition methods are necessary.

REFERENCES (1) Northcliffe, L.C. and Schilling, R.F., Nuclear Data Tables A7_, 233 (1970) . (2) Johnson, W.S. and Gibbons, J.F., Projected Range Statistics in Semiconductors, dist. by Stanford University Bookstore (1969). (3) Lindhard, J., Scharff, M. and Schiott, H.E., Mat. Fys. Medd. Dan. Vid. Selsk. 33_, 1 (1963). (4) Santry, D.C. and Sitter, C.W., proc. Int. Conf. on Electromagnetic Isotope Separators (Marburg), p. 505 Bundesministerium fur Bildung und Wissenchaft report K70-28 (1970). (5) Almen, O. and Bruce, G., Nucl. Inst. and Meth. 1_1, 257 (1961). (6) Santry, D.C. and Westcott, O.M., Nucl. Inst. and Meth. 97, 471 (1971). (7) Bergstrom, I., Brown, F., Davies, J.A., Geiger, J.S., Graham, R.L. and Kelly, R. , Nucl. Inet. and Meth. 2\_, 249 (1963) . (8) Reid, P.G.E., Sott, M. , Stone, N.J., Spanjaard, D. and Bernas, H. , Proc. Conf. Hyperfine Interactions (Asilomar, 1967) . (9) Santry, D.C. and Werner, R.D., J. Nucl. Energy 27_, 409 (1973; (10) Selin, E., Arnell, S.E. and Almen, 0., Nucl. Inst. and Meth. 56, 218 (1967). (11) Bister, M. and Anttila, A., Nucl. Inst. and Meth. 77_, 315 (1970). 14

PREPARATION AND CHARACTERIZATION OF ACTINIDE TARGETS

H. L. Adair and E. H. Kobisk

Isotopes Division Oak Ridge National Laboratory Oak Ridge, Tennessee 37830

Introduction

The preparation of both supported and self-supporting actinide targets in various geometry configurations having a well defined area and thickness is important to the researcher who may need such sources for alpha spectrometry, the measurement of nuclear constants, for neutron dosimetry, or as targets in transmutation experiments. The techniques used to prepare such targets include vacuum evaporation, vacuum reduction-distillation, arc melting, levitation melting, rolling, electroplating, and loading of accurately defined quantities of actinide materials into precision-machined capsules. The various actinide targets produced, together with their thickness range and methods of fabrication, are shown in Table 1.

The amount of radioactive material contained in each sample is determined either by weighing, counting (usually both alpha and gamma counting), or by destructive analysis. In most instances low geometry alpha counting and gamma counting is used to determine the activity contained in each target to ±1%. The uniformity of the oxide deposits or rolled foils is determined by alpha counting small sections of the sample.

Most of the samples are prepared by the vacuum evaporation of the oxide, the vacuum reduction-distillation of the actinide materials, or by loading known amounts of actinide oxides or metals in metal containers. These methods will be described in detail with only a few references to the methods of arc melting, levitation melting, rolling and electroplating.

Vacuum Evaporation

Most of the actinide targets fabricated are prepared by vacuum evaporation. The resulting deposits are normally oxides. Figure 1 is a schematic of one of the vacuum evaporation systems used. The material is evaporated from a tantalum crucible by radio-frequency (r-f) heating and the material is collected on a rotating substrate positioned above the crucible.

In addition to r-f heating, electron bombardment heating is used to evaporate actinide target materials. Figure 2 is a schematic of a commercially available electron bombardment source in which the electron beam is bent through an arc of 270° before, striking the evaporant. This heating source has been employed in the preparation of a large number of 238U, 235U, and 239Pu oxide targets for use in fission chambers. The material is evaporated directly from an oxide pellet, thus eliminating possible contamination from crucible materials. Over the past two 235 239 years forty-six U02 and thirty PuO2 targets have been prepared for use in fission chambers; materials were deposited as either 5- or 7.5-cm diameter spots on either 0.001- or 0.002-cm-thick aluminum backings.

Vacuum evaporation has also been used to prepare 252Cf sources suitable for either neutron or fission-fragment research. The material is vaporized from a tantalum

Research sponsored by the U.S. Atomic Energy Commission under contract with the Union Carbide Corporation. 15

Table 1. Types of methods used in preparation of actinide targets

Range of Method of Form of I'lenient thickness (iig/'cm') preparation Backing deposit

A in '1 to 500 0 '"'• a Ai, Pt, Ti AmO, 10.000 up a. b, d Sel f-supported Am

Cm .-I to S000 : a Al, Ti Cm,'.), ! S,000 up a. b Self-supported Cm

Of •'-' 1 to 10 a Pt. Ni. C, Au Oxychloridc

-'•' 1 tO 1 e Pt, Ni. C, Au Oxychloride

•<- ! to 1 f Pt, Ni Oxychloride

Np •- 1 to 12, 000' '"' a Al, Ti NpO,

Pu •' 1 to 5. 000":' a Al. Ti, Pt PuO, 1S,000 up a, b, d Self-supported Pu

U < 1 to 12,000"" a Al, Ti, Ni, Pt UC, 15,000 up c. d Self-supported U

Other backing materials can be used; those shown represent the backing materials used in preparing various targets desired by physical researchers during the past two years. k Numbers do not represent an upper thickness limit, but rather the maximum thicknesses of targets thus far prepared. Preparation code:

a - vacuum evaporation b - reduction-distillation "bomb" reduction (Ca) d - rolling e self-transfer f - electrodeposition - molecular plating

tubular crucible which is heated by resistance heating; details of this procedure are described in Ref. 2.

All actinide targets except 238U and235 U are prepared in glove box enclosures which are operated at approximately 1-in. negative water pressure with respect to the laboratory environment. The system used for preparing 239Pu targets by the electron bombardment process is shown in Figure 3. A schematic of the glove box system used in preparing Am, Np, and Cm targets is shown in Figure 4, while the system used for preparing 252Cf targets is described in Ref. 2. 16

SUBSTRATE DRIVE

PCU JAR COVER

•-C-PYREX BtLL JAR

RF INDUCTION HEATING COIL, ?IN D Ag PLATEP Cu fueirJG

Fig, 1. Schematic drawing of vacuum evaporation system used for depositing actinide materials.

ELECTRON BEAM

SHIELD EVAPORANT POLE PIECE

WATER COIL- COOLED CRUCIBLE

FILAMENT

ANODE PLATE MOUNTING BLOCK

Fig. 2. Electron bombardment source used for preparation of U and Pu fission foils. 17

Fig. 3. Equipment used for preparation of i39Pu fission foils by the electron bombardment process. DIFFERENTIAL PRESSURE GAGE (0-^m WC] 50 CFW HEPA F.LTER (TVPj STAINLESS STEEL GLOVE BOX •• FLEXIBLE COUPLING TO FILTERED EXHAUST SYSTF> .,- BALANCE •' ENCLOSURE SERVICE PANEL

INLET AIR DUCT

BAG PORT

MECHANICAL PUMP ENCLOSURE

T BAG VACUUM SYS EV PORT '•• J7 A : cc

DIFFUSION PUMP ENCLOSURE

!--jg_ 4. Schematic drawing of the system used for preparation of Am, Np. and Cm targets.

Vacuum Reduction-Distillation

In many instances the researcher requires actinide materials in their elemental "orm. Such actinide metals are normally prepared in large quantities by a "bomb" reduction of the fluoride; however, resulting metals normally contain a considerable number of impurities, e.g., Ca, Zr, Al, Fe, introduced in the process. Bomb-reduced metals can be purified by vacuum distillation, arc melting, and/or levitation melting; for elements such as Am, Pu, and Cm, which have a reasonably high vapor pressure at temperatures <2000°C, the vacuum reduction-distillation process can be used for preparing high purity actinide metals (<500 ppm impurities including oxygen and carbon). Reduction-distillation has been used to prepare 17 grams of 2^3Am, 10 grams of 241Am, 10 grams of 239Pu, and 10 grams of 21^Cm. In each case, the starting material in the form of oxide is mixed with a 50% stoichiometric excess of thorium. After the material has been thoroughly mixed, it is then placed in a chemical reactor-still as shown in the schematic in Figure 5. For the Am and Cm reductions, a tantalum still is used, but for the Pu reductions a tantalum carbide still must be used because Pu metal readily reacts with tantalum. Figure 6 is a schematic representation of the metal reduction system. Distilled metal is collected in a quartz dome from which it can be easily removed after the process is complete. The metal can then be rolled into thin foils or used for preparing alloys.6 A comparison of the typical analysis of 21

QUARTZ COLLECTOR DOME

OOME SUPPORT

RF INDUCTION HEATING COIL

BAFFLES

REACTION CHAMBER

WIRE STANDOFF JOIMT

RECEIVER

CERAMIC SUPPORT

: ig. ->. Schematic representation of reduction-distillation reactor.

Preparation of Actinide Dosimeter Samples

We have established a program for providing well defined oxide and metal accinide samples in support of the Interlaboratory LMFBR Reaction Rate Program (ILRR).8 Oxide samples of 238, 235 U, 237 Np, 239 Pu, 240Pu, and '•Pu as well as stable materials have been loaded into vanadium capsules and are used to monitor the neutron energy spectrum, flux, and fluence in both fast and thermal reactors. In addition to the oxide samples, metal foils of 235U, 238U, 237Np, and 239Pu have been prepared and sealed in aluminum. The procedures used for preparing the samples are described in Refs. 9 and 10. 20

RF INDUCilON HEATING COIL, If 1/8-IN D., Ag PLA- TED Co TUBING-

Fig. (,. Schematic drawing of metal reduction system.

Table 2. Spark sou. ce mass analysis of curium oxide and curium metal

curium (ppm) impurity O09 Cm meta 1

Al 50 30 Ba <20 5 Ca 50 40 Co 40 20 Cr 50 20 i Cu :o 90 Fe 40 60 Kn 20 20 Ni 50 2 Pb - 40 Total rare earths <400 210* Si < IC00 <20 Sn 20 30 la <2D

La, I; Ce, <|; Nd, 18; Sm, 8; Eu, 12; Od, 22; Dy, IOC; Ho, 50. 21

Characterization of Actinide Targets

The amount of radioactive material contained in each target or sample is determined by either weighing, low geometry alpha counting, gamma counting, or by destructive analysis. In some instances a combination of some or all of the above methods may be used.

Weighing is normally a good method for determining the amount of material contained in a sample; however, this requires knowledge of the chemical form of the material before isotope content can be determined. Thus, without a good elemental and mass analysis of the material, weights can sometimes be misleading for determining the amount of radioactive material present.

Most often the technique used for determining the amount of radioisotope contained on a target is that of alpha or gamma counting. For evaporated actinide samples, low geometry alpha counting is used to determine the activity present to <±1%. A schematic of the low geometry counting chamber is shown in Figure 7. For highly radioactive samples, a similar but larger counting chamber is used in which the geometry factor can be reduced to 10~^.

, ELECTRICAL PENETRATION CHAMBER COVER , DETECTOR C/.\\\\\\\\\\\\\1

APERTURE

3 in. 235 TO U02 VACUUM TARGET. PUMP

\ TARGET FRAME

CHAMBER

//////////// ////////777A

Fig. 7. Schematic drawing of low geometry counting chamber. 22

23 > A typical alpha spectrum from a - UO2 deposit is shown in Figure 8. The two large peaks between channel numbers 600 and 800 are the 4.773 and 4.772 MeV ziLU alpha peaks. The 23I*U content of the material, shown in the mass analysis in Table 3, is 0.029%. However, since the specific activity of 23L|U is much greater than the specific activity of 235U, the number of alpha counts from the Z3HU activity approached the total number ot alpha counts from the 23jU material. The remainder of the peaks shown in Figure 8 are due to 23jU and small contributions from 236U and 238U. Thi energy calibration is approximately 1.2 keV/channel. For 2:5 235 most of the UO2 deposits prepared the U content is determined initially by weighing. The samples are then placed in a low geometry alpha counter and spectra

500

400

300 [•

o o 200

100

200 300 400 500 600 700 800 CHANNEL NUMBER Fig. 8. 235U alpha spectrum for a 28 pg/crrr target.

Table 3. Mass analysis of 235U used for fission chamber evaporations

Isotope Atom

2 3 3L, <0.0001

0.029

99.91

0.016

238. 0.044 23

similar to the one shown in Figure 8 are obtained. By having an accurate mass analysis of the material, the 235U content can be determined by integrating the number of counts under the 235U peaks and subtracting out contributions from 236' and 238U. The 235U concentration is also determined from the mass analysis and the amount of 23i|U present. In addition, gamma counting is used to determine the 23SU concentration. Table 4 lists the results of alpha and gamma counting twenty- one 235UO;, deposits; deposits were prepared by evaporation-condensation onto a 7.5 cm spot on 0.2-mm-thick aluminum.

239 As stated previously, thirty PuO2 targets were prepared by vacuum evaporation- condensation for use in fission chambers. The amount of 239Pu contained in each targec was determined by low geometry alpha counting; the counter geometry factor was accurately determined by destructive analysis of one of the fission foils

Table 4. -'SU targets prepared for use in charged particle fission chambers

Target thickness Target thickness Target thickness based on J3''U based on "4U based en alpha data alpha data gamma data farget No. (|jg/cnr) (ug/cm2) djg/cnr ± 1. 0=* 1

63. 8 62. 9 64. 1 7 84. 5 83. 8 85. 9 8 85. 9 85. 0 86. 6 9 111.5 110. 3 1 12. 8 10 113. 0 110. 9 111.6 11 84. 1 83.7 85. 1 12 101. 4 99. 7 100. 8 14 111.5 110. 8 110. 6 16 101. 9 101. 3 103. 0 17 94:8 91.8 92. 9 19 99. 8 99. 4 100. 8 20 91. 1 90. 5 91. 1 21 103. 4 103. 1 98. 3 22 86. 0 84.7 86. 7 23 77. 7 76.7 77.9 25 85. 6 87. 4 86. 5 27 '.00. 2 99.7 99. 3 28 102.4 100. 6 101.7 29 101. 5 100. 4 99. 3 30 107. 1 105. 4 105.7 31 101. 2 100. 5 100. 5

Based on a destructive analysis. 24

Table 5. Mass analysis of 23<)Pu used for fission chamber target preparations

Isotope Atom

38Pu <-0. 0002 "Pu 99.978 40 Pu 0. 021 4lPu 0.0005 "Pu 0.0005

1000 •5.5 1.72 keV/CHANNEL

800 1 1 5.4

600 5.3

400 - 5.2

200- t 5.1

5.0 450 500 550 600 CHANNEL NUMBER

Fig. 9. 9Pu alpha spectrum for a 60 |j.g/crrr target. 25

using a couloraetrLc technique for analysis of the Pu content. Mass analysis of the starting material is shown in Table 5 and a z ; 'Pu alpha spectrum from a 60 ;;g/cnr ' ' JPu (as PuOy) deposit is shown in Figure 9. The spectrum shows the 5.105 and unresolved 5.143 and 5.156 MeV alpha peaks. Table 6 lists the thickness, as determined by alpha counting, of thirty ''-•'iV\i0y_ deposits.

For actinide materials contained in capsules which shield the alpha activity, gamma counting is used to verify the weights of the contained material. An accurate elemental and mass analysis is essential for this verification. The quantity of isotope present is determined initially to <.'.1% by weighing the material and after sealing in capsules, the weight is verified by gamma counting. All procedures used in these determinations are described in Ref. 10.

Uniformity of the vapor-condensed deposits is determined by alpha counting over small areas of the target. The system used for measuring uniformity is shown in Figure 10, while Figure 11 shows the results of the uniformity measurement of a 7.5-cm diameter ^JJPu02 target. The numbers shown in Figure 11 are the number of alpha counts/100 sec x 10"3 at 6-mm increments from the center of the circular deposit. The counting data at the most distant radial positions were obtained at positions of 5-mm from the outer edge of the deposit. As can be seen, the uniformity of 239Pu over a 5.1-cm diameter section of the target was ±2.5%, but at the most distant radial positions far more deviation was observed. These edge non-uniformities appear to be caused by aberrations induced by the circular mask used to define the deposit area during vapor condensation.

Summary

Many well defined actinide samples have been prepared by either the vacuum evaporation of the oxide or by the vacuum reduction-distillation process. Specific examples of targets produced and capabilities are described and many of the methods of accurately determining isotope concentration and uniformity have been described. 26

Table 6. "''Pu content of fission foils as determined by low-geometry alpha counting

Thickness Target No, (pg/cm2)

I 1 S 1. 9 i 106.7 4 81. 1 5 105. 0 b 111.9 7 15. 3 8 57. 5 9 118.6 10 143.6 11 96.7 U 7 9.6 13 87.9 14 144. 1 15 69. 3 16 77.6 17 73. 0 18 70.6 20 95. 1 21 103.7 23 88.2 24 60.4 26 89.7 27 87.2 28 101. 1 29 84.4 30 126.6 33 77.8 34 75.3 35 117.7 36 . 89. 9 27

TRANSPARENT CHAMBER COVER- DETECTOR ELECTRONICS

X-Y DETECTOR\ ADJUSTMENT /

_ » VACUUM CHAMBER 'POSITION INDICATOR T \ - I 5mmi '7 mm2 DETECTOR APERATURE

SAMPLE

MOLECULAR SIEVE TRAPPED 15 cfm MECHANICAL VACUUM PUMP

Fig. 10. Schematic drawing of the apparatus used in determining target uniformity.

180°

360° 0 Fig. 11. Uniformity measurement of a 7. 5-cm diameter Pu target 28

REFERENCES

1. H. L. Adair, Nucl. Instr. S, Meth. 113 (197 3) 545.

2. H. L. Adair and P. R. Kuehn, Nucl. Instr. 6, Meth. 13 4 (1974) 327.

3. J. B. Knighton and R. K. Steunenberg, Argonne National Laboratory Report ANL-7057, Part I (1965).

4. A. V. Hariharan, J. B. Knighton, and R. K. Steunenberg, Argonne National Laboratory Report ANL-7057, Part II (1965).

5. J. B. Knighton and R. K. Steunenberg, Argonne National Laboratory Report ANL-7057, Part III (1965).

6. H. L. Adair, J. Inorg, Nucl. Chera., 1970, Vol. 32, 1173.

7. I. D. Eubanks and M. C. Thompson, Inorg. Nuci. Cheia. Letters, Vol. 5, 187.

S. LMFBR Reaction Rate and Dosimetry 7th Quarterly Progress Report, HEDL TME-73-31, Dec. 1972, Jan.-Feb. 1973.

9. W. D. Box, Third International Symposium on Research Materials for Nuclear Measurements, CONF-711002, October 5-8, 1971, Gatllnburg, Tennessee.

10. H. L. Adair and E. H. Kobisk, Preparation and Characterization of Neutron Dosimeter Materials, a paper presented at the 1973 Winter meeting of the American Nuclear Society, San Francisco, California, to be published in Nuclear Technology. 29

THE CALUTRON AS A SOURCE OF TARGET MATERIALS*

L. 0. Love

Oak Ridge National Laboratory Oak Ridge, Tennessee 37830

On the occasion of the twenty-fifth anniversary of Oak Ridge, General Leslie R. Groves (who had been head of the wartime Manhattan District) visited the old calutron pilot plant and made the comment that when evaluations of the various methods for separating uranium were made, it was thought that the electromagnetic process would be impractical. He also commented that the committees assigned to the task of evaluating the various processes would have recommended against building an electromagnetic facility except that it was backed by Professor E. 0. Lawrence. In time, of course, the faith in Professor Lawrence's method was justified although it was proven that another process was found to be superior for providing large quantities of 235U, and the Y-12 Plant, consisting of over 800 alpha and almost 300 beta calutrons, was shut down. Fortunately for the nuclear industry, however, that was not the end of the story. Some farsighted people in what is now the U.S. Atomic Energy Commission and the Oak Ridge National Laboratory made the decision to use some of the machines for isotope separations and certain research activities, and, as a result, the pilot plant, containing two alpha 122-cm radius and two beta 61-cm radius calutrons, and one beta building, containing 72 calutrons, were retained. Today both of these buildings are used for isotope separations, and together they represent the most versatile isotope separating facility in existence. They have the capability of providing enriched samples of every naturally occurring isotope in the periodic chart and significant quantities of many radioactive species. It is from these unique materials that many of the targets for nuclear research are made.

The first stable element (copper) was separated in late 1945 and since then over 200 kg of enriched isotopes have been collected for use in science and medicine. The elements, both stable and radioactive, that have been processed to acquire these isotopes are shown in Figure 1. The separated material has been divided into three categories:

1. Research Materials Collection Pool for the AEC 2. Sales inventory 3. Unprocessed inventory

The first category contains the separated isotopes retained by the AEC (the prime customer and program supporter) for use in their sponsored programs; these are usually obtained on a loan basis for nondestructive experiments.

Category two contains material available for sale and is purchased by investigators who will normally use it in experiments that frequently destroy its identity. These isotopes are listed with prices and purities in the ORNL Research Materials Catalog (Separated Isotopes, Radioisotopes, Special Preparations), a copy of which can be obtained by request to Oak Ridge National Laboratory, Isotopes Sales Depart- ment, Building 3037, P. 0. Box X, Oak Ridge, Tennessee 37830.

Research sponsored by the U.S. Atomic Energy Commission under contract with the Union Carbide Corporation, Oak Ridge, Tennessee 37830. 30 • 3 FNO Mg_: Si . IT~ J283? __J 5757 'fi .344.S^:7 :;;4 5i' nr~ 2200 •' 2561 .186 4088

Cu • Go As Br-.. Kr ^ : 3517 |628 1663 ::-.•;;' 784 572 ; Rb Sr.'.: .'.: • Zr •'• • • • Nb Tc Ru . Rh 11Pd . -2314 658 8985 • .••:45' 64 Sn .. : . sb.,:::;

••••••••"" '531 :> 4932 '.384

: Cs Bo To •••••',•••" ;w .•;;...:•: --.--. pt ; ::I963 . :.;--294 602; :; \;j427- 370 "48 ,6.7 73

Au Po •648. "2539 S?88 Fr Ra Ac

Fig. 1. Calutron Processed Elements.

The third category contains an estimated 80 kg of separated isotopes which are in various stages of chemical purification and are supplied to investigators when they express a need for them. These enriched isotopes are available because the electromagnetic separator is a mass discriminator and separates the "unwanted" isotopes as well as the "wanted" ones. These "unwanted" materials, their estimated weight, and isotopic purity are listed in Table I.

The ORNL calutron facility as it exists today is depicted schematically in Figure 2. The pilot plant, with its two alpha and two beta calutrons, was used in the electromagnetic development program for the Y-12 Plant during the World War II effort to get235 U, but since has been involved in the separation of stable isotopes for peacetime uses. The remainder of the facility was placed into similar usage about 1960. While the use of the ORNL calutron facility may have been rather inconspicuoi".., many people believe that the separation of stable isotopes is one of the most important single programs for exploiting peaceful uses of the atom sponsored by the AEC; they also believe that the- pilot plant, like the graphite reactor, should eventually be dedicated as one of the nation's noted landmarks.

In order to increase the collection rate of the calutror, each machine was equipped with magnetic focusing lenses called linear shims. In addition to these original linear shims, two other types of magnetic focusing, Table L. Unprocessed StabJe Isotopes

Approx Approx Approx Approx Isotope Purity Quantity Isotope Purity Quantity (%) (g) (%) (g)

34Mg 99.9 1.160 92Mo 97.5 940 94 28Si 99.8 3,500 Mo 92.0 600 32S 99.7 750 "Mo 96.4 1,030 99.9 1,340 96Mo 96.6 1,140 40Ca <99.9 32,695 97Mo 94.0 630 44Ca 98.5 79 looMo 97.2 620 48Ti 99.?. 1,200 nocd 97.0 65 52Cr 99.8 2,500 1MCd 95.0 65 54Fe 95.0 820 116Cd ?~* 0 40 56Fe 98-99.8 15,000 120Sn 98.0 425 58Ni 99.8 4,750 128Te 99.2 345 130 60Nj 99.5 790 Te 99.5 475 64Zn 99.5 580 138Ba 99.8 750 139 66Zn 98.5 340 La <99.95 172 7»Se 98.6 48 140Ce 99.8 640 80Se 99.4 184 142Nd 97.5 195 86Sr 96.4 750 '""Nd 97.5 225 87Sr 85.0 350 M6Nd 97.5 160 88Sr 99.8 5,270 181Ta <99.99 1,200

SEPARATOR PROPOSED YOKE

N MAGNET EXCITATION • MAGNET YOKE COIL (D

SEPARATORS IN USE: (5)STAELE ISOTOPE SEPARATIONS mmm* 180°; RADIUS 61cm (R)RADIOACTIVE ISOTOPE SEPARATIONS I 1 180° ; RADIUS (22 cm illll'M 255°; RADIUS 51 cm '•' ••••••••• J 180°, n = 0.8; RADIUS 61cm

SEPARATORS IN STANDBY CONDITION : 255° i 1 180°; RADIUS 61cm

The Upper Blocks Represent the Two Tracks of Separators Housed in Bldg. 9204-3, Indicating the Relative Positions of the Three Types in use, While the Lower Blocks Represent the Four Separators (Two Types) Contained in Bldg. 9731. The Geometrical Designs in the Lower Left Depict the Beam Profile Produced by Each of the Three Types in use.

Fig. 2. Schematic Representation of 0RNL EM Isotope Separation Facilities. 3 2

have been incorporated into the overall program in order to increase the versatility of the separation facility. The first of these was the 255° modification (n = 0.5) which has an effective radius of ^40 inches ('102 cm) and increases the dispersion between masses by 1.4 times. These machines have been an excellent addition to the program since total ion output is similar to the calutron "bile assay purity is usuallv higher especially for masses greater than 100.

The second system added is a sector which is unique to the regular calutron in that the ion beam focuses outside the magnetic fieiri. The sector, which we refer to as the Oak Ridge Sector Isotope Separator (ORSIS), was mounted in a spare calutron tank and has a beam deflection of ISO" with inhomogeneity of n = 0.8. A schematic of this machine is shown in Figure 3. The objective of building this instrument was to wed the high-throughput features of the calutron with the high-purity capabilities of the European type separators. In addition to the focal point being almost 3 meters from that of the calutron ion beam, and as seen from the insert in Figure 3, the mass dispersion is over seven times as great. Currently, the ORSIS is being equipped with automatic controls for computer-assisted operation. A second machine of this type is under construction, and, when completed, will materially increase the capability of the electromagnetic facility to provide more samples of highly enriched isotopes in a single pass. Typical of the material expected is indicated in Table II, which compares the isotopic purities of the calutron with those obtained in the 180° sector separator.

Another desirable feature of the sector is its ability to make ion implants at energies ranging from a few volts to 160 keV. A holder that can accomodate 60 substrates is shown in Figure 4; similar devices have been used for making as few as two and as many as 100 targets in one run. It is not generally known, but between 15-20 thousand ion implants have been produced in the Oak Ridge separators. So far, =550 yg/cm2 is the quantity usually deposited per target, but efforts are being made to increase the quantity deposited; some progress has been made as can be seen in Table III. To the extent that this can be done, target costs will be reduced because the separation and deposition can be made in one step.

The electromagnetic separator will, if conditions are right, resolve the isotopes of each ion beam of an element into its respective masses. However, the ion production mechanism is complex and the degree of separation can be no better than that dictated by the energy spread existing in the positively charged ion beam. Since this spread is not reduced to zero the separation is not perfect. Figure 5 shows actual pictures of oscilloscope traces that recorded beam fluctuation during the initial stages of a calcium run. The time interval involved in the start-up of this run was about four hours and the monitored beam was ^"Ca. The pictures show the beam intensity fluctuations and corresponding focal property as it increased from an output of about 15 mA in the early stages to approximately 150 mA at its maximum. The disturbances during this period are one of the reasons for the less-than-perfect isotopic purities that we obtain. But even with imperfec- tion such as this, the calutrons still provide isotopic purities that are very high when compared to other methods. For example, 1*8Ca was enriched from 0-18% to >96%, and 46Ca from 0.0033% to about 50% in runs similar to those from which the pictures were made.

In most instances the attainment of high chemical parity presents greater problems than those associated with attaining the desired isotopic purity. For example, we collected about 100 grams each of >95% 29Si and 30Si in 1963 for growing single crystal wafer and for use in fabricating lithium-coated barrier-layer detectors. It was thought that single crystal 30Si would greatly improve the counting efficiency of neutron spectrometers if the boron cor.tant could be held to less than one part per billion. The problem of reducing the boron content in these isotopic samples to an acceptable level for these uses has not yet been solved. M, LIGHTER iAOTO

LK"*^''" ' ---"-'"••"a-^.-.-.i,.-.-

RECEIVER —•' SOURCE +35 kV —

; I _ J ...... 1 t' urn (3C° ORS 5 ? i 10 •i b< m

LUSPL WSION AT !A N Rf ( f VfcP

Fig. 3. Schematic of the 180° ORIS. 34

Table 2.

COMPARISON OF SEPARATIONS MADE WITH THE ORNL 180° (n ~- 0.8) SECTOR AND THE CALUTRON

180° ORSIS CALUTRON NA Total Beam Total Beam Isotope Assay Assay (%) Current Current C«) (mA) o) (mA)

Fig. 4. target holder. Table III

Samples of Targets Prepared in Oak Ridge Electromagnetic Separators

2 Isotope Back Energy (keV) Atoms/cm

U9Sn SS 0.050 ic19

158Gd C 0.200 io17

176Hf Ni 0.850 io18

3H Au 2.5 io13

85Kr Al 5.0 io13

69Ga Ge 10.0 io16

194Pt Ni 40.0 io17

U B Si 80.0 io16

31 P Si 160.0 io16

Fig. 5. Oscilloscope Trace of Beam During Calutron Start-Up. 36

Some sources of chemical impurities are inherent in the electromagnetic process. For example, elements present in the ion source components are heated to their vapor state and find their way into the ion production arc. These spurious ions are accelerated into the ion beam along with the wanted ions. Energy changes and charge exchange result in trace amounts of these unwanted atoms finding their way into the collector pocket, making problems for the chemist to remove, and if not removed, for the user to tolerate. The same thing happens when there are chemical impurities in the charge material. It is impractical to try to purify each charge to analytical grade material and, as a result, impurities in trace amounts may find their way into the collector pocket and again make problems for the chemist and the user.

On the other hand, to meet high isotopic purity requirements, it is only necessary to make second or third passes in the spearator using for charge that material which was separated before. It must be added that this approach to high purities is not without penalties because the enriched charge material for these runs is usually expensive and the low efficiency of the calutron (10-20%) means that the cost of a second and third-pass sample may be quite high. An example of this cost is seer, by examining the economics of a second-pass run made with about 12 grams of 6% is0W to obtain approximately 20 mg of >95% 180W at the nominal price of $200 per mg. In this case the charge was made up of material sputtered from the *^W collector pocket and deposited on the receiver defining plate, and was incidental to a 70,000-hour collection series made for the purpose of getting pound quantities of 95% !B1*W. The actual time of the second-pass collection involved only a little more than 10 hours, but getting the 12 grams of 6% charge would have made the collection prohibitively expensive if the runs had been made for the 0.14% naturally occurring 18UW alone.

An example of the technical ease with which the isotopic impurities can be reduced is a case where we had a request for 50 grams of 7-3-iU with the 232U content reduced to 1-10 ppb. The separation was made and the 232 was reduced to 4 ppb. The investigators began their experiment, then called back and said that they had missed their calculation — the 232 content was eight times too high. All that was necessary was to make another pass and the unwanted isotope was reduced to 0.1 - 0.2 ppb.

It has been gratifying to watch the interest grow in separated isotopes since the first separation was made in the mid-1940's. Figure 6 is a graph showing the trend in the use of separated isotopes based on the number of shipments made by years. The decline in usage in 1973 parallels a forced decline in separation effort. In 1969 we used 225,000 separator hours, and by 1974 the effort had dropped to less than 100,000 hours.

Another interesting, but not surprising, aspect of isotope separation has been the changing emphasis on various isotopes from year to year, with no predictable way of determining which will become important at a given period of time. Figure 7 is a graph showing the variation in interest with time in the iron isotopes. When Professor Mossbauer made the discovery that some energy was absorbed in the nuclei after a nuclear reaction, interest in 57Fe exploded. Several investigators wanted this 2.4% abundant isotope in hundred-gram quantities and since then we have collected kilograms of the iron isotopes. When the separation got under way, I called an investigator at one of the national labs and asked if he did not need kilogram quantities of 56Fe. He said no, but he would like to have two pounds of 5I*Fe and gave me the name of a friend in another lab to contact about the 56Fe. On learning that kilogram quantities of this 92% abundant isotope were going to be available, he wrote to the AEC Division of Physical Research and said that he could do an experiment with 7 kg, but he could do a useful experiment if he had 15 kg. Figure 8 is a picture of the chemist preparing a 15-kg sample of 99.9% 56Fe. 37

4000

3200

2400 LU 0_ 1600

800

1946 48 50 52 54 56 58 60 62 64 66 68 70 72 YEARS Fig. 6. Shipments of Electromagnetically Separated Stable 'sotopes. (1946-1973)

130 I J 65

120 iTJTJ f" ^ 60 - 54 Fe | I 110 56 55 Fe | ' 57Fe i 100 r 50 _ 58Fe 01 1 P p&l 1 ITRflW T! i " 5- 90 /|--r 45! 1 TANK HOURS A to - ! - < Q 80 40 g

- - X j , K y 70 i / 35 — / CO — - / f a: 2 60 30 g - \ - 50 V V r 25 § LLI — - DQ / 20 3 40 DISCOVERY OF ^ / "MOSSBAUER I 15 30 EFFECT" -~ 1 /

/ .-• 1 10 20 - ***•i, i / .^/ A 5 10 - .- 4^*?&h I i 0 0 48 50 52 54 56 58 60 62 64 66 68 70 YEAR Fig. 7. Trend in Separation and Use of Iron Isotopes Enriched in the ORNL Calutron. 38

Fig. 8. Recovery and Refinement of 56Fe.

I have mentioned milligram to kilogram quantities of separated isotopes and 50 Mg/cm?- targets from the calutrons, but the true versatility of the machine is perhaps more vivicly illustrated by the following story. At the time that we were preparing for the iron separation, we were also interested in seeing what we could do with milligram quantities of charge. An experiment had been planned that was to use 1 mg of uranium and we were discussing it when the charge preparation chemist commented that he had just ordered a ton of iron chloride and would order another ton when the separation got under way. The milligram charge of uranium ran for one hour and forty minutes, and the ton quantities of iron for almost 100,000 hours. Thus, the calutron is a versatile machine and illustrates the extent to which the ORNL Electromagnetic Program is usually capable of providing useful quantities of highly enriched isotopes from milligram to kilogram sizes of charge of a specific element for use in nuclear research and medicine. We dispense these isotopes primarily from an on-hand inventory, but if there are items needed which are not listed in the stock catalog, inquiries concerning those materials will be welcomed. 39

PREPARATION OF THIN ACTINIDE TARGETS 1-OR THE SYNTHESIS OK THE TRANSACTINIDE ELEMENTS*

R. Lougheed and E. K. Hulet

Lawrence LLvermore Laboratory, University of California Livermore, California

INTRODUCTION

? 4 ft ^49 In this paper we describe the preparation of Cm and "" Cf targets used for the study of element 104 chemistry and the discovery of element 106. The preparation of transplutonium targets for the synthesis of transact inide elements has presented many problems in the past. Because the reaction cross sections are very small, these experiments require targets that are deposited over a small area (typically 0.2 to 0.3 cm") with surface densities near the recoil range of the nuclear products. The scarcity of target material requires good deposition yields and total recovery of the target isotopes. The substrate and target must withstand the intense heating from heavy-ion beams, and the target must be free of chemical impurities that produce interfering nuclear products. o Previously, targets for such experiments were prepared by electrodeposition;" however, as larger amounts of Cm and ^Cf became available, the demands increased for targets with densities of 1 mg/cm or larger. Electroplated targets are especially difficult to prepare at these high densities and small target areas, especially with the limited number of substrates suitable for the experiments.

We have prepared 248^m ancj 249Q£ targets with surface densities greater than 1 mg/cm by sublimation of their fluorides onto healed metal substrates. These targets are free of detectable Pb, Bi, and ])g and adhere well at higher surface densities than electroplated targets. The alpha spectrum of recoil products from these targets no longer show alpha activities from the light isotopes of Kr, Ra, Po, and At which earlier obscured the detection of new elements. The deposits are silvery and visually take on the surface features and grain structure of the substrates. From visual observations of the deposits on various substrates, ESCA scanning, and the calculation of free energies, we suggest the following reactions on Be and Al substrates:

This work was performed under the auspices of the U.S. Atomic Energy Commission. 40

880°K 5Bc(e) + :An!',(g) £ 2An(c) + 3BeF_,(g) AF = - 15 kcal

798° K Al(c) + AnF-Cg) ;v_ An(c) + AlF.(c) AF = - 57 kcal

!;.\PHRIMi;.\TAL

Crucibles and Charge Preparation

The vitreous carbon crucibles (Fig. 1) were machined by the electrical- discharge technique. Carbon was chosen to reduce the possible contamination of the targets with heavy metals and for its chemical inertness. The vitreous form was used because it is nonporous.

Fig. 1. Vitreous-carbon crucible (4-mm I.D.) and Cap (1-mm orifice).

The crucibles were filled with - 100 p£ of HC1 solution containing the charge material. An excess of HF was added to the crucible to precipitate the fluoride, and the mixture evaporated to dryness using a fine stream of air and a heat lamp. A gentle air stream was necessary to avoid blowing fluoride particles out of the crucible at dryness. When the fluoride was thoroughly dry, the press- cap was fitted into the crucible. 41

Subi imati on

The crucible was supported in the sublimation apparatus (I:ig. -J by a base of pyrolytic graphite and heated using a 2.5 kIV, 43 kil~ Ri- power su]>ply. I'yrolytic graphite was used because it does not couple well to Rl:.

Fig. 2. Vacuum sublimation apparatus.

The crucible was slowly heated in vacuum (10 to 10 ' torr) to 950°C, as measured at the crucible top with an optical pyrometer, to drive off volatile impurities. The vapor pressure of the actinide fluoride was calculated to be ~ 1O'"J mm (Hg) at this temperature. Usually about 0.1 to 0.5 percent of the fluoride was sublimed during this prefiring step.

After the crucible had cooled the apparatus was brought to 1 atm of Ar. The target substrate was then inserted, the bell jar evacuated, and the substrate heated to - 500°C. The fluorides were sublimed by raising the temperature from 1000 to 1200°C (~ 10 " mm vapor pressure) over a period of about one hour. Slow deposition, achieved by effusion from the 1-mm-diam by 5-mm-long crucible orifice, was necessary to obtain a smooth target surface. The observed deposition rates agreed with rates calculated from the vapor pressure of Ami-"-'5 and sublimation calculations for our diffusion cell. ' 42

The targets were counted in low-geometry alpha counters to determine deposition yield, which ranged from 25 to 38 percent, depending on the distance of the substrate from the crucible. These yields are equal to or better than the yields expected for volatilisation from a 2~ source.

l;oi 1 Stack

Tig. 3 shows the substrate configuration for a 0.01-nun Al foil.

]"*-Water cooled copper clamp - Cu heat shield

Ni heating- :r:r> block Heater

TC guage- Graphite paper Al substrate Al rings Mica S/S collimator -S/S

Quartz shield

•Vitreous carbon crucible 7 8 9 10 H 1 1 f 10 cm

Foil stack.

The mica ring was inserted to insulate the collimator from the substrate. The Ni heating block was large so that it could act as a heat sink for radiant heat coming from the crucible which might overheat the substrate foil; otherwise the substrate reached 500°C from absorbing radiant heat with the crucible 4 to 8 nun away. A copper ring and water-cooled copper plate surrounded the foils to capture radiant heat and to clamp the foils in place. The S/S collimator and base were removable to aid the recovery of the target material. In addition, a quartz shield surrounded the upper portion of the RI-" coils to capture additional sublimed material. Recovery was nearly 100 percent; however, some target material (- 10%) was left in the crucible, but tliis amount does not increase after repeated use. The graphite paper (0.4 mm) and Al ring (under the Al foil) were not used in preparing targets on Be foils. These materials were added to prevent sticking of Al foils to the Ni heater and mica rir.g. The Ai foil was allowed to stick to the \1 ring and both were mounted in target holders for Siiper-III I AC bombardments.

DISCUSSION

Target Character i st ics

l-'ig. 1 shows JS!) ;.g of " Cm deposited on 0.013 -mm Be, with an average- surface density of 1.3 i;ig/cm~. liven at this thickness the curium appeared to take on the surface appearance of the substrate. If the substrate was not heated during deposition, the material formed large crystals which flaked off, much lik-- thick electroplated samples. Because the crucible was close to the substrate, the deposits were considerably thicker at the center. Pinhole camera pictures ivere taken of two Cf targets to measure density variations. The density at - 90 percent of the target diameter, as defined by the coliimator, was approximately a third of the density at the center of the deposit when we used a crucible to substrate distance of - 4 mm and a collimator diameter of 4.7 mm. Later targets were made at greater distances to achieve more uniform densities.

Fig. 4. 250-pg 248Cm on Be.

The sublimed trifluorides deposited on Be and Al substrates were very ndherent. This adherence was demonstrated when unintentional, localized focusing of the *°0 beam on a *-"Cf target raised a small "volcano" in the target, and still no -4-T.f was lost. However, Cf deposited on a Ti foil began to peel within two hours after preparation even though similar rare earth targets were adherent months later. We calculated that the free energy for reduction of Cfl:_ by Ti is positive and, therefore, no reduction should occur. ihe failure of this reaction to occur may explain the poor adhesion in this case. b Rare earth fluorides deposited on Be were analyzed by x-ray diffraction, I.SCA,' and Auger electron analysis.' X-ray diffraction showed the bulk of the material to be X rifluoride. Auger electron analysis of the target surface showed the presence of carbon and oxygen iv addition to the rare earths and fluorine. Some ot: the rare earths at the surface were possibly in the form of oxides or oxyf luor i des . l-.SCA, which has an effective penetration of - 7.0 A, indicated that tlv; rare earths were present in several oxidation states, and that a portion of the rare earths were not chemically bonded, except perhaps as inter- metal lies. Beryl 1 ui;;i was also observed in the liSCA work, even though the deposit thickness was more than a hundred times the liSCA penetration depth. Although the .\ rfaces of Be and Al foils were oxidized, the actinide fluorides apparently rt icted at SuO to 900°K with the metal substrate.

Target Purity

As noted earlier, traces of llg, I'b, and Bi have obscured the detection of new elements • 1 the past, even though extensive chemical separations were performed c . the target elements. We purified -'*°Cm and '" ~Cf by using ultra- pure reagents and ion-exchange chemistry. However, there is evidence that ion- exchange resins contain small amounts of Pb which cannot be completely removed, o and which elutc from the columns at constant but low levels. Further decontamination is possible using solvent extraction, but this was found to be unnecessary because of the additional decontamination provided by sublimation. Fig. 5 provides a comparison between the alpha spectra of products formed from bombardments of a 60-ug " 'Cf target prepared by electrodeposition on a Be foil and of a 259-;_ig " Cf target prepared by sublimation onto Al. These data were normalized so that the yields of 25t>No were identical. The electroplated target contained 10-30 ng of Pb compared to < 0.2 ng of Pb and more than four times as much californium in the sublimed target. Extensive chemical separations had been oerformed on the yCf used in the electroplated target to reduce Pb contamination. A minimum alpha background in the region of S.7 to 8.9 MeV was particularly important since this alpha energy range showed the alpha decay of ii)wRf, the alpha-decay daughter of 106. Small amounts of lead would produce Po and At alpha activities in this energy region. 45

io5 ' Fin ?4Q f\ 254 " LLL Vaporized Cf taroet 104 •/' •'", 250 --LBL Electroplated 249Cf /' V 'A Fm target ' / , 256No i \

OCA 9KCJ ,' . /

loo. HiiLJiL ,, .- 6.8 7.2 7.6 8.0 8.4 8.8 9.2 Energy (MeV)

I-'ig. 5. Alpha spectra of nuclides produced from electroplated and sublimed targets of 249Cf.

The choice of substrates is severely limited in these experiments and is determined by the mechanical strength, thermal conductivity, energy lost by beams of charged particles, and freedom from interferences from nuclides produced by nuclear reactions in the substrate metal and impurities. High purity Be (0.013 minj has been the substrate of choice for these experiments. Californium and curium were deposited on Be in producing earlier targets and those were shown to be Pb-free. However, in the synthesis of element 106, which used ^ Cf, Li was made from the reaction of 0 ions with Be. The Li (t-i/2 = 0-85 s) decays with the emission of a high-energy beta particle which is in coincidence with IOK- onergy alpha particles from the decay of an excited state in ^Be. Because the energies of these particles were summed in the alpha detectors, they constituted a background radioactivity seriously interfering with the detection of alpha particles from the decay of ^ 106. Californium-249 targets were then prepared on Al to eliminate this background. In part, the discovery of element 106 was made possible by increasing the -'^(.'f surface density and by reducing Pb, Bi, and Jig impurities in the target.

SUMMARY

Our sublimation method has proved to be a valuable new procedure for preparing thick, adherent actinide targets having a high chemical purity. Deposition yields are acceptable, especially since the charge material is easily recoverable. 46

Observation of the deposits on different metal substrates, the hSCA data, and free energy calculations indicate that the very good adherence to Be and Al substrates may be due to a surface reaction of the actinide fluoride with the metal substrate to form a thin layer of actinide metal or an intermetallic between the substrate and the deposited fluoride. We are continuing the physical and chemical analyses of these sublimed deposits.

ACKNOWLEDGMENTS

he would like to acknowledge Arthur Bolson for constructing and maintaining the sublimation apparatus.

REFERENCES

1. A. (ihiorso, E. K. llulet, J. M. Nitschke, J. R. Alonso, R. W. Lougheed, C. T. Alonso, M. Nurmia, and G. T. Seaborg, submitted to Phys. Rev. Letters (Sept. 197-1).

2. J. E. Evans, R. W. Lougheed, M. S. Coops, R. W. Hoff, and E. R. Hulet, The Use of Electrodeposition Methods to Prepare Actinide Targets for Cross- Section Measurements and Accelerator Bombardments, U.S. At. Energ. Comm. Rept. CONE-711002, 21 (1971). Nat. Tech. Inf. Service, U.S. Dept. of Commerce, Springfield, Virginia.

3. D. Brown, Hal ides of the Lanthanides and Actinides (John Wiley 5 Sons, Ltd., London) 1971.

4. S. Dushman, Scientific Foundations of Vacuum Technique (John Wiley 5 Sons, Inc., New York) 1962.

5. D. Schulz and A. Searcy, J. Chem. Phys, 36^, 3099 (1962).

6. J. R. Peterson, Univ. of Tenn., Knoxville, Tenn, private communication.

7. C. A. Colmenares, Lawrence Livermore Laboratory, Livermore, Calif., private communication.

8. K. A. Gavrilov, Yu. S. Korotkin, and Ya. Shukurov, Preprint of the United Institute of Nuclear Research, P13-3380, Dubna (1967).

9. A. Ghiorso et al., Lawrence Berkeley Laboratory, Berkeley, Calif., private communication. 47

Tailoring of targets for a tandem accelerator laboratory G. Sletten The Niels Bohr Institute University of Copenhagen, Denmark Abstract The organization of a target laboratory serving the nuclear physics research at a tandem van de Graaff accelerator is described. Emphasis will be put on the layout of the laboratory and the mode of operation. The working force is about 40 h per week shared by two technical assistants, and they are supervised by a physicist who on the average spends about 1/3 of his time on target-related problems. The standard techniques employed will be presented briefly, but selected topics like heavy ion sputtering of actinides and the preparation of multilayer targets will be described in detail. A main feature in the setup is a general target bank or library for the whole laboratory. T.t has been attempted to have a ready made target of as many isotopes as possible of most elements for any experimenter at all tines. The basic framework of this system has worked successfully for about 10 years. Introduction Papers given at target development conferences mostly deal with specific techniques and preparations. Therefore, I have felt it important to describe a complete target laboratory where all these single operations are carried out. I also feel that it has a certain value to present a technique or preparation with some reference to the experiment for which it has been designed. This will therefore be the spirit of this talk. The target laboratory of our institute came into operation during 1961 when our first tandem van de Graaff was ready for experiments. With a staff of a dozen Danish physicists and as many visiting physicists, the importance of a standardized and well organized target supply was realized. One of the physicists on the staff took the responsibility for development and daily care for all target related problems, and it was stressed that this person also ought to be active in the experiments at the accelerator. In the years that have followed, we have learned to appreciate this decision, and it is probably also the reason why we have a very flexible and well organized target service today. With an active experimenter as leader, the understanding for different requirements concerning a target has been appreciated by those making it, and many unnecessary misunderstandings have been avoided. The Target Laboratory In the following, I will describe the main features of oux target laboratory and how it is working. One of the first things that was decided was to have one standard target frame, fig.l. It is made of 0.5 ram aluminium and has a window of 0.8 cm . The frames are numbered from 1 up to about 3000. 4 P.

u u

E E QD

12mm m

L74154 Fig.l. Standard target frame. All targets have their place in a target library in boxes, each one containing 200 numbers. The keys to this library are a target file and a target catalog as indicated on the diagram, fig.2. The file contains all available isotopes of all elements with separate cards for each. On the card the index is by frame number. The following columns list method of preparation, backing, thickness and different comments like isotope sample (ORNL nr.), date of preparation etc.

Target File Target Catalog Index by element Index by frame no.

Target Library Targets ordered by frame no.

High Vacuum Radioactive Storage Storage

Fig.2. Organization of the target collection. 49

The target catalog is a listing of numbers from 1 to 3000 and by sach number there are columns for the isotope, method of preparation, backing and so forth. There are many targets that corrodize rapidly in air and cannot be stored in the target library. These targets are all stored at a pressure of lxlO~6 ram Hg in 3 glass jars, each one 30 cm wide and 12 cm high. These targets are ail replaced by dummies in the target library, indicating in which vacuum chamber they are stored. Radioactive targets and actinide targets are stored separately in vacuum jars or dry boxes, but they are all registered in the general file and catalog and have their dummy in the target library. The experimenter can now localize a target of a given isotope by the file and rapidly get all relevant information about "t. If the user later on should forget what the target contains, he can by the number enter the target catalog and identify it. After an experiment the targets are returned to the target laboratory whether it is damaged or not, and inspected by the personnel there before they are put back into storage. Targets that are damaged, are thrown out and erased in both the file and the catalog. The number, however, is used over again and in some cases even the frame. This might be a weak point in the procedure, but as long as the re-distribution of frames and numbers is restricted to those who scrap the damaged targets, we feel that the method is safe.

Isotope File and Certificates

Isotope Storage Preparation's Logbook

Target and Chemistry

Laboratory

Target File Target Catalog

Fig.3. Target laboratory organization, In fig. 3, the connections between the input of isotope and the target library are displayed. The log book is particularly important because the experimenter can get an index to it in the target file and find out about details in the preparation that might be important for his results. Everybody that has responsibility for a target laboratory knows that many physicists want to have their own targets, with their own name on the box, in their own drawer of their own office desk. One therefore would argue that a system as described above would break down because of long term loans and people having the targets, they were interested in, stored in their own rooms. The thing to do is to convince all experimentalists about the advantage of a common pool of targets. He will himself have access to all targets available in the laboratory , and at the same time the target makers will be able to spend more time on developing new and rare targets. In general, our system works well, but it is necessary from time to time to give people a reminder. Our motto is that we will do anything to promote physics, and with that in mind we think that the rigour of our system is justified. The working force of the target laboratory consists of two technical assistants both working part time and together covering about 4 0 hours per week. They work under guidance of a physicist that devotes about 1/3 of his time to target related problems. The target laboratory itself is physically divided in two, one for inactive targets and one for actinide targets and other radioactive targets. Both sections are equipped with vacuum evaporation units and heavy ion sputtering apparatus as general tools.

CONDENSER PLATE (rUNGSTEN)

METALLIC CONDENSATE

THERMAL BAFFLE (fANTALUM)

REACTION MIXTURE

FILAMENT SPIRAL ITUNOSTEN)

HIGH VOLTAGE 0 - 2000 VOLTS

Fig.4. High efficiency electron bombardment assembly for T=3000°C. Here set up for meta]lothermic reduction. In addition to this, the inactive target laboratory has the following general equipment: 1. Pack roller for production of self supported metal targets. 2. Evaporation unit for carbon foils . 3. Argon glove-box with pack rolling equipment. 4. High frequency oven for reduction of oxidss and melting. The vacuum evaporation units are all built in the laboratory, ref.l. The one used for inactive work is shown in fig. 4. It employs electron bombardment and can handle very small amounts with very high collection efficiency. The radioactive evaporation unit is displayed in fig.5 and has many features in common with the inactive one, ref.2. What is special is that parts can easily be taken out and changed or stored away v/hen they get contaminated. The whole thing is of course contained in a glove box shielded with absolute filters.

Fig.5. Electron bombardment unit for radioactive materials. The heavy ion sputtering apparatus is shown in fig. 6 and is identical for active and inactive work,, ref.3. The radioactive sputter unit is contained in a glove box as shown in fig.7. 52

< i

Fig.6. Heavy ion sputtering apparatus with high collection efficiency. The beam energy is 10 keV and a typical current about 1 mA. The beam stop can be rotated to obtain uniform targets. Fig.7. Glove-box with heavy ion sputtering equipment for actinide elements. Preparation of Radioactive Targets After this general description of our target laboratory I will carry on by mentioning three major achievements in more detail. They all have to do with radioactive preparations although the first one concerns an almost stable isotope: 1. preparation of "* 1 Ca targets 2. actinide plunger targets 3. actinide, lead multilayer sandwich targets 1. Ca Targets The '''Ca isotope with a half life of 8-101* years was produced in a Savannah River Reactor in the U.S.A. and mass separated at ORNL by Dr. L.O.Love and coworkers. We obtained two batches of about 2 mg "* : Ca each and prepared 3 sets of targets from this material. Two preparations were carried out at the 1 mg scale and the remaining material in one. The isotope was received in dilute nitric acid solution and the calcium was precipitated as the oxalate. After conversion to the oxide it was vacuum evaporated onto 30 yg/cm2 carbon backings from a tantalum crucible. The calcium oxide is reduced to the metal by the tantalum crucible and the deposits were shiny metallic films. To collect as much of this unique material as possible, 18 carbon foils were mounted per evaporation. The result was about 10 targets with thicknesses about 100 yg/cm2 and dozens of thinner targets down to a few micrograms per square centimeter. With the 2 mg batch serveral targets of 200 to 300 ug/cm2 were made. With thick targets it is important to avoid contact with the atmosphere because after oxidation reentry into vacuum would 54

cause peeling due to expansion of CC>2 and I^O. We therefore transferred these targets to an argon glove-box without breaking the vacuum. There they were stored in an atmosphere containing less than 5 ppm of 02 and 1^0 and have been transferred to the experimental area in argon filled containers. These targets have been involved in experiments not only in our own laboratory, but also on this continent, and all thinkable scattering and transfer reactions have been made with them. They have been used in Munich, Groningen, Brookhaven, Rochester N.Y., ref.4, and Los Alamos to mention the most prominent ones. 2. Actinide Plunger Targets The research programs dealing with fission require actinide targets, and we have solved the safety problems connected v.'ith those by establishing a special target laboratory for actinides. As indicated above, we use the same standard techniques as j.n the inactive laboratory, but all apparatus and desks for manipulation are enclosed in glove-boxes. Actinide targets on carbon and metal backings can as a rule be quite successfully prepared by vacuum evaporation, ref.2. For some preparations, however, the large amount of heat from electron bombardment is hazardous. This inconvenience can however be over- come by the use of heavy ion sputtering. The recoil distance method has quite successfully been applied to the study of fission isomers with half lives of a few nanoseconds, as the lower limit. By application of plunger targets it has been possible to study isomers in the 10 picosecond range, and one has thereby been able to complete the systematics of fission isomer half lives in even-even plutonium nuclei. To look for fission isomers is to look for a needle in a haystack because the probability of creating one is a million times less than the probability for prompt fission which takes place at the same time. In the recoil distance method we take advantage of a geometrical separation of the delayed fission events from the prompt. The detection system works in such a way that only those nuclei which recoil out of the target and then decay by fission are detected as true delayed fissions. The uniform and very flat surface of a plunger gives a sharp time zero, and the recoil velocity determines the time scale. In the reactions used most frequently the recoil velocity is about 0.6 nm/ns. When we now talk about isomers with half lives of 5 to 10 picoseconds, it means recoil lengths of about 3 pm. There- fore the limitations on the half life sensitivity are determined by the flatness of the target. Fig.8 shows the suspended surface of a plunger target. The interference picture gives an idea of its flatness. An ideal object would give straight parallel interference fringes whereas deviations from straight lines indicate an unevenness in the target surface. Deviations of the order of the distance between the fringes corre- spond to a difference in height of 0.3 pm. In the centre this particular target is planer than 0.5 pm. Fig.8. Interference picture of the suspended surface of a plunger target. For details see text. With a recoil velocity of 0.6 ram/ns and this quality of plungers, it is the thickness of the deposited target material that puts the lower limit for half lives. Typical thicknesses are 30-60 ug./cm2 of actinides. A requirement for this special type of plunger is also that the target material is restricted to a central spot on the suspended membran and of the same size as the accelerator beam spot (diameter ^ 1 mm), fig.9.

Fig.9. Plunger target with a 2lt2Pu deposit prepared by heavy ion sputtering. The plutonium isotope is restricted to the central area by a mask. 56

The material must therefore be deposited only after one has stretched the flat face of the plunger. Great care must be taken to shield off the rest of the plunger tube because any actinide material on the outside of the tube could fission by secondary neutrons coming from the beamspot just a few millimeters away. These fissions would create a background and could completely mask the real effect. The plunger is prepared by stretching a piece of 1 mg/cm2 Ni-foil over the outer tube by means of a teflon disc, fig.10, ref. 5. As an extra precaution we sometimes apply a small amount of glue on the outside of the tube down at the base. Then the inner tube is carefully expanded and the resulting flat surface inspected by microscopy.

Fig. 10. Plunger target holder with 2.3 mm outer tube diam. 1.Teflon ring. 2.Inner,expansion tube. 3.guiding pin. 4.Outer tube with alignment flange. 5.Nut for expansion of inner tube.

Thereafter a mask with a 1 mm diameter central hole is fitted on to the outer dimension of the plunger body and adjusted under microscope until it is less than a tenth of a millimeter from the suspended membran. The plunger is then ready for deposition of the actinide isotope. The isotopes are deposited by argon sputtering of oxides or metals with the plunger flat surface 25-30 mm away from the sputtering centre. To obtain a 50 pg/cm2 target of plutonium in this collection geometry we have to apply 0.5 mA of beam for about 50 minutes or on integrated current of about 1500 millicoulombs. 57

During an experiment the beam might heat up the target foil so much that wrinkles may appear. It is therefore "advisable to have a spring mounted between the expansion screw and the inner cylinder of the plunger to allow for expansion. We also insert a beam collimator just behind the expansion screw to stop any stray beam at that point and take excess heat away. We now have experience in preparing plunger targets of uranium, neptunium, plutonium and americium. The amounts of isotope needed for a preparation is of the order 5 mg. This does not mean that so much material is consumed to make a target of the type described above. Five milligrams is just a convenient amount of material to work with and is sufficient for several preparations. We have for example sputtered targets of 2'*''Pu from about one milligram of isotope. We have alsc carried the plunger technique one step further by going to sloping surfaces . The reason is that one has wanted to study angular correlations of fission fragments from picosecond fission isomers. The traditional plunger gives the possibility of measuring at 90°, but we have produced plungers that have their suspended surfaces 20° to the beam axis. 3. Actinide Sandwich Targets The other type of sputtered actinide targets we have developed recently is multilayer sandwich targets. A problem in measuring angular correlations and perturbed angular correlations of fission fragments is the attenuation of the correlation by extranuclear fields. These fields originate partly from the crystal structure of the medium and partly from the creation of vacancies in the atomic shells as a result of radiations. The extranuclear fields can attenuate an angular correlation almost to isotropy and the target makers' problem has therefore been to find a medium which reduces the attenuation to a minimum. It has been observed that y-ray angular correlations are very well preserved when polonium nuclei recoil into a lead lattice, ref.6. We have therefore adopted lead as stopping material, and we let the recoiling reaction products stop in lead. Since the recoil energy is low in these experiments, one must keep the thickness of actinide down to about 20 ug/cm2. Otherwise the recoil would stop in the actinide itself and probably be subject to strong extranuclear fields. At the same time, the cross section for producing the isomers is so low that a 20 ug/cm2 layer of actinide with a lead stopping layer on top would be unacceptable. We have therefore developed a target consisting of 40 layers of actinide interspaced by 40 layers of lead. The backing is a 2 mg/cm2 copper foil and the actinide isotope and the lead are deposited by heavy ion sputtering. We have for example 233Pu in one cavity of the beam stop and lead in another cavity 5-10 mm away. The target frames with the copper backings are mounted about 20 ram away in a cylindrical geometry and are rotated around the beam stop by an electro motor. The beam of argon ions is now focused onto the actinide to a 1 mm2 spot and a layer of 20 pg/cm2 of 239Pu is very evenly deposited on the rotating backings. We then remotely move the whole beam- stop with substrates until the lead coincides with the beam axis, and we sputter 70 ug/cm2 of lead on top of the plutonium. 58

After this we go back to plutonium and the whole sequence is repeated 40 times. The sputtering is controlled by a beam integra- tor which can be preset to a certain number of millicoulombs and reproduction is very simple. The number of millicoulomb that gives a certain thickness is determined in a short test sputtering. The yield of actinide can be determined by a-counting of its natural radioactivity and the lead yield by elastic scattering of a-particles at the tandem accelerator. To sputter a 20 pg/cm2 layer of 235U from the metal takes 15 minutes, and the 70ug/cm2 lead layer 5 minutes. Effectively, one would therefore have to spend 13 hours on this type of preparation. We make 5 targets at a time in this way so they each represent almost 3 hours of work. On the other hand, the result is such a major break through in the study of fission isomers and intermediate states in fission that we feel the work is justified. I want to acknowledge the very valuable assistance of Sonja Dahl and Jette S0rensen and financial support from the Danish Scien- tific Research Council. References 1. L. Westgaard and S. Bj0rnholm, Nuclear Instruments and Methods, 4_2 (1966) 77-80. 2. G. Sletten, ibid. HT2 (1972) 465-468. 3. G. Sletten and P. Knudsen, ibid. 1£2_ (1972) 459-463. 4. J.L. Gallant, ibid. 1£2_ (1972) 477-483. 5. Annual Report, Nuclear Structure Research Laboratory, University of Rochester, 1973. 6. C.V.K. Baba et al., Phys.Lett., 43B (1973) 483. 59

I'P.EFARATICII OF THIN ACTII.'IDE METAL LISKS VSI:;G A MULTIPLE DISK CASTING TECHNIQUE V/illiani V. Conner CLEL.ISTEY RESEARCH ALL LEVELOFLELT Frccecs Chemistry LOW CHEK1CA1. U.S.A. Rocky Flats bivision Golden, Colorado

ACKLOWLELGEMELT

T:.e infc2'::.atioi. contained in this article was developed during the course of work under contract AT( 29-1) -HOC v;ith the u. S. Ate!.-.:"•.c Energy Commission.

AirSTLACT

A casting technique has teen developed for preparing multiple actiniae n.etal disks which have a miniinum thickness of 0.006 inch.. This technique was based en an injection casting procedure which utilizes the weight of a. tantalur. rnetai rod to force the molten metal into the mold cavity. Using the proper mold design and casting parameters, it has been possible to prepare ten 1/2 Inch diameter neptunium or plutonium metal cLl.ks in a single casting. This casting technique is capable of producing disks which are very uniform. The average thickness cf the disks from a typical casting will vary no more than C.001 inch and the variation in the thickness of the individual disks will range from 0.0001 to 0.0005 inch.

ILThCLUCTICn

The Rocky Flats Plant has been involved for several years in the preparation of accinide metal targets and target materials (12) for research and diagnostic purposes. ' The target materials are usually highly purified both isotopically and chemically. Since these materials are expensive to prepare and are available only in limited quantities, the techniques used to prepare targets from these materials should generate as little scrap as possible. Precision casting techniques Lave been developed at Rocky Flats for preparing small actinide 60 metal targeis. These casting techniques we:°e developed to eliminate the large amounts of scrap generated Ly rolling; ;ro- (ii > cedures formeriy used for preparing targets. '

C't.e of the techniques developed was an injection casting process v;hici. utilizes the weight of a tantalum metal rod to force the molten metal into the mold cavity. A recent development program has resulted in a new mold design which makes it possible to cast multiple actiniae metal disks which have a minimum thickness of 0.006 inch. Using the new mold design and the proper casting parameters, it has been possible to prepare up to ten 1/2 inch diameter disks in a single casting. This paper describes the meld design which was developed for preparing these thin metal disks and the procedures and casting parameters which are necessary for successful castings. iiXFIiRIKEIITAL

Equipment The furnace used for the multiple disk castings was a System VII general purpose metallurgical facility manufactured by Vacuum Industries (Figure 1). The furnace consisted cf a tilt-pouring 'i inch diameter by 6 inch high water cooled induction coil enclosed in a water cooled vacuum chamber. The coil contained a quartz insulator sleeve and a 1/H inch thick sleeve of WDF graphite felt insulation manufactured by Union Carbide. The power for the coil was supplied by an inductotherm 15 kw motor- generator unit. Temperature measurements were obtained using a chrcmel-alumel thermocouple. The vacuum system consisted of a !-i inch diffusion pump with a mechanical roughing pump.

Several different mold designs were evaluated for the multiple disks castings. Most of these mold designs proved to oe only partially successful, that is. only a few of the disk cavities v.culd fill. Other designs were successful insofar as filling the disk cavities was concerned, but the dimensional tolerances produced were very poor. The mold design which finally proved to be successful is shown in Figure 2. This design was for a split mold capable of producing ten 1/2 inch diameter disks per casting. The mold was designed with a 0.020 inch thick by 1/8 inch wide central ;:.crue which extended from the metal reservoir in the bottom 61

O CO

-P c

(Ll EH 3 62

THERMOCOUPLE CYLINDRICAL WELL MELTING SECTION

METAL RESERVOIR

VENTS

DISK CAVITIES

CENTRAL SPRUE

GATES

Figure 2 .

Cross Section Showing the Design of trie Multiple Disk Kold 63

of ij.e ::.vl:i|-.;; stc-t,io!i to U.e Lotto::, of ti.e .\:c2ci. The dj.;k ^av/'ties v:u2;c- joined to ti.e central sprue with tapered ;;u';t::. -!.« me Ids ;;t'i s fabricated from ^lack pole graph! It/ , and the '. >.ner surface:: of ihe melds were coated with a C&F^ ::.cld ccatinr ! ;-t'j arat i 01. before being used. Ti.e tantalum ,:.t-tsl red useci to J'crcL- the molten metal into the i:iOld cavity was O.cor^ inch I:. ^Jaj.cic:1 and 2-1/Z .inch long. The weirl.t of the rod v;as i;10 rrar.o. '.:-_• L r,d uf the rou v;l..i ch contacted ti:^ :;,olten ::.f:tal wa;; coated v.'Jt ••.; ..'u::- t.i.iu ccr.iti;,|" LeToi-e each cactiii;".

: I'OCtJlii't V:.e r.etal to le cajt v;a:; placed in the :::eltinp section of ti.e ::.oiu an-j i i.e.- ::.ciu v;a^ placed in tiie inductior. coil. Ti.e coil v;aj "-llted at a 'l'..° uiirle, and the tantalum metal rod wac piaceu in the upper portion of the melting cection of the ri.cld. ".he vacuum c:,ai:.i.er v:at' then sealed and evacuated. The mold waa heated over a IC to li- minute period to tetween 800 and 8^0°C. The induction joil and ;:.old v;ei'e then moved to a vertical position, allowing the tantalum metal rod to 'drop onto the molten metal which forced the metal into the aio'i-: cavitiec. The mold was allowed to cool under vacuum.

The multiple dick ca;;tir.r technique has been uoed to pr-epare aipi:a p.iutonj urn, delta plutoniuri, and neptunium metal dick?. Ti.e product from a typical caotir.L iu si:ov;n in Figure 3- Thio figure ahowc ten delta plutoniui:. ir.etal dir.kt; as they appeared after they were i-c-movfcu fro::, the ir.old, but before they sere separated from the central s;rue, gates, and vents. These disks were 1/2 inch in uli-moter arc 0.C11 inch thick. The disks were separated from the rest of t:.e casting using a precision ground tool steel punch.

\':,b dimensional accuracy obtainable with the technique is illustrated by the data given in Table I. This table gives the weights ana tiiicknesses of ten 1/2 inch diameter alpha plutonium metal disks. The charge for this casting- contained 19-7 grams of alpha Plutonium and the casting produced ten well formed metal disks. Ti.e average thickness of the individual disks ranged from 0.0060 tc O.CCC'3 inch. The spread in the thickness of the individual uisks ranged from 0.0001 to 0.C005 inch. 64

Figure 3-

/:!.[ eared After i-.en.cva] Irrc::. ;. j.t- Kol; 65

S';;.ilar result:; v.ere obtained frcr. ca^tir.r:; u:.;;r.t~" :\eptur.iur. rr.etal . The ciii:.e:.ric;i;j cf the di;.;k:; obtained f'rcr. cr.e of the r.eptur.iu::. ca:.:t .1 r./;'i= are a!.cv;r. in Table II. T:.ic cactin;-; j r-cauceo ;er. 1/2 Ir.c:

dian.etei- ;:e:jtur;i u::; metal a j :i'••:::, the averare thickne:;:; cf v:hlch ;-a:.red froi:. 0.0102 to C.G112 inch. The sj.reaa in ^he t:.io/.ne-j.z if ti,e.:e a:, .-k:: rari;"e(i fron. C.CCGl tc C..OOCr,. 1:.c:..

TALL-.-: : A:.;L Ln^I^lOlii; C.I-' 1/2 IIX1!! I•iAKJ-.Ti-.;-. ALPHA PLUTCIJIUF: I-iHTAL L'lCK;.- Dick Thici-ineL: 5 ^j-i-ead i: u t • (c) I-hin .(inch) Kax . (inch. Thicknecb(1: I .3127 0 .0060 r..0062

'-• .3120 0 .0060 u .0062 0 • 3009 0 .0059 0 . C0£2 C\ n •' 0 6£ .0057 .0062 r 0 • 3033 .0060 0 .0062 0.0002 r • 321? 0 .0061 0 . 0065 -3129 .0062 0 .0063 r. .3131 0 .0061 .0064 0.0003 .2970 0 .0060 0 .0061 0.0001 .3014 0 . 0 0 61 0 .0061 0.0002

TABLE II AND D:MEI:EIC;-:S OF 1/2 IKCH !;EPTUI;IUI>- KETAL DISKS

U-L Lit u x ^ JJ. x Ilo . Kt.(£) I-Iin . ( inch ) Kax. (inch) Th i c kn ess(ineh) 0.6429 0.0109 0.0112 C.C003 2 0.6129 0.0101 0.0104 0 .0003 3 0.5975 0.0101 0.0103 0.0002 4 0.6209 0.0103 0.0104 0.0 001 0.6535 0.0110 0.0113 C.00 03 £ 0.6573 0.0110 0.0112 0.0002 0.6593 0.0110 0.0114 0.0004 t' 0 .6 j 91 0.0107 c .0109 0.0002 0.5919 0.0101 0.0103 0.0002 - r c f. £ 0.0107 0.0112 0.0005 66

:.'.. :....:' :;.le uisk casting technique. Ti.ese ir.elude ti.e desirr. f i:.': ,• ":' a; i.; *; c i.'.o 1 a , ti.e s 2 s e 01 t J.C c a s t.. n g 0 nar; "c ^ an o t ne

y . '. •.. .. '.'.' -;• c t • •. t o a be f Ci'e a t a c c e s s f u j d e s . /' r v.ra s u e Y o 1 o! t_- c; .

.-e central sprue and ti.e disk cavities. Other me Id uesigns v;ere •..•.t~d ;•:'•.;. the gates entering the disk cavities at various anrie ;.ese mo.: us produced results which varies fro::, completely unsuc- t. s:f,>i to partially successful. The final meld design (see igure 1) utilized gates which entered ti.e totter.; four ui;:k uvities at ;i;° above the horizontal and gates for the toy six is/ cavities which entered at 3OC below ti.e horizontal. i.e vent ing of the disk cavities was critical to the success of .. s tC'cnniisus . oeveral unsuccosst ui vonimg systems .'.ere test ea

- • > - ' *" •. . :e .: t, ;; L '- .- ^ :. O ,% : I ^_ n ^ ^. ^ ni ^ t_ »v^ ^j u^: v Cl^'^cu . 1,' S U n j i. „ o /stem the uc-ptr of tite vents between the disk cavities were . tCi inch le.'.s ti.a.n the depth cf the disk cavities. however, .e depth of ti.e vents from the top two disk cavities were tapierec - O.^'-t inch and ti.is depth v;as i:eld to the point where the vent; _-t tne top surface of ti:e mold. This venting arrangement jiieveu the internal pressure in the moid and allowed the molten :;ta.: to flow through the aisk cavities and up the vents. This .•:.'.'.:.; system solved many of ti.e problems encountered wit],

;••-•"•.' ious meld designs. ince this tL-cnnique was developed for casting actinide isotopes . 1c:. are highly purified both isotcpicaily and chemically, it was :.; ortant to i:,inimize the ai:.ount of material required for the -.;tl;,g charge. A large portion cf the charge is required to ill the metal reservoir at the bottom of the melting section :' ti.e mold. Several attempts were made to decrease the volume ' tne metal reservoir, but the molten metal did not flow •C! erly and ti.ese castings were ur.cuccecr-ful . However, the •ta'i whic:. forms '. n ti.e metal reservoir is not lost. This 67

metal can be recast into feea ingcts and used fcr subsequent castings. A charge containing 20 to 23 grams of me.tal was found to be CT.tir.un. for' these castings.

The casting temperature was also found to be very important. If tht- temperature was L'0°C too iov:, the disk cavities would not fill completely. i\ temperature which was 20°C too high would result Ir. excessive quantities of metal being forced out of ti.e v^ntc. Ti.e optimum temperature varied somewhat depending upon ti.e element being cast, but for plutonium castings, the optimum ": omperature was round to lie between 820 and &it0°C.

une practical application of ti.is casting technique involved ti.e i ref. araticn of neptunium metal disks foi- use in ti.e liquid metal ."'ast breeder reactor neutron dosimeter program. These disks were prepared in cooperation with the Tar-get Preparation Center at Oak hiuge national Laboratory (ORIJL). The most recent order was foi' thirty i/1' inch diameter by C.006 inch thick neptunium metal disks Ike starting material foi' this project was from a Latch of high i urity ".'pX'L1 wi.ich liad been puri"ied previously1" at hocky Flats i'oi- the Isotopes Pool at ORIIL. The 95 gram Latch of :ip0o was converted to metal and the metal button was cast into a feed i ; . g O t .

Tre first two castings were made at c2C to c30°C and produced five and four disk;:, respectively. The temperature was increased to the c1f0-830c'C range for the subsequent casting's which all produced ten disks. The average thickness of the best thirty disks varied fro;:: 0.00^3 to 0.006? inch, with the spread in ti.e thickness of ti.e individual disks ranging from 0.0001 to 0.0005 inch.

Cne concern with the multiple disk casting technique was the potential contamination of the metal with CaP0 mold coating or graphite scraped from the walls of the mold by the tantalum metal rod. Another possible contamination source was the tantalum rod itself. A sample of the central sprue from the last neptunium disk casting was analysed for impurities by spark source mass spectroscopy. The results of this analysis are shown in Table III along with the impurity analysis from the NpOg feed used for this project. The C, Ca, and Ta content of the metal was lower than in the oxide while the increase in the F content of 68

TAELE III SPARK SOURCE MAES SPECTRCGRAFk]'J IMPURITY ANALYSIS Concentration. Ccr.centrat ior len:er.t ' in IipO2 Feed Jr. .;.p Metal

Al 1C c 3 o o Ca 3t Cl 130 C r 1

In 13 < C.2 K 12 < COG Mr 5 < C.CC

55 < o. 3^ < 1 7 0 N.A. - 260 3 9 < O.CC 2 5 < 0.3 25 21 15 ^ < 1 1 H < 1 1 0.4 3 3 TOTAL €^G 6 92

t listed were not detected. 69

the r::etal v;as very small. The Fe, ;.:i. and Cr center.t cf ti.e me; v;as higher than in the oxide. Ti.it; increase v;as attributes to pickup from the inconel anu stainless steel equip::.ent uced ir. the conversici. cf the oxide tc metal . The oxide cor.taii.eu a total cf C46 pp n; of detectable impurlti es. '-'he netai sample cor 1 faired L^'2 i]!. . total of detectable iri.puri t ler., :,';t if c.xy("en Iz excluded, ti.e- total ::;.p:a-:ty level j :: i-educt-U to h"^2 v\v...

The :;.ultij. le dioi; ca^tiiI(" technique pi-cvide.: a ir.ethoc: fo2- pj'e- p.ai-ni£' tijiii actir;ide ::.eta] di^ki; v;iti,out t\,t prcLler:..; at;t;ociate; v;itli a 2'cilir;.;- and punching- proceed. TJ.IL; if, especially ii.-if oi'tant in ti.e caue cf neptuniui:. metal v;hich ic difficult tc roll Lecauce of its very brittle nature. This technique ij capable of producing dicks v:itii ^ccd uir.enci&nal tclerar.ee:; and without addinr objecticnal impurities tc ti:e r.etal .

1. V.'. V. Conner and S. C. Procter, hFf-lJllC, The Lev; Chemical Co., Lecember 23, 19^9- k. V,. V. Conner and D. L. Laaoo, RFr-lc^ic, Ti.e Dov: Chemical Co. , May 17, 1972 j. V.7. V. Conner, Nuclear lr.strur.er.es and Methods 1C2 (1972) pp. kn-k23. M. V;. V. Conner- and ~j. L. Baaso, HFF-1^![.-., ':":. e Dov; Ci.en.leal Co., Cecemter 30, 1969- 70

EVAPORATION OF TARGET MATERIALS OF LOW VAPOR PRESSURE ON VT.KY THLN CARBON BACKLNGS BY ELECTRON-BOMBARDMENT

P. Maier-Komor

Physics Department, Techn. Universitat Mtinchen, Germany

INTRODUCTION

Thin ami uniform targets are needed for high-resolution spectroscopy, i.'xcpeciaj ly with heavy ions. Typical target thicknesses are of the order 5-50 ,.g/i:m" which precludes the use of self-supporting foils. Therefore thin backings have to be used, mainly carbon backings. The preparation techniques are unprobleiiiat: ic if the isotopes can be evaporated at moderate temperatures from a boat or heated crucible. Damage to the carbon backing can be avoided by a thermic shield and by limiting the evaporation rate. Metals with very low vapor pressures can only be evaporated by electron bombardment from a cooled crucible. Because of the large primary beam energy special precautions have to be followed in r-rder to prevent damage to the backing material. We have studied the influence of energy dissipation in the vapor source on the preparation procedure, especially heat conduction, convection and radiat i on.

EXPERIMENTAL

Using isotopes the amount of evaporation is very small. Therefore, when using commercially available electron guns only a small fraction of the emitted electrons actually impinge on the evaporant. Typically the cross section of the beam is of the order 3 x 7 mm and consequently the effective intensity is too small to evaporate metals of low vapor pressure. We are working with a Veeco type VeB6 electron source which utilizes the Pierce principle. The electrostatic deflection angle is 40° (Fig. 1). In 7 to 13 cm distance from the source the diamter of the beam is about 3 mm. The maximum acceleration voltage is 20 kV and the maximum power is 6 kw. Since the manufacturer doesn't offer a crucible capable of withstanding the high power density produced by this source (85 kw/rm ) we have developed a suitable water cooled crucible. The crucible is simply a flange with a spherical cavity (Fig. 1). It is bolted to the cooling water feed-through with an 0-ring seal. Is is therefore very easy to replace, making feasible the use of separate crucibles for different isotopes. Use of a liner was ruled out because of the difficulties involved in controlling the contamination of the evaporant. 71

Fig 1 Constructional Details of the Electron Gun 1 Substrate Mount 2 Copper Crucible 3 Evaporant I* Electron Source 5 Deflector 6 Water Cooling

0 B 20 30 (0 50mm

In order to reduce heat loss due to conduction in the crucible itself we also tried crucibles made of tungsten and molybdenum. Both were turned on a lathe and iiad the same geometry as the cupper crucible. The tungsten crucible proved to be unsatisfactory since the high temperature gradients involved tend to crack this rather brittle material. The formation of these hairline cracks is signaled by rapid deterioration of the vacuum. Molybdenum on the other hand, is very resistant to this effect. However, although the heat conduction of molybdenum is only about the third of that of copper, the measured evaporation races were but slightly higher than those obtained with the copper crucible. Heating of the target is mainly due to two contributions: 1) heat radiated from the evaporant 2) condensation energy, i.e. heat liberated as the target builds up. Heat absorbed due to radiation from the evaporant remains fairly constant during the evaporation process. In the beginning the carbon backing is quite transparent and absorbs little heat. As the target material builds up the heat 72

is reflected from the shiny surface. With ruthenium targets, for example the temperature rise is about 40°K x

(S = > • i-(T4 - T ); T = 2900°K i. = 0.25, absorption coefficient •; 17) o

neglecting the heat radiated by the target itself. Here the distance between the crucible and the target is about 12 cm. On the other hand the temperature increase due to the condensation energy approaches the same value (40° s ) for a condensation rate of 0.05 iig/cm" s. Evidently, higher condensation rates should be avoided. However for some metals (e.g. niobium) condensation rates of this order are only obtained at temperatures considerably above the melting point. With a water cooled crucible this requires a drastic increase in input power. This becomes evident if one considers the heat losses after exceeding the melting point. Below the melting point the effective contact area between the evaporant and the crucible is very small so that the thermal resistance of the interface evaporant crucible is quite high. Above the melting point the evaporant comes into more intimate contact with the crucible. The effective contact area increases with rising temperature due to the decrease in surface tension of the molten evaporant. Consequently the heat transfer to the crucible increases over-proportionately with rising temperature. Conversely a decrease in temperature as induced for example by interrupting the electron beam, leads to a considerable increase in heat resistance to the crucible.

Above the melting point the conduction losses also increase due to convection within the molten droplet. The corresponding energy flux rapidly rises from zero at the melting point to the order of magnitude of the pure conduction loss and then seems to increase as a linear function of temperature. In order to obtain adequate evaporation rates with a minimum of input power we have developed a control system which permits use of a pulsed electron beam. The filament current is kppt constant and the beam is pulsed by switching the high-voltage. The energy impacted to the evaporant is controlled by varying the duty cycle of the pulses. The beam is interrupted as soon as the evaporant :s completely molten. The intensity of the electron pulse is determined by measuring the evaporation rate with a quartz crystal thickness monitor. The upper limit is calculated in advance considering both the condensation energy (2) and the heat radiated by the evaporant . As a rule of thumb, both of the contributions should not result in a temperature rise in the target exceeding 80°K s . The beam must be switched off long enough so that the molten evaporant can solidify and that both the evaporant and the target can cool off. With these techniques we have prepared plane, homogeneous targets of Mo, Ta, Nb, W, Re, Ru, Os, Zr, Hf, etc. on carbon backings of 3-5 i'<*/cm . Up to now thicknesses ranged up to 100 lig/ra'', however it is possible to achieve greater thicknesses. 73

REFERENCES

1) R.E. Honig and D.A. Kramer, FXA Review Vol. 30, No. 2 (1969). 2) L.R. Roller, Scientific Foundations of Vacuum Technique, New York, London, John Wiley 4 Sons Inc. (1962). 3) P. Maier-Komor, Nucl. Instr. and Meth. 102 (1972) 485. 74

SELF-SUPPORTING ISOTOPIC CHROMIUM NUCLEAR TARGETS

Charles-Albert Bouchard Laboratoire de r.iysique nucleaire Ur.iversite Laval, Quebec - P.Q.

1. Introduction 50, We needed a few months ago ' Cr targets to do an experiment of reaction cross section for the bombardment with a-particies. We required self-supporting targets in order to reduce the backing contribution to the yield of y-rays. We tried to produce targets of chromium by electro-plating of chromium on a copper foil (.0002") and by dissolving the backing by floating on a trichlo- roacetic acid-ammonia-water solution (100 g : 500 mil : 500 m£). But this method required 2 grammes of CrO for 15 m£ of aqueous solution. Therefore such a procedure was judged too expensive for isotopic preparation. We thought then to use the vacuum evaporation method. Chromium is very brittle; its ductility star+c- only at 200 C, which makes self-supporting targets difficult to prepare. Targets of enriched (95%) 50rr have been evaporated in our laboratory. This paper describes the procedure used.

2. Electron gun source and apparatus We use the e-gun source and apparatus made by Therrnionics Laboratory Inc. (Model 150 - 0030). Figure 2.1 illustrates the schematic of the evaporator crucible and glass slide holder.

Figure 2.1 Evaporation geometry

1 - large crucible; 2 - water cooling conducts; 3 - electron emission filament; k - permanent magnetic field; 5 - tantalum sheet; 6 - small tantalum crucible; 7 - holder; 8 - teflon insulator; 9 - glass slide; 10 - copper block; 11 - heater. 75

iltctror.s are emitted from the filament at the front of the crucible. T:v.".-y are bent by a permanent magnetic field and foeussed in the center of a "r'.ailio block used as a large crucible. This block i; continuously cooled by -•old water. Filament is negative '-U000 V) with respect to the crucible; max;mur. emission current is 7f;0 mA giving 3 kwatts of output power. 0:; the large crucible we put a tantalum sheet in order to avoid evaporation of iir.rurities. A crucible of tantalum was placed on the tantalum sheet.

The crucible is a small cylinder cf tantalum 1/8" in diameter by 5/It" long. AT. the top ve have drilled a hole l/l6" in diameter by 3/32" in depth. '••!c chargv the crucible with 10 mg of CroG_ the first time and with 5 i.ig the ot:v-r times, "here it- a reduction of the oxicie in the crucible. The same crucible cannot to used more than 3 times before it loses its ability to act as a reduci:.,™ efit'iiyst.

3. Substrate preparation and target evaporation A small quantity of IlaCJi (~ 1/1* cm^) was evaporated in a separate evaporator. The "aC& was first melted in air. Five glass slides were placed 2 inches at ove a tungsten boat. After evaporation of the IlaC?-, we could see the interference colors by reflection of light. One of those glass slides with a light coating of NaC£ substratum was placed '.' inches above the crucible. The pressure in the chamber was lower than 10"1' Torr. The current of the electron gun was adjusted to around 10 mA when the targets were evaporated. During evaporation the back-pressure of the di ffusior: pump goes from 10 y of Hg to 60 y of Hg. The evaporation time is between ;-0 seconds and ho seconds. The glass slide with its substratum rests on a teflon insulator and is maintained at 300°C by means of a heated copper block placed on top of the glass slide. The temperature is maintained via a resistance heater. The heater was made by machining a series of parallel grooves in a piece of LAVA stone (grade M;, which was then fired at 1900 F for 1 hour. A 2 foot length of resistance wire is placed in the grooves and the current is controlled via an autc-transformer. The temperature is monitored by a thermocouple device attached tc the copper I lock. After evaporation the target is slowly cooled down to 100 C; it is cut to required dimensions and floated off on varm water (60 C).

"• ?a_";-;jt condition and conclusion

Gix targets (lA" in diameter) can be made from one glass slide. Them thicknesses of the targets we have prepared ranged from 115 yg/cm^ to 16; yg/cm^. Tj.oy were measured by the change of range in air of 5-5 Mev a-particles from

The Rutherford diffusion of 3 MeV alp/a particles shows that the targets contain oxygen (ratio oxygen to chrom. am ~ uQ atomic percent) . The .-itri.Tii^th of th.e targets is remarkable: we have us^d a 300 nA t ;an; of 3-00 MeV alpha particles on the same target for one day without visible damage, except that there is carbon deposition.

The percentage of successin making the targets is around 39^- ^he cost of ••aoh target is around SlO, mainly due to the expense of the J Cr isotope from Oak Ridre.

Trial.ks are d'.i° to Dr Maynard High for his enrcuragement during the jr^par.Lion of the targets. 76

Cryo-puniped gas target for the study of radiative capture reactions

K.V.'.Allen, S.F.Dolan, A.R.Holmes, T . J .M. Symons , F.Watt and C .."•!. Zimmerman N'uclear Physic Laboratory, Oxford, England

and

A. V. .Lither land and A.Sandorfi University of Toronto, Ontario, Canada

Th-.-: study of resonances in radiative capture reactic::s is a powerful technique for investigating, with very high resolution, the properties of virtual states of light nuclei. However, because the elecfromagneric interaction is much weaker than the strong interaction, the yield of resonant -,-ravs is much less, normally by two or more orders of magnitude, than iv.c yield <>f y-rays accompanying inelastic scattering and other charged panicle induced nuclear -reactions. When solid targets are employed, usually as thin films evaporated on to gold backings, background and contaminant ,'-r;r yields are troublesome even with very clean beam line conditions. Carbon contamination, which gives rise to 6.13 v-rays from the reaction1-C(a,ii)1 ^0 ]/ for r::i > 5.1 Mev and 4.43 Mev v-rays from inelastic scattering in C are particularly difficult to eliminate.

These contaminants are largely eliminated by the use of transmission type gas targets (see, for example, Litherland et al. (1967)) which have tl? following advantages:

(i) background is greatly reduced because the beam is tra emitted through the target and stopped many metres from the experimental location where the v-rays are detected

(ii) the thickness of the target can be easily controlled by varying the gas pressure

i.iii) pure elements, e.g. N,", 0.;, Xe, etc. can be used, thu:. reducing y-ray background still further and eliminating parasitic energy loss wlv'ch in turn reduce the effective thick target yield.

(iv) gaseous separated isotopes such as 'G0, lsN, -:Xc, etc. can be recovered and re-used.

Tile main problem with di f f eren. ial ly pumped gas targets is tha" high pumping speeds are vequired in the neighbourhood of the target apertures for a high gas throughput. Fcr this reason, Hoots blowers have often been used for the first stage of a differential pumping system, although they are expensive and noisy, and can give rise to contamination of the vacuum system. We have used cryo-pumps, which we find to be clean, quiet and convenient to ise.

We have carried out experiments with both single ended and transmission ty;M' targetsargets. In the single endend' :1 target ,(Fig. I) \.-c used a cryopumv* (type \TK 5000) manufacturef.uredd by ^eybo^eyboll dd-Heraeus •*

Leybold-lleraeus Cmbl!, KG, KHln. We are indebted to Mr. A. Simon for his advice in connection with the use of the Vl'K 5000 and other heybold equipment iiel inn is led to the inner cooling plate f5) via the manifold (6) and proceeds from there via the hiii]ar wound Lube coil (3) ard the radiation shield (7) to the exhaust gas outlet (1) which is connected to a mechanical flow control valve. The coil (3) serves not only as a condensing surface for the gas being pumped, but also as a partial radiation shield for the inner cold plate. Moth the inner and outer cold surfaces are provided with sensing elements for vapour pressure thermometers (2) and (4) which indicate the temperatures of tthh e shieledd andd colcldd plate respectivelrespctyy and influence the helium flow by feed- back to the pressure controlled flow valve in the helium line. The maximumm pumping speed for nitrogen with the cold plate at 3°K and the coil at about •' i°K is given by I lie manufacturers as 5000 .''./sec. In our experimental arrangement, which is probably condtic t anee limited, we have measured speeds of about 1000 ../sec at rather higher temperatures. Pumping speeds of about 800 ../sec for neon have also been observed. Pressures up to about 1 torr have been used in the target, although it has been more usual to operate ;)t 100 to 300:.. No significant diminution in pumping speed has been observed even after tin.- condensation of 40 atmospheric litres of nitrogen.

In the transmission target 0'ig. 3), we have added an annular cryo-pump d in association with Oxford Instruments^"'

ii [-.

!! • \A FARAD4Y CUP

1 fjn;

i I T.I]

GAS T*V?GE:T

Transmission gas target

Transmission target-

Here tlu' cold surfaces surround, and are attached to a tube 12.5mm in diameter through which I'he ion beam passes. The pumping speed of this cryo-pump lias not yet been measured, but from the surface area it is expected tu be 2r:t)C) ./sec for neon at '- 4°K. The rise in pressure in the region of the I'araday cup is . '1 x 10~r> torr for a pressure of '•- 0.3 torr in the target, which is consistent with a high pumping speed. Both cryo-pumps used "« 0.5

(.'-) 'I he Oxford Instrument Company, Osney Mead, Oxford, England. We are indebted to ,')r. Ian Herbert for his co-operation and for the design • • i this punp. __)•+.

LOCATION f=O« SECOND CRYOPUMP

DETECTOR

GAS TARGET INSULATED COLLIMATOR

GAS THERMOMETER

DIFFUSION PUMP

!'ig. ! T'K- singiu ended gas target for angular distribution measurements using a Leybold cryo-purap

Tiie operation of the pump, which is based on tlie flipping (1962) cntinuous flow pr::iciple, can be understood in greater detail from Fig. 2

ig. 1. The Levbold cryogenic -ondenser VPK 5000 (dimensions indicated are in nir.i.) ].'; litres ieu;.: helium per i.^ur, -iepenc i ng r n npt-r.-u i ng temperatures.

ie ;>.as recovery system used with t!;e single ended target is shown -n ii-a 1 ! v i :i rig. •'(. The \'!'K yDfif) rry-punu ;s allowed Co warn up

1s7 DIFFERENTIAL PUMPMG RtSIOH

MERCURY DIFFUSION PUMPS IB 2ND DIFFERENTIAL PUMPIM6 SE6I0H 7AB6E1 CHAMBER

« IIEEOLE VALVE

LIQUID NITR06EH SORFTIOH PUMPS

i'it;. -i (-as recycling and recovery system slowly am: the evolved gases, as well as any small quantities of gas iccunu1 iU ed in the backing volumes behind the mercury diffusion pumps, which ;rvo-pump, are condensed by a helium cooled cold finger. When ,i i 1 tile gas has been collected, the appropriate valves are closed and the ;'old finger is allowed to warm up. The gas then expands into the 20 litre storage volume and associated transfer lines, ready for recycling through the target. When the experiment is concluded, a multi-stage diaphragm pump is used to empty the transfer lines into the storage volume.

By careful control of temperature during the warm-up phase, gas purification can be achieved by fractionation. For example, with neon, by warmir.g the cryo-surface to 27"K, when the vapour pressure is '- 740 torr, essentially pure neon can be transferred while impurities such as nitrogen ami oxygen remain behind on the pump. The Oxford Instruments pump is provided with very convenient instrumentation which enables the temperature t ..i be coptirolied very accurately while the gas transfer is taking place.

Some experimental results which illustrate the performance of the transmission type gas target will now be described. For these studies, we used tiie singly charged helium beam from the Oxford vertical Van de Graaff. This is a high quality beam (semi-angle of divergence after magnetic analysis :s 10 'radian at " 5 Mev) which was easily focussed through insulated eol 1 ima tors and the target apertures, which were usually '. 3mm in diameter) In :•" i g. 51 '.) the yield curve in the region of the 6 , E- = 5057 Kev resonance 80

YIELD YIELD CURVE OF 6 8.78 McV STATE IN 20Ne

60O0 Ea = 5057 KeV

8.78 6' 5000

4.25 — 4000

1.63 3000 -

i 1 2000 -

1000

14980 14990 15000 15010 15020 KHz

Gamma ray yield curve (5 Mev '• F,, • 4 "lev) for the. ']"0(,-,)- •• \'e reaction in the neighbourhood of the 6 + resonance.

'• n ''" ^ + • is shown, while Fig. 5(b) sliows the energy spectrum of the decay • rays measured with a 4" x 4" Nal detector. The energy spectrum is a (>>,as in - gas out) difference curve, but the. background was, in fact, very sna ] ' • The pressure in the target was 150.. of natural oxygen.

• e yield curve shows that the target was clearly "thick", and the •]se shovs that the energy resolution was '. 1 Kev. The yield curve was obtained by the beam modulation technique (Alexander et a!. 19 72, Wormald and Takacs, 1974) and has insufficient channels per unit NMR frequency interval to demonstrate the resolution clearly. In Fig. 6, we sK>w the re.so! ot ion of a close doublet in '"F at 5603 Kev corresponding (.o Y, , - 1326 Kev in tiu- reaction '''N + a. The existence of this doublet was first recognised by K.P.Jackson et al. (1974) from accurate measurements of the -. ray decay energies using a Ge detector. The. vitld curve • obtained with the gas target shows that the doublet separation is I . f- Kev, in excellent agreement with the Toronto measurements. 81

DECAY OF e\ 8.78 McV STATE IN 20Ne

2000 3000 4000 5000

Ev (KcV)

Fig. 5(b) Gamma rav energy spectrum from the 6+ resonance in 82

5600

DOUBLET IN 18F AT 5603 KeV

Ea= 1526 KeV N' * a GAS TARGET p = 0.05 TORR

AEa = 2.1 KeV x < AEa = H 2.1 KeV = 1.6 KeV

3300 390 440 490 CHANNELS

Fig- 6 Yield curve of the reaction ' ''N (a ,y)l ? 'F in the region of the close doublet in 18F at 1526 Kev,

In conclusion, we wish to thank the staff of the Oxford workshops, under S. Tolan, who fabricated the target equipment described here, and the Van de Graaff operations team, under the direction of H.R.McK.Hyder, for their excellent co-operation. Financial assistance from the Science Research Council is also gratefully acknowledged.

RetereiK-.es A.E.Litherland, R.W.01lerhead, V.J.M.Smulders, T.K.Alexander, C.Broude, A.J.Ferguson and J.A.Kuehncr (1967), Can. J. Thys. 45, 1901.

0. Klipping and W. Maschor (1962), Vakuum-Technik, _M_, 81.

T.K.Alexander, B.Y.Undervood, N.Anyas-Weiss, X.A.Jelley, J.Szucs, S.P.Dolan, M.R.WormaUi and K.W.Allen, 1972, XucJ. Phys. AI97, I.

M.R.Wormald and J.Takacs, 1974, Sue 1 . Inst . & Meths., 113, 26.3.

K.P.Jackson et al. (1974). To be. published. 83

Thin Film Analysis Using MeV Energy Ion Beams*

I. V. Mitcheli

ABSTRACT

With a low energy accelerator one has available three complementary methods for analyzing thr> composition of thin films, these being the methods of nuclear reactions, ion scattering and ion- induced X-rays. These techniques are now well-grounded and permit fast, non-destructive, quantitative and depth-sensitive assays.

The basis of each method is briefly outlined and illustrative examples follow, including:

(a) profiling of 4He implanted into metal foils (to be used as targets in DSAM lifetime measurements) by the method of helium ion backscattering;

(b) determination of N and O incorporated into sputter- deposited Mo films, by the methods of ion backscattering and nuclear reactions; and

iii.niusr r i p L not r^cc i vt'd . 84

M A Sensitive ethod for Determining I'ranium and Ti1Orium Contamination in Nuclear Targets A.M. Chung, F.T. Diamond and A.F. Litherland 1'niversity of Toronto, Toronto, Ontario, Panada

'luring the past three vears we have been observing the fission of light elements such as '"Mg by bombardment with high energy electrons. The very low cross section of this reaction, of the order of tens of picobarns in the region F.e = .10 -/n Me\'' , and the high background of gamma rays, scattered electrons and other lightly ionizing particles from the other reactions make the use of conventional semi-conductor counters impractical. However, the Polycarbonate Damaged Track (PDTi detectors are verv useful in these measurements. The development of these l'DT detectors will be published elsewhere'. Thin, about 2-10 \.m, PnT tietecLott are first vacuum-coated willi a layer of aluminum (' 150 A thick) on one side. These aluminum-coated PIT detectors were then mounted on frames which surround the nuclear target. The large active area of these P^T detectors makes possible the detection over approximately - sterrad- ians. Scattered electrons, gamma rays as well as reaction products are incident onto these PDT detectors. Permanently damaged tracks made by these ionizing particles form the basis for detection of the heavv ions. The degree of damage depends on the energy loss of the particular ionizing species that produce the damage, l^ien these PUT detectors are floated, witli the aluminum side facing up- wards, on an anueous etching solution of concentrated NaOH the etching speeds aiong the damaged tracks are much faster than the etching speed of the undamaged bulk material. furthermore the etching speed along the damaged track depends exponentially with the dF./dX of the ion species along the tracks, The time required to etch through a certain thickness of POT detector (which is related to the inverse of the etching speed along the track) can then be used for ion identi- fication. When a track etches through the unwelling \Ta0H etches the aluminum coating futther, thus leaving a visiblp spot on the aluminum which can be seen easily by the naked eye. The etching speed along the damaged track depends also on the "'aOH concentration, the temperature, etc. of the etching solution. n the case of our measurement, ft um thick PUT detector are usually used. The. etching was done with 7.1 Mormal *!aOH solution at 21 C. Mission fragments from the olectrofission of uranium and thorium will be etched out between 2 to ft hours, Fragments such as oxygen and carbon will not be etched out until '<.• 20 and "« 60 hours, respectively. Single Tracks due to gamma rays, electrons as well as other lightly ionizing particles such as protons, aloha, etc. will not be etched cut at all due to their small dF./dX in the PDT detector.

One of the maior difficulties that we met in the early stage of our experiment was the uranium and thorium impurities in our target. The electrofissiop 35

cross section for uranium in the energy range of Fe = 20 to 50 MeV is about 2 mb -. The electrofission cross section for light elements such as "'Mg is only about 50 pb. More than a few parts per million of uranium and thorium impurity in the target will then give serious background problems. Consequently the PDT detectors proved to be a sensitive and effective method for determining the uranium and/or thorium in thin .nuclear targets. As an example, our experience wirh ~~ 20 yA of 30-45 MeV electrons shows that a four hour irradiation yields about 100 fission fragments from uranium and/or thorium in a 200 ug/cm' thick 2''Mg target.'1 This corresponds to about 2 ppm of uranium and/or thorium with a PDT detector "-< 6 cm x 23 cm in size and ^ 15 cm away from the target.

References 1. A.H. Chung, Ph.n. Thesis, University of Toronto 1°73, unpublished. 2. W.T. Diamond, Ph.D. Thesis, University of Toronto, 1974 and to be published. 3. I.C. Nascimento, The Proc• of International Conf. on Photonuclear Peactions and Applications (Asilomar, Calif., U.S.A., 1973). 4. Rolled target obtained from the Oak Ridge National Laboratory, U.S.A. 86

SOME MASS MEASUREMENT PROBLEMS

Janet S. Merritt

Atomic Energy of Canada Limited Physics Division, Chalk River Nuclear Laboratories Chalk River, Ontario, Canada KOJ-iJO

Faced with the problem of determining the thickness of a target, one perceives that an uncomplicated approach is to measure its mass and area and take the quotient. This paper examines the mass measurement aspect of such an app roach. The first question encountered is the selection of a balance. There is a wide choice commercially available: the classical 2-pan balance, the single-pan substitution balance with a direct reading optical scale and the more recent electronic balances, some of which have 2 pans and others are substitution. In addition to examining the specifications for capacity, size of weighing chamber, readability, convenience and precision, it also is necessary to consider whether an adequate balance room can be provided, because most types of balance will not perform satisfactorily in an unsuitable environment and the degree of adverse affect varies.

Generally desirable features for precise weighing are as follows " : a sufficiently large room so that the presence of the operator will not significantly alter the room temperature or air currents, a heavy vibration free balance table th.it does not contact the walls of the room, as these readily transmit building vibrations, absence of direct sunlight particularly on or near the balance, location and selection of lighting fixtures so that they do not heat tin..1 balance or the objects to be weighed, and airconditioning that minimizes draughts, controls humid !ty, maintains the temperature constant to ;. 0.5°C and allows a small positive thermal gradient, i.e. warmer at the top, Preferably the balance rmi::] should be located in the basement or on the ground floor and well away from e 1' ".'a tors .

I'm" the classical 2-pan lialar.ee all ol these balance room qualities are essential and it is worth noting that a negative thermal gradient within the balance case has been observed to give large instabilities, e.g. changes of 0.6 ing (?) with a i.'OO-n balance when the top was 0.4 C cooler than the base " . Unfortunate- ly most air-conditioning systems blow cool air across the top of the room and create the undesirable thermal gradient; this can be corrected by directing the air I low with baffles and low speed fans. Some scientists prefer to turn off the a i i'-.oii'.l i t i on i ng system for the weighing pel od i. I it will give conditions of slowly rising room temperature with the temperature change of the balance itself lagging slightly behind C). The substitution-type balance (2 knife-edge single pan) is less sensitive to thermal gradients and draughts and thus has become the choice of many workers. Nevertheless a well designed balance room is advantageous for this type of Valance also. A sturdy balance table is a necessity, as the substitution-type balance with an optical scale is very sensitive to tilt and so must be "naintained level. Another important requirement for these balances is temperature control to reduce error caused by zero drift. Figure 1 shows this effect for a particular model of substitution balance that gives readability of 1 ug and has a capacity oi 20 g. The zero drift is not negligible and seems related to the temperature fluctuations which are indicated by the triangles; closer temperature control clearlv would be worthwhile.

Temperature ] Cooling on ! p Cooling off j ~ J24.I 5 1 o ! 0) i24.0 |- :23.9 ""

i 1 10 20 elapsed time (min)

Figure 1: The effect of temperature fluctuations upon the zero reading of a Mettler M-5 balance (reproduced from Nuclear Instruments and Methods 11J2 (1973) 329)-

The electronic balances a.'e reported to be less sensitive to environ- mi'iiL.-i 1. conditions. However, the sail effect caused by air currents acting on the pan '.s as serious for these balances as any other, so that the usual precautions Tor the control of thermal disturbances should be taken. For those balances that dn not have knife edges, however, the vibration-free requirement is not severe. Another consideration that may affect the choice of balance are the maintenance requirements. All balances should be serviced and cleaned by an 88

expert at least once a year. A knife edge can be seriously damaged by a small particle of dust or dirt lodged between it and the plane on which it rests during a weighing. Thus cleanliness is important. At the first sign of performance deterioration, the knife edges and planes should be cleaned to prevent further damage. Apart from mechanical damage caused by dirt, any accumulation of dust, dirt or films on the pans or the weights will cause error. This is a good reason Tor putting a "NO SMOKING" sign on tha door to the room. It also is a reason viiy the substitution type balance has come to enjoy such widespread use, i.e., the built-in set of weights is enclosed and protected from room dust, as is the weight manipulation mechanism; thus the weights do not have to be cleaned and recalibrated nearly as frequently as a set of weights that is hand manipulated for every weighing, such as for the classical 2-pan balance or most electronic balances.

Other reasons for the popularity of substitution balances, whi h also pertain ior the electromagnetic balances are the speed of response, direcL reading and relative independence of the sensitivity with load. Speed is valuable not only as a time-saver, but ' o because it reduces error cauised by zero drift or mass changes taking placf < . the sample being weighed, such as oxidation or evaporation. An interesting study of the basic design features and the relative advantages of the substitution-type and classical 2-pan balance was published by Peiser who strongly recommends the substitution balance for routine weight calibration work.

The U.S. National Bureau of Standards (NBS) has established various classes of weights (4) for which the materials and construction details of the weights are described, as well as their allowed limits of deviation from their nominal values. Most balance manufacturers and suppliers of sets of weights have chosen to produce weights that conform to the NBS requirements for these various classes, probably in order to market their products readily in the U.S.; but fortunately the result is a well defined set of tolerances for the weights used in most commercially available balances. As an example the NBS tolerance require- iiienLs for ciuss M are given in Table 1. Sets of weights meeting such specifications are commercially available. Calibration service according to an NBS procedure can also be purchased. In many countries tiie national standardizing laboratories have cooperated in providing weight calibrations to satisfy particular scientific, needs.

The material used for the fundamental standards of mass, the international prototype kilogram, is a platinum iridium alloy. The recomn.ended material for high quality weights in the 1 g - 1 kg range is a chrome nickel steel; smaller weight pieces are made of tantalum and in the mg range aluminum is used. Brass weights electroplated with gold or rhodium are no longer favoured because they were found less stable to changing atmospheric conditions. It is important that 89

Table I

\'BS Tolerances for Class M Weights

Denomination Individual Group tolerance (rag) tolerance (mg)

100 g 0.5 50 0.25 30 0.15 not specified 20 0.10 10 0.05

5 0.034 3 0.034 0.065 2 0.034 1 0.034

500 nig 0.0054 300 0.0054 0.0105 200 0.0054 100 0.0054

50 0.0054 30 0.0054 0.0105 20 0.0054 10 0.0054

5 0.0054 3 0.0054 0.0054 0.0105 1 0.0054 0.5 0.0054

The deviation of a weight from its nominal value shall be small enough: (1) that it falls within the individual tolerance listed and (2) that, no combination of weights within its group shall differ from the sum of the nominal values of the combination by more than the amount listed under the group tolerances. 90

weights be handled with forceps whose tips ace of a softer material than that of the weights themselves; ivory tipped forceps are often used. The forceps should be sufficiently long that the operator's hand does not heat the weights nor enter the weighing chamber during manipulations. Also important is that the weights be kept clean. As well as taking normal precautions about cleanliness, a daily wipe with lens paper is a good idea. More serious cleaning is advisable from time to time and various methods have been advocated. Among them is one in which the weight pieces are rubbed with a cleaned wet chamois cloth then dried with a cleaned dry chamois ; the chamois cloths ara cleaned by soaking in soluble detergent solution then rinsing many times with pure water to completely remove the soluble detergent. Table II shows an example of some high quality weights that were cleaned in this manner; before cleaning the weights looked bright and shiny yet this cleaning procedure removed considerable dirt from 2 of them, then gave an acceptably constant weight value after repeated cleaning. Some laboratories advise cleaning with a jet of steam, particularly for 1-piece weights . Wiping with a soft camel-hair brush is frequently done, and here again, the brushes should be cleaned in detergent, rinsed with pure water, then alcohol and allowed to dry

Table II

Results of Cleaning Weights

Before 1.000066 1.999985 1.999973

After 1st 1.000042 1.999961 1.999975 cleaning 1.000040 1.999962 1.999972

After 2nd 1.000041 1.999963 1.999976 cleaning 1.000041 1.999963 1.999973

1 day later 1.000039 1.999963 1.999975 1.000040

All materials used - weights, brushes, forceps and objects to be weighed - should be placed in the balance room for at least an hour before weighings are performed so that they will be in thermal equilibrium and therefore not cause any disturbing convection currents or buoyancy changes. It has been claimed that improved balance performance is rbtained if the operator sits in front of the balance about 15 minutes before commencing a series of weighings. 91

As a further precaution against thermal disturbances a reflecting blanket of aluminized mylar has even been worn by the operator when performing weighings using a balance with readability of 0.1 [Jg. The performance of any balance should be tested routinely to ascertain whether any problems have developed. Some excellent test procedures have been described in the literature ' that give good outlines for simple tests and are designed so that the sequence of operations is representative of actual weighings. Such procedures test the direct-reading scale to see if the deflections over the lower and upper halves are equal. A difference of more than 3 standard deviations is a clear indication of trouble; it may be caused by electrostatic effects, dirt or damaged knife edges. Other tests should be run to monitor the precision, the course of the zero with time, and the sensitivity. Supposing that after running such tests, one finds that the zero drift is significant and follows some pattern, such as temperature fluctuations shown in Fig. 1; this is a source of possible systematic error. Accordingly the appropriate weighing procedure to follow is one that takes it into account, e.g. the ASTM recommendation in which the zero is observed just before and right after a pair of weight observations; the average of the 2 zero readings is subtracted. The set of weights should be tested and if necessary recalibrated. Many weight calibration procedures have been published, both for sets of weights ' as well as for the built-in weights of substitution balances ' . Also the direct reading scale should be calibrated. An example of this is the procedure of Bowman et al who observed the response of the indicating scale to the addition of the same small weight at different points along the scale. Small non-linearities are common in the optical scales of substitution balances and calibration will provide suitable corrections that can be applied; however electromagnetic scales are fundamentally linear and so any observed non-linearity indicates trouble that needs the attention of service personnel. Having tested and calibrated an instrument and taken some observations on the weight of a target, the final problem is to compute the target's mass. First the corrections to the weights used for a given observation are applied. Also any correction for the on-scale reading is applied. Then, except for specialized work where the weighing is done in vacuum, the correction for air buoyancy ' must be computed. Some examples of the magnitude of air buoyancy corrections for observations versus steel weights are given in Table III. The relationship of the true mass of an object, m, to the observed reading ID , when it is weighed in air versus weights of known mass is given by the equation

m = mg I 1 + °frDJT--^ - )r) I ] (1) m w -i where p is the air density, p is the density of the object weighed and p is 92

Table III

ICx,inples of Air Buoyancy Corrections

Object weighed % Correction

I'l.itinum - 0.009 Germanium - 0.002 Aluminum + 0.035 Beryllium + 0.05 W.-iler + 0.10

the density of the weights. The air density can be computed from observations of. the temperature, pressure and relative humidity and it varies by about 3Z in any one location for the normal fluctuations in atmospheric conditions ' ; typically at sea level it is 1.18 mg cm and at elevations of 1500 m about 1.0 mg <:m . For a rigorous treatment of the air buoyancy correction, the densities of the object weighed and the various weignts used should be determined; this is a great deal of work; furthermore it frequently happens that weights of several materials will have been used in performing a weighing, e.g. stainless steel for the _* 1 g weights, aluminum for the < 50 mg weights and tantalum fcr the intermediate weights, and this adds to the complexity. So for routine work this becomes impractical. Thus the apparent mass basis was established to simplify the p'roblem and yet remain sufficiently accurate for most work ' . The basis is this: that weights made of various materials are adjusted so that they appear equal to their nominal values when compared in air with weights of the "ideal" density. The normal conditions specified for air density are 1.2 mg cm at 20°C; in the U.S. the specified "ideal" density is 8.4 g cm at 0°C (corrected tor thermal expansion it is 8.3909 g cm at 20°C) and most commercial equipment is adjusted on this basis; however, in some countries 8.0 g cm has been selected as the ideal density . Thus it is important that the user ascertain from the manufacturer's data on what basis the weights are adjusted so that the appropriate density value (8.0, 8.39 or some other) is used in computing the buoyancy correction according to equation (1). Adoption of the apparent mass /. basis adds an inaccuracy of < 10 ""% unless the air density is considerably lower, -4 e.g. at elevations of 2000 m the error introduced is - 3 x 10 %; even this is acceptable for most purposes and very likely is a negligible part of the error in determining small masses. Workers in many fields have reported precision such as is shown in Table IV. With sensitive instruments, precision of a- a few pg frequently is obtained. For our purposes one concludes that the accuracy of mass determinations of targets should be similar to this precision if care is taken in maintaining the performance capability of the balance and in its calibration.

Table IV

Precision Reported by Workers in Various Fields

Precision Load Type of Balance

(12) Bowman et^ al. ± 60 Mg 1 kg Substitution

BIPM(16) ± 1.5 Ug 1 kg Classical 2-pan

Aimer et al. few in 10' 100 g Semi-micro substitution (9) Lashof & Macurdy + 44 yg 40 g Semi-micro substitution

Marinenko & Foley ± 3 Ug 2 g Micro substitution

Henir.s(18) * lug 5-18 g Micro substitution (19) Green ± 2 pg 1 g Classical 2-pan (12) Bowman et axal.KJ ~^J ± 0.5 pg 2 g Electronic (19) Green11 ^^ ± 0.1 pg various Electronic 94

References

1) L.B. Macurdy, "Measurement of Mass", in "Treatise on Analytical Chemistry", Parr. I, vol. 7 (eds. I.M. Kolthoff, P.J. Elving and E.B. Sandell; Wiley, New York, 1967) p. 4247. 2) L.B. Macurdy, J. Res. Nat. Bur. Std. 6J5C (1964) 135. 3) H,S. Peiser, Report 45th Nat. Conf. "Weights and Measures", U.S. Nat. Bur. Std. Misc. Publ. 235 (1960) p. 45. 4) T.W. Lashof and L.B. Macurdy, U.S. Nat. Bur. Std. Circ. 547 (1957). 5) P.E. Pontius, U.S. Nat. Bur. Std. Tech. Note 288 (1966). 6) L.B. Macurdy, "Do It Yourself Plan for Mass Measurement", Instr. Soc. Am. 20 Hi Ann. Conf. (1965) 14.8-5. 7) H.E. Aimer, "Weight Cleaning Procedures", NBS Report 9683 (1968). 8) "Standard Methods of Testing Single Arm Balances", ASTM Designation E319-68 (1968). 9) T.W. Lashof and L.B. Macurdy, Anal. Chem. 26 (1954) 707. 10) "Balances, Weights and Precise Laboratory Weighing", Notes on Applied Science No. 7, National Physical Laboratory, U.K. (1954). 11) H.E. Aimer, L.B. Macurdy, U.S. Peiser and E.A. Week, J. Res. Nat. Bur. Std. 66C (1962) 33. 12) H.A. Bowman, R.M. Schoonover and M.W. Jones, J. Res. Nat. Bur. Std. 71C (1967) 179. 13) P.E. Pontius, U.S. Nat. Bur. Std. Monograph 133 (1974). 14) P.K. Faure and J.E. Gledhill, Anal. Chem. _30 (1958) 1304. 15) P.H. Bigg , J. Sci. Instr. J36_ (1959) 359. 16) "Proces-Verbaux des Seances", Comite International des Poids et Mesures, 2e Serie, Tome 41, 62e Session (1973), Bureau International des Poids et Mesures, Sevres, France. 17) G. Marinenko and R.T. Foley, J. Res. Nat. Bur. Std. J75A (1971) 561. 18) 1. Henins, J. Res. Nat. Bur. Std. 68k (1964) 529. 19) E. Green, Medzinarodne Syrap. Metrologie (INSYMET, Bratislava, 1972) p. 61. 95

MEASURE OF UNIFORMITY OF FISSILE MATERIAL TARGETS BY THE FISSION TRACKS METHOD J. KWINTA and F. AMOUDRY CEA

Centre d'Etudes de Bruyer es-le-Chatel B.P. n° 61 - 92120 MONTROUGE FRANCE

INTRODUCTION

Neutrons flux densities, fission yields and fission cross-section measur ements require the knowledge of the number of atoms and (or) the fissile nuclei density per area unit of a target. In many cases, the uniformity of the target is important arid its determination is often necessary even in the case of very thin films. Because of its great sensitivity, the ionographic (or fission tracks) method constitutes a good operating process for fissile material distributions measurements. If a target of fissile material is irradiated with an isotropic neutron flux, the smallest variation of the deposit density is converted into a variation of the density of the fission tracks on the ionographic detector.

Results presented in this paper were obtained with uranium targets of 30 to 40 (jig/cm thickness irradiated with neutrons. The objectives were the following :

1) The selection of a suitable technique for targets preparation which gives ade- quat results with respect to uniformity and fabrication cost.

2) The calibration of an automated tracks counting device. Our Laboratories facilities allow us to prepare the fissile material targets by electrodeposition, vacuum evaporation and electrospraying.

EXPERIMENTAL

I- Targets preparation The specifications were the following : - isotope deposited : 238 U - diameter of the deposit : 0. 8 mm - thickness : 30 - 40 u.g/cm^ of uranium - substrate : platinum disk 10 mm diameter , 0.2 mm thickness - chemical nature of deposited material : oxide - yield of deposition : maximum. The techniques employed are now well known and the apparatus were thoroughly described (1). We want only to remind here our operating conditions concerning each method. 96

a - electrodeposition : electrolytic solution : ammonium oxalate 0. 15 M ; pH - 8 volume : 1 ml voltage : 12-16 volts current density : 0. 33 A/cm anode : platinum wire rotating at 600 rpm.

1) - vacuum evaporation :

pressure in the vacuum chamber : 10 mm Hg iempvratuiv of filament : 2 000°C starting material : U3 Og area of deposit on the filament : 2 mm filament to substrate distance : 10 mm

c - electro spraying : solution sprayed : uranium acetate 1 g, 1 in methanol capillary material : glass capillary diameter : 0.450 mm anode : tungsten wire 0. 390 mm diameter high voltage : 7500 volts.

Electrodeposited and electrosprayed targets were calcinated at 650"C during 15 minutes.

II- Fission tracks production and measurements :

The targets were put in contact with ionographic detectors (Makrofol foils 60 (j.m thickness) and the assembly was irradiated with neutrons. The time of irradiation is calculated in order to produce several hundred fission tracks observable in a microscope field. The fission tracks density on the detector is given by the relations :

D = p -N>5~ (0t) for a thin layer

Dt " ?R ' A?^r"' (0*) for a thick layer

with D = fission tracks density

P = detection yield

(j- = fission cross section of the considered nuclide

MA = number of fissile nuclei per cm of target (thin source)

>vf = number of fissile nuclei per cm of target (thick source)

R = mean free path of fission products in the target

(

After the appropriate irradiation time the detectors were separated from the targets and etched during 40 minutes in a 25% by weight potassium hydro- xi de solution. The goal of this etching is to permit a sx "ficient magnification of the tracks in such a way that a microscopical observation is feasible. Each optical field area was 6. 25 x 10"^ cm . Fission tracks were counted on photographs. The overall magnifying of the system was about 256. In these conditions each target area scanning was made up with 400 photographs. The determination of the fission tracks density on a diameter of a target was made up by counting the tracks on 20 to 24 photographs. This method is very tedious. However, the counting of absolute tracks is necessaiy to calibrate automated and more sophisticated devices such as photodensitometers (2) or quantimeters. With such apparatus a very rapid determination of the Hpnsity of the tracks can be achieved on the ionographic detector itself.

RESULTS

Figure 1 shows the " cartography" of electrosprayed and vacuum evaporated targets.

The variations of the density of the tracks are represented with darkness diff erences.

• -• • • B_U •« F^ a B_B . -i *"* .£ a"a a a"' a a a-a > -a a *w a •.• a p_a a • a _ • a B n • p^a a •_• - a •_• a m a a • I • a a a a a a a "a a B'"J i •"• a •"• a . " " a"e • -••Mi •>• a B-- a »-a a a~ «-a * B-B • •>• B •»• a B-a • a-a - B ' nil aa a^B a njm a a a a • a ,.- a >-_• a II a • a a a e •_ ~ i a »"a iuiiiu'i a .-J"B " i a B~H a B-a a i»>d a p-a B a-a - i a e-a II.«II • a-'B • a-a - a-« a a •yaijtfk •.•"•• a a a i a"a a B-a a B-a a •«• a B-a a B--B a-B a »«a a B-B a »-a a a-» B-a » ••I • ITW-B : jrw wjrw WJM a ~._a i a a • < 1II • ••III* II "i-*-* &"• B-I a ; •< a a-« n-li B'-a 11(111 p,# B I"«"B M - B'B B B"I Mlix tli'a • a MlHlMlH- B-B laiiiiaaiiiiii " . • •"• ik'M a p'a a a"a • i-di'- a s>- t -I it,'iectro sprayed Vacuuiii evaporated

Figure 1 - Cartography 93

For quantitative determinations the distribution of the tracks on a diameter of tiio target was plotted. This is shown on figure 2. Results are as follows : 1) The- electrode posi ted target is non uniform and the distribution non symetrical. The relative variation of the thickness between the side and the zone where thickness is maximalisll07o.Irregulariti.es of deposit exist and are probably caused by insufficient surface treatment. Our experience based on thn preparation of thousand electrodeposited sources of actinide elements fcr analytical purposes shows us that it is very difficult to obtain good umfurmi- ties when using low voltage current.

2) The vacuum evaporated deposit appears symetrical with few irregularities. The deposit is curved and the relative variation of the thickness between the side and the center of the target is 59%. This is due to the fact that a pood deposition yield was researched, so the substrate was positioned near the filament. It is obvious that with usual operating conditions i. e. with a low [\ro) deposition yield, the layer must be more uniform.

3) The general distribution profile of the electrosprayed target is satisfactory. Maximum irregularities of thickness are within - 7 "!<> of the average value. The accuracy of the counting of tracks is about 3%. The edge effect was evaluated by scanning a side of the target. The density of the tracks falls from the average value to zero along a distance of 60 \x m - 10 )j.m. This is comparable to the mask thickness : 100 f^m. The mask was a metallized mylar foil.

CONCLUSION

1) The fission tracks method appears a valuable process for measurements of the uniformity of fissile material targets. Its sensitivity is very good and uniformity of samples as thin as 1 (ig/cm can easily be determined with a good accuracy. By adjusting irradiation times and neutron flux one can measure the uniformity of practically all the usual thickness of targets. The method is non destructive provides the target offers sufficient mechanical resistance to be in contact with the detector.

2) Electrospraying is a technique convenient for the preparation of fissile material targets. The fact that non reproductible results were obtained in different laboratories indicates that it would be useful to standardize the operating conditions of this method.

Acknowledgments :

Authors are grateful to Mme GRENIER , wl. GENDRE for fission tracks determinations and to MM. PIEL, BOULIN and PERRLSfE for targets preparations.

References :

(1) Third international symposium on research materials for nuclear measurements - Gattlinburg 1971 - ORNL CONF 711002.

(2) MOURGUES M. , RENAVOT J. M. , IAEA - SM 160/45 - p. 627 to 650. 99

o c

Z Q

ELECTRO5PRAYED rig c

XI D

-X—•*-—K^- _

O VACUUM EVAPORATED

f.y

3 > a L.

L. O

< c

9. 2 TRACKDENSITY DISTRIBUTION 100

Preparation and Handling of Calcium Metal Targets

.] .D. Stinson Division of Phi's i cs, .'iational Research Council, Ottawa, Canada.

Abstract A technique for the preparation of calcium metal targets fro.ii CaJcium (.'.iriwnate is described. The method of handling and mounting these targets in a imn- react i vu environment to maintain purity is given.

I nt rodnct i on Over the past five years a technicje has been developed for the preparation and handling of calcium metal targets. The technique de-scribed differs frara those usually reported, in that special attention is paid to deposition and handling to maintain target purity. The targets are prepared by the decomposition of CaC0_ and the subsequent reduction of CaO, and are gold backed. Thicknesses range from 5 to 160 ug/cm2. These targets have been used for (u,y) studies at the Van Je Oraaff Accelerator installation at the National Research Council of Canada.

Vacuum System The vacuum system used is a modified commercial unit which consists of a 50-cm bell jar and a 10-cin oil diffusion pump. A water cooled chevron baffle and a liquid nitrogen chevron baffle provide the trapping for the work chamber. A zeolite trap between the fore-pump and diffusion pump prevents oil backstrearo- ing. Roughing of the work chamber is accomplished with a separate zeolite trapped pump. This system is capable of providing a base vacuum of better than 10 ' Torr after six hours of pumping and permits the evaporation to take place at about 10"G Torr.

Substrate Preparation J"hc substrates are cut from 0.25 mm gold ".ieet of 99.99% purity. Two sizes of target backing are used: one si^e is 2.5 cm x 2.5 cm and the other a 1.9-cm diameter disc. Cleaning of the gold is performed by chemical treatment. The gold substrate is boiled for 5 minutes in each of the following solutions: 10 lvT-% KOH, 50 Vol-% HC1 and 50 Vol-% MHO . Chemical carry-over is prevented by using separate beakers and by following each solution with a 3 minute boil and rinse in de-mineralized water. 101

Source Preparation The source used is a closed boat type, 5 cm long with a 3 Rim aperture in the centre. The source is fabricated from 0.125 ram tantalum sheet to form a S mm diameter tube with flattened ends. A pre-firing of the source at 2150 C for 5 minutes in vacuum is performed to clean the tantalum and deoxidize the surface. If tiie charge of CaCO, to be used exceeds 15rag, tantalum chips are included in the source to assure complete reduction of the CaU.

Transfer Chamber A stainless steel chamber, lig. i, 13 cm long x 5 era diameter, equipped with a bleed valve and flange is useu to remove the target from the bell jar. The chamber, with the substrate mounted in the open end is suspended by a short length of chain from a beam attached to a sliding, rotary feedthrough. Pre-heating of evaporant is accomplished with the chamber in position A. In ihis position the substrate is protected from condensable impurities by a stainless steel shield. Evaporation takes place with the chamber in position B. After allowing the apparatus to cool for 20 minutes, the chamber is swung to position C for sealing. The sealing operation is performed by lowering the chamber onto a metal disc with 0 ring seal, closing the main vacuum valve and bleeding in ar^on to atmospheric pressure. The chamber containing the target is then removed from the vacuum system. The sealed chamber is clamped shut and along with the target mount is placed in the interchange lock of an inert atmosphere enclosure, Evacuation and back filling of the lock with argon allows introduction of this equipment to the glove compartment.

Evaporation The decomposition of calcium carbonate and reduction of calcium oxide is described by the following two equations:

CaCO, •+ CaO + CO,f (1J

5Ca0 * 2Ta •+ Ta,CL + SCa- (2)

With the chamber in position A, Fig. 1, the source temperature is slowly increased to evolve the moisture and C07 from calcium carbonate. Raising the lemperature too fast will blow the salt out oF the boat. The temperature of the boat is gradually increased to 1190 C, at a rate that limits the pressure in the work chamber to 10 5 Torr. The evolution of CO, begins at approximately 600 C. The CO ceases to be evolved at about a temperature of 800-900°C. A decrease in pressure to about 5 x 10 7 Torr signals the cessation of CO evolution. The time for dehydrating and decomposing the calcium carbonate will vary from IS minutes to 102

•to minutes uVpen Jing on Che quantity of CaCCL in the boat.

fT TRANSFER CHAMBER A-PRE-HEATING B-EVAPORATION C-SEALING t t

With the chamber in position B, the temperature is increased to 1650 C to start reduction of the CaO and calcium deposition is observed. This temperature is held for 5 minutes within ±25 C. During the deposition process, pressure readings will increase, reaching a maximum of 3 x 10 ' Torr. The temperature figures quoted are readings taken with an optical pyrometer and have not been corrected for emissivity.

Inert Atmosphere Enclosure Two different inert atmosphere enclosures have been used for mounting targets. The first enclosure, a commercial glove box, used an inflatable balloon and valve system to replace the air with argon. This unit used large quantities of argon to obtain a dry atmosphere, and especially so after a period of nonuse. 103

Dew points of -As C were attainable, but i jcause permeative transfer of moisture was high a new enclosure was built. The enclosure built at \.R.C. is of stainless steel construction and consists ot~ a cylindrical tank with viewport, glove ports and interchange lock. The glove ports are blanked off and are evacuated simultaneously with the main compartment. The valve system allows the gloves to be stored in vacuum or after back filling with argon. A vacuum ol' 10 ^ Torr is attained with a zeolite trapped pump. Argon is back filled through the saint? trap to give an initial dew- point of -55 C. Circulation of the argon through the zeolite trap for further drying i1'. provided for by a carbon vane pump. Dew points of -Co C have been reached after ."50 minutes of circulation with the glove ports sealed. While usin, tin: gloves, the gas circulation system limits the rising dew point due to permeation through the gloves at approximately -55 C.

Transfer of Mounted Target » The mounted target is transferred to the accelerator bea;n line in argon. lwo types of target mounts are in current use, each, requiring its own method of atmospheric isolation during transfer. The first mount used employs a square target with indium seal to close a water-cooled cavity. The cooled cavity is fed water by supporting tubes attached to an 0 ring flange. This type of mounted target is scaled in a polythene bag prior to the transfer from within the enclosure to the accelerator. The flanged beam line which accepts the target mount is brought to atmospheric Pressure using argon just before transfer is initiated. Argon is made to flow out of the beam line at a considerable rate to prevent the entry of air. The plastic bag is opened in the argon stream and slipped over the beam line flange. The target mount is plugged into the beam line and after stopping the flow of argon, a seal is effected by immediate vacuum roughing. The second mount used, employs a round target which is sealed directly to the gold substrate by an annular knife' edge. This mount consists of two close fitting concentric tubes each mounted on its own flange. The knife edge annulus, sandwiched between the closed end of the outer tube and the gold substrate on the one side, and the open end of the inner tube on the other side, effects a seal on the gold when the flanges are bolted together. Cooling of the target is achieved by circulating water through pipes connected to machined passages between the two concentric tubes and across the back of the target. Atmospheric isolation of the target during transfer is accomplished by means of a stepped plug with 0 ring seal inserted in the beam aperture of this mount. The plug acts as a check valve and is put in place before the mount is moved from the main compartment, to the interchange lock of the enclosure. By evacuating the argon in 104

the lock to -100 Torr and by back filling with argon to atmospheric pressure, the

check v.ilve is made to seal the mount. The sealed mount is installed on the

accelerator beam line which has been previously flooded with argon. Vacuum

roughing causes the check valve to fall to the bottom of the beam line where a

metal shield provides protection against accidental beam impingement.

General Comments

Researchers using these targets determine thicKness from widths of resonance

peaks in the yield curve. Adjustments to target thickness are made as a result

of these findings. In agreement with. G.'l'.J. Arnison1 it has been found that 15 ::

.if CaCO.. produces a 40 ..g/cm' deposit at a 6-cm source to substrate distance.

1-or i <.,<) studies the important contaminants affecting target purity are

the 11M\ ~ elements. These elements give high background because of large cross-

sect ions lor i reactions.

GoJei sheet purchased from one supplier as 99.999% pure was unusable because

it contained a low level aluminum .impurity. Gold from another supplier'- of

J'J .-J'J \, purity has been used for these targets and found to be acceptable. The

gold from both suppliers was chemically cleaned by the procedure previously

doscri !>ed.

Care must be taken to exclude all other sources of contamination. On one

occasion, boron from distilled water contaminated a target. The distilled water

jn'oduced by a borosilicate glass still had been used instead of the usual de- mincra'i;ed water. Fluorine from a teflon bushing in the vacuum chamber was another source of contamination even though the teflon showed no evidence of heat damage .

Chemical analysis of target materials has shown that important contaminants are well below 2 ppm,

More than SO targets have been made by the method outlined here and improvements in equipment and technique are continuing.

Acknowledgment

The author gratefully acknowledges the assistance and encouragement of

Dr. U.S. Storey and Or. IV. R. Dixon. Thanks are also due to Mr. D.C. Elliott who mounted the targets and gave invaluable technical assistance.

1,1) United Kingdom Atomic Energy Authority AlVRIi Report Xo. 0-32/67 (2) Johnson Matthey $ Mai lory Ltd. 105

PERFORMANCE OF THE UPGRADED CRNL MP TANDEM ACCELERATOR

N. Burn & L.B. Bender Operations Division Chalk River Nuclear Laboratories Atomic Energy of Canada Limited

ABSTRACT

In 1967 an MP Tandem Accelerator was installed at Chalk River. The accelerator was upgraded for 13 MV operation by the installation of a set of high gradient stainless steel electrode accelerating tubes in September 1972. Further upgrading took place in June 1974 when the original charging belt was replaced by a new charging system, a "Pelletron" charging chain. The effects of these changes on the overall performance of the accelerator are reviewed.

1. Introduction

Electrostatic accelerators and Chalk River have had a long and honorable association. In the middle 1940's the Electrical Engineering Department of the National Research Council in Ottawa began building a 3 MV vertical electrostatic accelerator which was subsequently turned over to Chalk River in 1949. By February 1952 this machine had produced 10 pA beams of 2.5 MeV protons. Next, a request in October 1954 to High Voltage Engineering Corpo- ration (HVEC) from Eric Paul for a 10 MeV particle accelerator led first to a small development contract from Atomic Energy of Canada Ltd. (AECL) for a 6 MV "tandem" accelerator and shortly thereafter, in September 1956, to an order for the 5 MV tandem now known as the EN. After successful tests at the HVEC plant in Burlington in the summer of 195R the first EM tandem was moved to Chalk River where on the 20th of March 1959 it produced a 1 yA beam of 8 MeV protons. Although proton beams up to 12 MeV were soon obtained, it was not able to produce the helium beam dearly wanted by the Chalk River physicists. Their disappointment was however soon lessened by getting 100 nA of O ions during the summer and observing resonance structure in the 12C + 12C system early in 1960. 106

The EN tandem No. 1 served well in Chalk River until 1966 when it was dismantled and shipped to the Universite de Montreal where, rebuilt and improved, it still forms the basis of a vigorous research program in nuclear structure. In due course, MP No. 3 went into service at Chalk River in April 1967 and five years later, in September 1972, was the first to be upgraded to 13 MV by the installation of stain- loss steel electrode accelerating tubes newly developed by HVEC. Two years later, in June 1974, the MP was further upgraded by the installation of a "Pelletron" charging syster. manufactured by National Electrostatics Corporation (NEC).

The MP tandem was installed in a new building parallel to the old EN machine room with a new target room between them as shown in Fig. 1 .• •

3560-J

GAMMA GONIOMETER

OTUS ARRAY OF 7 GAMMA // GAMMA TABLE // COUNTE

OD3 BROAD RANGE SPECTROMETER

ON LINE ISOTOPE SEPARATOR MP TANOEM CONTROL RACKS

-BEAM TRANSPORT CONTROL

Fig. 1 Layout of Chalk River MP Tandem Facility The former EN target room and one end of the old machine room. then became target rooms 3 and 2 respectively for the MP. Together the three rooms give approximately a dozen semi- permanent target locations.

The principal pieces of experimental equipment are indicated in Fig. 1. The high resolution, broad range, high transmission QD" spectrometer is installed and measurements of the aberrations are in progress. The window- less, differentially pumped gas target, the 17" Ortec scattering chamber, the seven counter Lotus goniometer, the seven gap orange spectrometer and the computing facilities have recently been briefly described by Hardy, McDonald and Milton.

Ion Sources

Many ion sources have been used with the MP since 1967. At the present time, the principal sources are an HVEC Heinicke-Penning radial extraction source and a General lonex LCE source. A helium recovery system is available for use with the LCE source or the earlier HVEC charge exchange duoplasmatron. Any of the above sources can be used with an Ortec chopper-buncher system installed in 1973 which allows pulsed beams to be obtained with pulse widths less than 1 ns for p and 5 ns for 0. Pulse separation times are variable down to 400 ns.

3. Accelerator Tube Upgrading

The accelerator tube upgrading program began January 16, 1972 with the removal of the old aluminum electrode tubes and culminated with the completion of the acceptance tests of the new stainless steel electrode tubes August 31, 1972. Since the new tubes were guaranteed to operate at 13 MV with a reasonable expectation for operation to 15 MV, it was decided to test the terminal and column structure to 15 MV before their installation. It was quickly found that sparks between the underside of the high energy column and the tank floor limited operation to much lower voltages. The best performance war always seen immediately after closing the tank when voltages as high as 14 MV were 108

sometimes reached. Invariably, performance deteriorated after two to three days with tank sparks occurring at random intervals and random voltages above 10 MV. There was no evidence of improvement through conditioning. The fact thec essentially all of the sparks occurred to the tank floor under the high energy column implicated the belt dust tnat was present in copious quantities in this region and belt debris ras finally concluded to be the major cause of high voltage breakdowns. The dust hypothesis was tested by placing a sheet of fiberglass window screening along the bottom of the high energy column. The screening was put inside the rings to prevent most of the belt debris from reaching the floor. The screen improved the voltage holding properties considerably,'so that after six days of operation the accelerator could sti.ll hold 13.1 MV. To reduce dust generation, the black belt, which had only been in service a few months since December 1971, was replaced by one of the new "tan" belts.

Following the installation of the new belt, the tank was closed on May 13, 1972. Shortly thereafter, voltages of 13.6 and 14.0 MV were held for periods up to 21 hours. The voltage was then raised to 14.6 MV and then to 15.1 MV where it remained for two hours. Finally the voltage reached 16.2 MV where runs of 40 minutes and 22 minutes were made between breakdowns. Subsequent runs of 12 and 10 hours were made at 14.6 MV and the last test was a 24 hour run at 15.1 MV, which wa^ terminated without a breakdown. Following the installation of the new belt, only 12 breakdowns were recorded in 160 hours of operation above 13 MV.

These final tests were carried out with 92 psig 2 (0.73 MN/m ) SF, containing 5 percent air and 5 to 10 ppm of b moisture. During pump-out it was possible to hold 15.1 MV until the pressure reached about 75 psig (0.62 MN/m ). The chief effect of lower pressure seemed to be a greater ten- dency for repeated breakdowns. In routine operation where the terminal reaches 13.5 to 14 MV during conditioning, a 2 SFg pressure of about 80 psig (0.65 MN/m ) has proved adequate. . 109

The only important factor in the occurrence of tank sparking is the nature and amount of belt debris generated in the accelerator. Other factors such as gas purity and pressure and the surface quality of the column rings and tank wall are not critical within reasonable limits. This has been confirmed by all our operating experience since the upgrading. We found that under dust free conditions, operation at 13 MV is possible for days at a time without tank sparks.

The installation of the new tubes and the associated terminal equipment took about one month.

In the initial tests, the tubes worked well up to 10 to 11 MV. Several weeks were then spent in an attempt to condition the tubes to 13 MV. This procedure stalled around 12 MV with many tube sparks and no advance. Another approach was finally tried in which the voltage was forced up rapidly. About 10 tube sparks resulted over a two hour period but the conditioning point was increased to 13 MV for the first time.

Following the recommendations of HVEC, conditioning had been carried out in the absence of stripper gas. When the Oo flow was established in preparation for a proton beam at 13 MV, it was found that the conditioning point dropped back to about 11 MV. Oxygen was found to produce the same effects as before, whereas the presence of N_ appeared to improve the voltage holding properties. It has now replaced 0^ as the stripper gas in the terminal.

The introduction of N2 as stripper gas was the last change required to permit reliable operation at 13 MV and the acceptance tests were completed shortly thereafter. The acceptance criteria consisted of two hours of steady operation with the following proton beams and energies: 5 uA at 6 MeV, 10 uA at 15 MeV and 5 yA at 26 MeV. These criteria were easily met using either a direct extraction duoplasmatron or a Heinicke-Penning ion source. The transmission from 3.5 MV to 13 MV was greater than 40% for protons from both sources and was 4 0% for oxygen from the latter source. However, for He beams from the HVEC lithium 110

vapour charge exchange duoplasmatron, the maximum transmission is about 20%, which suggests a larger emittance for this source.

•I. Charging System Upgrading

The accelerator tube upgrading program in 1972 had shown that the major voltage limitation on the accelerator was imposed by belt dust. It was therefore decided to further upgrade the accelerator by replacing the charge belt with a "Pelletron" charging chain system. This system consists of 6 chains, 3 in each end of the accelerator. The chains are made up of steel cylinders or "pellets" 1.25 inches long by 1.25 inches in diameter, joined together by nylon links (Fig. 2). Each chain is capable of carrying at le^st 100 microamps of charging current up to the high voltage terminal.

Fig. 2 The centre Pelletron chain being installed on the drive sheave. Ill

The charging system upgrading program began on flay 26, 1974 with the removal of the original belt, drivernotor guides and other associated hardware. Within three weeks, the new Pelletron system had been installed and was ready to test. However, a multitude of minor mechanical problems delayed the final testing. These problems were mainly associated with the idler pulleys which are used to guide the chains down the inside of the accelerator structure. Satisfactory temporary solutions for all of the problems were found and the accelerator was returned to routine operation on August 2, 1974.

Effects of the Upgrading Programs

Prior to the installation of the high gradient tubes, the accelerator had operated routinely at voltages up to 10.5 MV mainly using light ions. For six months immediately following the installation of the new tubes, the accelerator operated at voltages up to 13.7 MV with more than 67% of the operating time being spent at voltages in excess of 10.5 MV using a wide variety of heavy ...s. Because of belt dust uxubierus, the maximum operating voltage was thpn deliberately limited to 12.5 MV to avoid tube damage. The effects of the tube upgrading program on the accelerator operating voltage and the accelerated ion types is illustrated in Figs. 3 & 4.

The benefits of the charging system upgrading program in 197 4 were immediately obvious in the much lower frequency of breakdowns at high voltage; equally significant was the remarkable reduction in terminal ripple. Following the Pelletron installation, the operating voltage distribution curve was again almost identical to that for the period immediately following installation of the new tubes; the accelerator has operated at voltages up to 13 MV and more than 70% of the operating time has been spent above 10 MV but with a significant reduction in the number of breakdowns. The lowest monthly average breakdown rate previously observed above 10 MV was three breakdowns per day; this rate has dropped to one breakdown per day. On one occasion, the accelerator operated continuously for ten days at voltages up to 12.6 MV without any breakdowns. With the Pelletron, the unstabilized terminal ripple has been "educed almost a 112

3580-B

0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 TERMINAL VOLTAGE (MV) —

AFTER UPGRADING

1.4 2.5 3.5 4.5 5.5 6.5 7.5 8.5 95 10.5 11.5 12.5 13^ TERMINAL VOLTAGE (MV) — 13.7

Upper histogram - typical distribution of the operating time at the indicated terminal voltages prior to the accelerator tube upgrading.

Lower histogram - the same distribution for the eighteen month period following the accelerator tube upgrading. 113

BEFORE UPGRADING : 15-

O ZIO

XI 'H 2D VHe6Lrui lni I2C I4N l60 l80 i9F » ION SPECIES —-

AFTER UPGRADING

'H2D3He4He 13C I4N l60 l70 l80 l9F24Mg ION SPECIES —

Fig. 4 Upper histogram - typical distribution of the operating time with various ion species prior to the accelerator tube upgrading.

Lower histogram - the same distribution for the eighteen month period following the accelerator tube upgrading. 114

factor of 50 from that observed with a charging belt. Ripple as low as 50C volts peak to peak has been observed at 10 MV.

Terminal Equipment

During tha accelerator tube upgrading program in 1972, several pieces of electrical and electronic equipment were installed in the high voltage terminal. Power for this equipment is supplied by a "GEORATOR", a permanent magnet 110 volt, 3 phase, 400 Hertz generator, which is belt driver, by the terminal pulley.

Because of the low vacuum conductance of the new accelerator tubes, it was necessary to install a terminal pumping system. The major stripper gas load is pumped by four "Ti-Bail" sublimation pumps; at the maximum stripper gas flow rate "Ti-Ball" lifetimes have been in excess of 350 hours. At each end of the stripper canal, one litre per second conductances limit the strapper gas flow into the accelerator tubes.

In order to utilize further the full capability of the upgraded accelerator, it was obvious that foil as well as gas stripping should be available. A National Electrostatics Corporation foil changing mechanism holding 114 foils was installed immediately ahead of the stripper gas canal. Carbon foil thicknesses ranging from 2 to 3 0 pgm/cin have been tested and 2 ugm/cm foils have been chosen for regular use. These foils do not exhibit significantly shortGi lifetimes 2 than 5 or 10 pgm/cm foils, yet they have lower multiple scattering, resulting in increased yield by as much as a factor of 2 for the heaviest ions. A technique has been developed for mounting these foils using a cellulose nitrate coating that enables them to be mounted on frames and installed in the accelerator terminal with a success rate near 100%. The coating evaporates off the foil during the first minute of exposure to a heavy ion beam, as revealed by an increase in intensity during that period. Considerable experience with foil lifetimes has now been accumulated. For injected negative ion beams of about 3 microamperes typical foil lifetimes are: protons or helium >2 days/ oxygen ~6 hours, sulphur R;1 hour and bromine (1.5 uA) «15 minutes. These lifetimes are measured as the tLnc until 1/2 the original intensity it obtained or else until complete rupture of the foil. Beam profiles arr>. readily observable on exposed foils, and on thicker foils (30 Ug/cm ) some unusual patterns have been observed (see Fig. 5). This Figure also .^IUJS the excellent focus attainable for short times in the horizontal plane (beam diameter «l/4" (6.4 mm.)). The larger diameter in the vertical plane may be from an asymmetric focus produced by a gradient change across a slotted aperture in tub<_ 1 or may be from, vertical motion induced by gradient chances in the inclined field profile. Thinner foils, exposed for longer periods usually show uniform darkening of the area behind the aperture protecting the foils (Fig. 6). Rupture of the foils tends to occur near the edge of the darkened region (Fig. 7).

2 Fig. 5 A 30 yg/cm carbon stripper foil which has been exposed to an ion beam in the high voltage terminal of the accelerator. A fine rippling is evident in the area of the foil that was bombarded by an ion beam. The diameter of the foil is 10 mm. 116

Fig. 6 A 12 Ug/cra carbon stripper foil which has been exposed to an ion beam in the high voltage terminal of the accelerator. The uniformly darkened area was defined by a 10 mm diameter protective tantalum aperture normally located in front of the foils. The diameter of the foil is 15 mm. 117

Fig. 7 A 2 yg/cm carbon stripper foil which has been 24 exposed to a 5 nA Mg beam. Rupture of the foil has occurred near the edge of the darkened region.

Experiments are in progress to test the effects of annealing the foils prror to installation and to examine the source of foil thickening, through the use oi C foils.

7. Summary The installation of the HVEC stainless steel, high gradient accelerating tubes increased the useful voltage of the CRNL MP Tandem accelerator from 10 MV to better than 13 MV. These tubes provide transmission of 40% or more for protons and heavy ions from a direct extraction duoplasmatron or a Heinicke-Penning ion source and about 20% for He beams from a charge-exchange duoplasmatron. It was found that O_ stripper gas is unsuitable for high voltage operation whereas M appears to have beneficial effects for high voltage performance . 118

Although the NEC Pelletron has been in routine operation for only 1000 hours so that it is too soon to evaluate its long term reliability, it has already produced, a significant improvement in the overall operation of the accelerator.

The Chalk River Tandem was the first KP to be equipped with both the stainless steel high-gradient electrode accelerator tubes and a Pelletron charging system.

8. Acknowledgements

The authors would like to thank the management of Operations Division for their continued support during these upgrading programs, the Physics Division for their cooperation and technical assistance and especially the MP Tanden operating staff for their perseverance with the actual hardware .

9. References

1. Manufactured by National Electrostatics Corporation, Middleton, Wisconsin.

2. J.C.D. Milton, G.C. Ball, W.G. Davies, A.J. Ferguson and J.S. Fraser, AECL-3563 (1970).

3. A. Olin, T..K. Alexander, 0. Hausser, A.B. McDonald and G.T. Ewan, Phys. Rev. D8(1973) 1633.

4. J.C. Hardy, A.B. McDonald and J.C.D. Milton, AECL-4596 (1972).

5. Manufactured by GEORATOR Corporation, Manassas, Virginia, Part No. 36-041.

6. Manufactured by VARIAN, Palo Alto, California, Part No. 916-0005.

7. This technique was developed at Chalk River Nuclear Laboratories by J.L. Gallant. 119

Sample Preparation and Definition at the

Central Bureau for Nuclear Measurements

2440 GEEL/BELGIUM

J. VAN AUDENHOVE, V VERDINGH, H. ESCHBACH, P. DE BIEVRE 120

i nt rodU'. t inn

Sample Prcpiiralim: L iect rospray ini; methods L i (Mrophorcsr Sed imciital ion 1 4. Layers prepared by vacuum evaporation I. i . 1 Boron layers 1. 4 . Metallic Uranium layers 1. 1 . >, L,K-layers 1. r 1 L'F.j foils 1. 4 Layers for oscillation measurements 1. 4. 6 Highly enriched s t a b1e isotopes 1. n. Layers prepared by triode sputtering 1. 1) Solut ions M e 11 i n i; a n d c a s 11 n i;

Sample Definition Classical approach I'.oron standard layers 1 . Uranium layers 1 .>, 2*tlPu Oi Synthetic preparation as a definition procedure 2 l Synthetic blends of elements l. i Liquids 2. ) l. l Solids I.I. I. Synthetic blends of isotopes 2.2.2. 1 . Liquids l.l.l.i.. Solid s 2. 3. Work in progress 2. J.I. Standard reference layers 2. 3.2. Standard reference alloys 2. i. 5. Thickness measurement 2.3.3. 1 . Stylus method 2.3.4. Physical behaviour of samples 2. '). T. Determination of traces of neutron poisons

References 121

!nt roduction

The Treaty of Rome setting up in ] 9 TT the European Atomic Energy Com- munity [Kuratom1 provided for the foundation of a Central Bureau for Nuclear Measurement s (CBNM'. Work on its establishment started at Geel (Belgium) in 1959. CBXM as a Nuclear Standards Bureau was planned in close collaboration with the European American Nuclear Data Committee and is primarily concerned with the promotion of accurate measurements of parameters of interest to nuclear energy. It therefore covers the fields of neutron, radioisotope s, stable and fissile isotopes measurements. The execution of these measurements very soon required the preparation of the necessary well defined samples and nence it was decided in 196] to establish a major facility for a sample preparation and assaying at CRNM to cover the needs of such samples for precise nuclear measurements relevant to nuclear energy. Since that time this Sample Preparation Group supplied about 30.000 samples covering about 1200 orders from 12 countries. The samples have been used mainly in:

y^arjia^le A c c e 1 e r a t o r s: Measurements of c oss - sections, studies of fission fragments, neutron beam filters. . . Reac to r Physics: Neutron flux mapping and dosimetry, fuel element studies, burn-up studies, reference samples for the determination of boron equivalents of construction materials used in reactors. . . _\tetallurgy, Chemistry, Mass Spectrometry, Activation Analysis, •• • - Alloys with certified composition, ultra pure metals and semi-conductors doped with certified amounts of elements. . . - Solutions, compounds mixtures, containing enriched isotopes of stable and fissile elements. . . Normally each sample is prepared and assayed according to the requirements of the applicant and samples are not kept in stock. It can be safely stated that in the past 15 years, CBNM has played a prominent role in the promotion of accurate measurements of parameters of interest to nuclear energy. The staff of the CBNM presently numbers 170, approximately 4 5 of whom are qualified academic scientists. About 10 have been involved in the preparation and assaying of samples. A review of the present state of the art of a number of sample preparation techniques as well as recent advances in definition methods at CBNM will now be given.

The preparation and assaying of radioactive standard sources and solutions at CBNM will not be treated here (31). An extended survey concerning this subject is in preparation (32). 122

. Sample Preparation

Table I lists the different sample preparation methods which are in use ,it Cl'.MN.

X_ajjlil I_ Summary of available preparation techniques Klectrospraying

;le; trophoresis Evaporation in vacuum ,ind nil ra-nigh vacuum - resistance heating - electron bombardment - high frequency levitation Cithociic sputtering Powder techniques - sedimentation - pressing - sintering Self transfer: ^ ""Cf sources Melting and casting - Resistance, Induction and electron bombardment heating - Alloying by high frequency levitation (crucible free) /{oiling (conventional and in vacuum) Machining "lectrodeposition banning

1.1. Electrospraying methods (I1 (2) (3) (4) (5) One of the main tasks was the preparation of thin layers of elements or isotopes available in very small amounts (ug-mg range) on large surfaces. The supports had very often to be very thin so that fragility increased the difficulties for preparation. Many of the isotopes were scarce and in addition often very radioactive and special handling techniques were necessary. Electrospraying methods were devised and improved for the preparation of such layers. This technique makes use of the fact that a solution or a suspension is forced through a narrow capillary by a high voltage (3 - 20 keV). A spray of tiny droplets is projected onto the backing that can be any metal or any conductive thin foil (metallized plastic, carbon, etc.). The solvent has to evaporate during the passage from capillary to backing. The acceleration by the high electrostatic field is enough to assure good adherence of the deposited particles to the backing. For the preparation of homogeneous layers of large area (up to 1000 cm ) various devices are used to move the backing. Fig. ) and 2 show two equipments as used at CBNM. In the first arrangement a complete and relatively cheap mechanical device is placed in the glove box, in a second arrangement the substrate holder is moved via coaxial rotating seals so that motors and gears remain outside the glove box. ]23

Fig. 1 : Electrosprayinn facility.

Fig. 2 : Electrospraying facil;ty (Motors and gears outside glove-box). 1 •?£

>m'!H 1'iialK available high volldji1 supplies capable of providing : to !'' k \ are ! i s e (1,

l '.i::-- r'.rl:uniil -Ic'.iil.-s of several capillary systems developed at CBXM h ; •. • • < • < • n r >' ] > o r! t • c j. •.'• • vo.spraving has proven its utility and broad applicability not only •'i ill :ul iblf iiimpoiiiids (e.g. acetates of the actinides) but also to the •:>:• ,-,i';LT, "f suspended particles (e. g. oxides of the actinides and the 1 . . elements I'.oroij, metallic tungsten isotope layers on C foils, etc. . i. (hi- method lends itself particularly well to the preparation, of layers '.ir. .ci-. thin foils it is one of the most widely used methods at CbXM - • (• mure {nan 10 years.

I h'- Josses during preparation are very small, the equipment is sample, M.I-XJIC! sivc .iiiri easily adaptable to glove box work. ' .iit'ipes of elements like I', rare earths, Th, U, Pu, Am and >.'p are ::. -.•:! |v.l., ted I'U rrently . : ,'. t :.- of 1000 cm" are being produced on any metallic or metallized iili-'i. support (1 'I n £ •' cm"- ', or carbon support. ' i\ e:-H of oxyrles up to IT mg U •'c i n - and boron up to 30 rag B/cm~ .-.:•• prepared bv suspension spraying. . ne noi:i()u<'!ieits of the eleetrosprayed deposits is better than i's. :'.;!•. eases better than Z"•> has been reached. Work is in progress •w i'-'ijruve the proven homogeneity. n ;• ': the ratio element to compound can be improved by calcining • u '. h < - o x i (i e . .1 h i: i- .sufficient information on the actual composition of the layer, -IM:[)]CS for analysis are taken from the spraying solution and from a '•iiiC iroi.'id the sprayed deposit (wi.tnt.ss samples sprayed during the ;, "in i r it i'.'ii of tne sample itself or duplicate spare samples made under 'n-- s.i'r;e •. o;;f]i ti on K

i..'. !•'. 1 ii ! rophorese (3) ( n! ((•>) '."if limiiinq factor in the electrospraying method is the amount of ii:.ste rial (hat i.an be deposited (a few milligram per cm ). The high '.oli.icc clei-t rophoretic deposition proved to be a versatile coating ::n tho'l for amounts up to IT mg'cm . !'he ciethod :is adapted to our purposes, worked quite successfully for the preparation of various samples of natural and highly enriched uranium.

! :. Sefl i m ent a tion : or tne preparation of thicker enriched isotope samples in oxide form ]j to several g•''cm-, sedinientation teenniques are used. finely divided • III.M1I.TS a n1 allowed to settle from suspensions or are compressed directly ir-.tn ippropriate containers. In many cases these samples are canned ;::f:i'i' \ac!ium or inert gas. i'he In nniqiio was used for tne preparation of various thick enriched .^otope samples, e.g. - r.i re earln oxide samples in the g/cni"" range - molybdenum oxide samples (various isotopes) - •. opper oxide samples (various isotopes) - •'' .\m samples - 100 mg/cm2 - '-']_Pri samplos TO mg - 1 g/cm - ' ''\p samples - 100 mg/cm'' - !. rjnium oxide samples (various isotopes) 125

i'xtreme i.irc is taken as to preserve the puritv of the material ,iM(i its quantitative rt'iovcry after use of tne samples.

1. 1. !.avers prepared by vacuum evaporation:

I . -1. 1 . Horon layers I'*''

i'lOron i s an important rcferenee substance in neutron n i easu r cmenl s. Trie cross section ~i'> In,'""') ami branching ratio nave bei-n investigated bv maiiv autnors. About 12 years ago the CB.W1 started with the fal) rii atinn of standard layers of natural Moron by deposition by cie-.tron bombardment and weighing under vacuum.

Tnc Livers requested and used nad to be deposited on Au-i oaled cuartz disks. The Pioron layers nad a diamete r.-, of 3K mm and tne thickness r.HU'i1 of B-deposits wjs: -10 - 1 ^0 nc/em". The maximum tniekness \ ariation of tne layer is 1 "<• and tne number of B atoms per en;*" is known u-itn an accuracy of 0. 5"<-.

I.-L2. Vfetailic Uranium layers (10) (|p (] 2'•

The induction levitation melting technic|ue nas been usn] to acnieve evaporation of metallic Uranium in UH-vaouum.

The L'-films {up to 300 ug/cm on metallic or quartz backings) nave total masses of a few nig determined by weigning in UH-vacuum witn an accuracy of 0. ] "'••. Tne main isotope amount is defined with an accuracy of 0. ^"o (uncertainties of all defining techniques adcied up).

1 . i. 3. l.iK-layers (1 3^

Li-K layers up to IS mg/cm are being prepared by vacuum evaporation from a Pt-boat source. r'xperirnents at CR.N'M have shown that the Li-content of the layers is practically equal to the one of the starting material and tnat it can be determined to within _+ 0.2"'".

A possible change of the isotopic composition of natural Li during evaporation might be a problem.

i . -!. 4. VV foils. .23 5 i or precise measurement of tne U fission cross section, U F layers have been used, prepared by evaporation of L"F from a Ta-boat on vyns. The amount of material on these layers is small (S 2 rag) and yc-t nas to be accurately defined with respect to total Uranium content and isotopic com position.

1.4. ri. Layers for oscillation measurements

f'ne ('H*VM regularly receives requests for well-defined metallic layers for pile oscillation measurements. Typical examples are given in Table II for layers of Au, In, Co, B

1 . 4. 6. Highly enriched stable isotopes. for the preparation of metallic layers of highly enriched stable isotopes by evaporation under vacuum, a series of different vapour mini-sources are in use at CBNVf.

The evaporant is heated by resistance neating; electron bombardment - or HF induction, - heating of a tubular crucible allowing maximum colli- mation of vapours.

About 1 S"-» of the total evaporated amount can generally be collected in form of a layer with diameter of 20 mm and a homogeneity of about 6%. 7

10- 9

Fig. 3 : Mini-3ource evaporation arrangement.

Ta-cruci'ole containing the 6. Ma_c evaporant (and the reductant) 7. Sustrate .;u^pended on Ta-f(lament of the electron gun n^icrobalance V - rod S. Ta-ribbon oven (degazing and heating of the substrate Crucible position can be changed from outside q Thermo couple -.""eramic isolation 10. HT - power supply and filament heating Table II : Layers Prepared for File Oscillation Measurements

Element Au Au In In Co B B

Substrate rnaterial Quartz Al Al Quartz Al Al Al

Form of backing Cylinders Plates Plates Cylinders Plates Cylinde rs Disks

Thickness 0.25 - 5.0 0. 1 - 1. 0 0. 1 -3.0 0. 05 - 1.0 1.5 - 7. 5 0.2-2.0 0. 2 - 0. 4 range (Aim)

Accuracy of mass 5.5 - 0. 3 5.0 - 0. 2 6. 0 - 0.2 1.0 - 0. 1 1. 0 - 0. 1 5. 0 - 0. 5 0. 4 - 0.2 determination (°i)

Max. thickness 3. 5 3 3 7 5 9 2 variations {%)

Evaporation Resistance he;.ting Electron Electron bombardment technique Mo - filaments bomba rd - Single source 5 evaporation sources ment Distance : 500 mm 3 sources

Dimensions Plates : 25 x 100 x 1 mm Cylinders : 0 25; length 100 mm; wall thickness 2 mm Disks : 0 30; thickness 0. 5 mm 128

I'hi- major part of iKiii-usciblc evaporated m.tlfrijl cm easily lit' stripped off from .i judiciouslv plaicil water cuoli'd i-oppcr condenser. Keiiuiti.ins of oxides can also he carried out directly in the mini - sou rce. A tvpic.il mini - sou rce ,t r r.iii gemen t i -S shown in Fig. '•.

Tnis technique h,is been applied at (". hN\[ for " MR, " Mf;. ' Mg (), ' '-Sm i ' , ' '"'Sm (> - •' .1 > I. ^. Layers prepared by triodi' sputtering;. 2 I'riode sputtering is used for trie- preparation of thick (up to Z mg cm*"! I .(, W, I't. Ti layers on glass, 'iii.irlz, carbon and metallic sub- - ! rales.

! . i'. Solutions

Well defined solutions had to lie prepared for in-ulroii measurements, .iiul for i aliljratinjj mass spectrometers (mixtures of enricned and de- pleted i sot ope s . Some typical solutions whicn nave been prepared at ("BNM are shown in Table III and 111 ' 1 . In most of the above cases no standards are available and therefore high purity .substances were dissolved and ihe euantitv 'f clement present determined by several independent methods.

1.7. Melting and '"lasting (14'

The metallurgy laboratory developed preparation techniques and definition methods in response to numerous requests for alloys witn very specific ,in<} accurately known composition for: iccir.ite neutron flux measurements intended to be used in form of small disks or wires, - ihe determination of "cross sections" by oscillation measurements in form of disks, tubes or rods, - reference alloys for analytical chemistry and solid state physics.

In most cases Ine user wants to know, to within 1 - 0. S% the quantity of alloying element in an amount of material which in some cases is as small as 20 mg.

The levitation melting technique has proved to be successful for fast(and qu.inlitative) preparation of pure rind homogeneous alloys having a well defined composition.

This technique is extremely interesting for standard reference materials because it allows in many cases additive preparation of very well known alloys and has therefore be developed and refined for that application.

The principle of levitalion can be summarized as follows: - the high frequency field in an induction coil produces eddy currents in a piece of metal placed in this coil. These currents set up a magnetic field and as a result of the interaction between this field and the HF field of the induction coil the metal begins to "float". If tne top turn of the induction coil is wound in a direction opposite to the other coils, a magnetic field with a zone of lesser intensity is set up inside this coil at its centre; it is in this portion of the intern space that the piece of metal tends to become centred.

1'igure •! shows a drop of molten metal in an induction coil. Fig. 5 shows one of our levitation melting facilities installed in a glove-box and user! for tne preparation of Al-Pu and U-Pu alloys. Table III : Solutions in quartz containers

Concentration Definition Isotopes Chemical form Solvent (mg • cm ) Main isotope Main element (+ <) (± "•")

6 Li hydroxyde or H2O or DO 15 < 0. 1 0. I sulfate

7 . t 1 J-,i 15 £ 0. i 0. 1

nat. . Li " " !5 2 (on Li) 0. 1

I0 1 ' B H3BO3 or 4 S 0. 1 0. 1

Na2B4O?

1] 11 B '• 4 •-• o. i 0. 1

natB 11 1O '• 4 0. Pon IV) 0. 1

U any enrichment Nitrate -I NO, 2 00 • o. i 0. 1 ofu"5 Table 111,'I : Solutions in quarts containers

Element or isotope Chemical form Solvent Concentration Requested definition mg element 0' g solution 60 In 1»4 (S°4>5 D2° 0. 3

Au KCN Complex H2O 0. 5 - 15 0. 2

Cd CdSO4 D2O 20 0.3

Co Sulphate 4 - 40 0. 5 o H2SO4 Hf HfF, Acid 0. 5 - 8 0. 5 4

In In2(SO4)3 Acid 0.3- 3 0. 2

Ag Ag NO3 Acid 2 -20 0.2 H Cs Cs2CO3 z° 4 - 50 0. 2

Th Th(NO3)4 Acid 20 - 100 0. 5

Au + In Complexed 20 5 - 0. Zb 0. 062S Au : 0. 3 In :0.2 . 131

it;- "1 ; l^rop of molten metal in an induction coil.

Fig. 5 : Levitation facility for the preparation of Al - Pu on U - Pu alloys. 132

The .sd vantages of levitation melting ari' numerous !. The ni.-iti-ri.il is molted wilnout physical contact so that there is.no risk of contamination from the crucible. .'.. The intensive stirring to which the drop of molten metal is subjected, mainly under the action of the magnetic field and the rapid solidification after casting of small amounts (up to I SO g) into a cold mould, puaran- tee 1 good homogeneity cf tne alloy. >. The process is quantitative and reproducible. The losses that occur during the preparation of a lar^e series of alloys are so small '.hat ihe composition differs by less than ] "'•• from the one which is calculated from the masses of the components before alloying. •1. Samples can be made quickly. 133

2.. Sample Definition

To perform absolute measurements of physical constants in general and nuclear parameters in particular, the sample quality and definition are of rssenlial importance. A precise and accurate nuclear measurement requires a precise and accurate knowledge of the number of atoms of the nuclids to be measured and of the impurities present in the sample. 2.. 1. Classical Approach One currently used approach in sample definition for such a purpose hence is a) to determine the rna_ss_of the sample (this requires metrology) i>) to determine the number of atoms of the element concerned in the given mass of sample (this requires analytical chemistry) ( ) to determine the number of atoms of the nuclide concerned in the element (this requires mass spectrometry of the element used) (}) to determine the impurities present in the sample especially the neutron poisons (this requires analytical chemistry)

Table IV : Definition and characterizing niethods

Weighing : - conventional - weighing in ultra-high vacuum - weighing in pressure controlled glove boxes (under development) Quart/, oscillator thickness measurements X - ray absorption and fluorescence P h < > t o - d e n s i t o m e t r y ' Stylus methods i nterl'erometry Rutherford scattering Density measurements Analytical chemistry Classical wet chemistry - gravimetry - titrimetry Klectro-analytical methods - coulometry - polarography (including differential and pulse polarography) Hmission spectrographic analysis Atomic absorption Activation analysis Mass spectrometry Isotope analysis in solids, liquid and gaseous samples of stable and fissile isotopes. Quantitative assay by isotope dilution of concentrations ranging from 100 % to 1 ppm and of amounts down to 1 mg with accuracies ofS0.3 %. Others a - P - y - counting and spectrometry techniques Metallography Electron microscopy 134

Many samples have been character i zed by this classical approach which will continue to be used extensively (15)(!b). A lew of the more interesting sample character i/.ai ions might be given as illu st ration. Z. 1 . 1 Boron standard layers (8) A complete clelinition ot B-layers prepared by vacuum evaporation is shown in Table V .

Table V : K ementa! Boron layers

Layer \\7 >, Ii71 15 64 13 69 • V ncertainty

Total mass (tvl) depositor and determined under va- 412. o 525. 0 787. 2 49?. 8 2:.« cuum (u g) Metallic impurities(ppm) 2215 2171 42 64 1431 20 ppm Carbon content (ppm) 1080 1450 1 100 2500 20 ppm Total mass of boron deposited Mj>; (^ g) 411.2 523. lc 78 3. 3 492. 1 2 '-R 4 2 Calculated surface area 1 1 . 357^ 11.3597 11.4 54 11.453 0. 005 cm Surface density of boron 2 Z (ug./cm ) MB = MB/O 3 6.21 4 6. o: 68. 39 U. 97 0. 20 ng/cm Isotopic content 'OJJ (At. %) 20.448 20.330 20.127 20.230 0.04 at f, Surface density of 13 6.857 8. 670 12. 75 8.050 0. 04 ijg/cm tig/cm2)

Total uncertainty

10 Total Mass of B(pg) 77.87, 98.484 146. 01 92. 19 0.4-0.5% 0 A nice confirmation of the absolute characterization is given by the perfect linearity of the function mass '"B vs. counts taken from the neutron absorption measurements. See Fig. 6. The performances of the U.H. vacuum balance developed at CBNM (23), (24), (25) to determine the mass of these boron standards (as well as other standard layers : e. g. U) have recently been improved.

Previous performances Present performances

Max. load 3 g 10 g

Accuracy 1 a(n=10) ± ' Mg + 3 ug Backing temperature 250° C 250° C

Table VI : Performances of U.H . vacuum balances 5x10'

Fig. i> Calibration i-urve for the tour boron standard layers 2.1.2. Uranium layers (8), (12) Characterization of evaporated metallic vj -layers by three independant essay methods yields nicely coinciding elemental assays as shown in Table VII. Table VII: Assay of elemental U-layers in mfi

Sample 1 sotope dilution Weighing C. P. coulometry

128. 1 1 1 . 82 11.81 11.84 12H. 2 1 6. 25 1 6. 54 16. 34 1 M.2 12. 55 12. 57 12. 59 1 55 15.05 13.02 1 3. 09

ln another series, U-layers ranging from 1 . 7 to 5 mg were counted by : - Low Geometry and specific activities determined with the results of a) mass determination by weighing b) elemental assay by isotope dilution mass spectrometry. Results are summarized in Table Vill for U - layers prepared for the determination of Z34U - T 1/2 (17) (18) (2Q). Table VIII: Mass determination on uranium layers

Sample Total Mass Low geometry counting rate(dps) deposited (jj g) Total mass (mg)

Weighing Isotope dilution A%

1 2777.2+0.4 1885.34 1890.79 - 0. 3 -i 3547.7+0.5 1886. 78 1882.91 + 0.2 3 2743. 8^0.4 1885. 02 1889.70 - 0.25

5 2785. 1+0.5 1880.30 6 5001.5+0.7 1818.50 7 2460.8+0.8 1885. 59 8 3599. U0.5 1JG

Very rt'Li'iit work ha^ been d(.)in' on a now dete m i inat ;on oi the ~ ' JL- hall- life at CHiVM and somr of the .sample definitions for that program are showing mo.it recent improvrmi-nts in ai'curai y ol dei inilion.

Table IX : Isotopic A11 a I y s i ^

Atom ". A c c u r'a c \'

2', •; • 0 . (!0.i() 2 34 0. 00 1 I) 2 55) I' 0. DO^i 2 it o. o i i •; o. no.'.i)

Table X : " {' .Sample l)i';'iii:ti.in

isotope dilution assay Constant potential coulometry (including total uncertainty)

1 . 934 3 i_ 0. 0029 (0.15 " 4 . 40 57 _+ 0. 00 6 6 (0. 1 5 ". 1' * 1 T - 4 10.94 5 i 0.017 (0.15" 10. 924 _f 0. 027 (0. 25 ".'.,) 14. 91 3 +0.022 (0.15' 14.88 5 + 0. 0i7 (0.2 5 "',)

24 i 2.1.?. Pu Co samples (tor neutron cross section measurements) (18) Canned samples of several g •-*' ' PuO^ (rented from USAKC) have been defined to a total uncertainty oi 0. 5 •% on Ihe " Pti content. A number of the characteristics are given to illustrate that definition in Tables XI, XII,XIII. Table XI :_ Characteristics of the requested and prepared sample s

'Type of sample A era Retjuested Sample sample thickness thicknes s cm ") (nig Pn cm"") (mg Pu cm"'

a Tl T2 T 3 c fl n il Table XII : Isotopic Analysis

Isotopic composition Atom Accuracv 238 Pu 0. 009 2 39 Pu 0. 8 68 240 Pu 2. 968 241 Pu 94.6 58 + 0.100 24, Pu 1.497 137

Table XIII : Impurities determination by spectro^raphic analysis

Element Material as received After Am separation (ppm) (ppm)

Na 250 < 30 Be < 11 < 20 Mg 250 1 5 Ca 300 < 250 Ba <18 < 10 l>> 24 5 Al 68 < 25 In <23 < 10 Si 750 720 Ge < 7 < 6 Sn <23 < 10 Pb 250 15 Bi < 7 < 6 Ti < 50 < 20 V <23 < 20 Cr 40 15 Mn 7 3 Fe 64 5 75 Co <20 < 10

Ni 68 25 Mo <23 < 20 Cu 62 5 Ag 3 < 1 Zn 250 50 Cd 23 5 138

2.2. Synthetic preparation as a definition procedure ;

Tin? procedure described under 2. ] is an a posteriori and mostly destructive definition approach. A fundamentally different way of tackling a definition problem is the onv which uses synthetic preparation methods ( mixing , dissolving, melting ) of samples by combining <|uantitatively and qualitatively known materials. This approach has the advantages of a) high selectivity : one selects on beforehand (he components required and only those li) techniques of synthesizing ( e.g. mixing, dissolving ) do not offer major or fundamental difficulties c) weighing ol components to be synthesized gives high accuracies of compo- nent concentrations in the synthetic samples, provided losses can essen- liallv be kept at zero d) highly pure and well known materials can be used in the syntt.esis.

2.2.1 . Synthetic blends of elements 2.2.1. 1 . Liquids Solutions of dissolved elements or their compounds can rather easily be prepared from pure materials and then belended into mixtures which have an accurately known composition from the gravimetric values of the synthesizing process-in this case mixing of solutions . The homogeneity in the dissolved state usually does not present ; in1 difficulty.

Example : Li? SO in D.,O for Li standard cross section measurement .

Table XJV: Li SO in D9O

Isotopic Analysis Li Element concentration

' Li 99. OS 0 10 (21 . 42S - 0. 030) me Li,' — ' g sol ' I.i 0. 9S i 0 10

2. 2.1.2. Solids

The procedure however is not limited to the mixing ol solutions . HF-levitation mixing or alloying of molten metals without noticeable loss and crucible free during the entire process , also provides a way of characterization by synthesis . A number of examples are given in Table XV . 139

Table XV : Certified composition by quantitative alloying

1 1 Alloys Isotopic Composition

1 A Atom a,'o Ace.

"" Pu 0.008 Al - (7. 50 J 0. 07) wt "'„ Pu 9Pu 94. 500 J_ 0. 050 40Pu 5.24 8 Al - (10.0 + 0. !) wt % Pu 241 Pa 0.2 30 24ZPu 0.014

234U 0.783 .1 - 21 . 50 + 0. 20) wt % U 3DU 93.192 - 0.020 Al - (1 6. 14 +_ 0. 1 5) wt f U 0 2j6U 0 227 Al - (10. 7 + 0.1) wt % U 238U 5.798

2.2.2. Synthetic I.'ends of isotopes Along exactly the same lines as for elements, synthetic blends of isotopes can be made in order to arrive at samples with highly accurate isotopic compositions. To achieve this, one carefully defines amounts of separated isotopes and blends them. 2.2.2.1. Liquids This technique has been applied in recent years to prepare calibration samples for mass spectrometr/. For the Boron case (stable isotopes 10 and 11) that approach was chosen and characteristics are given hereafter in Table XVI. (26)

Table XVI : Isotopic and chemical definition of the enriched isotopic solutions

ll '"H U atom ratio KA 28.571 • 0.024 Ktt 0.01 72S1 : U.0tXK>20 {n 41 (II 4)

'"U atom eonc. (96.6183 j 0.0027)% ( I.MM : 0.00201% :MS atom cone. 1 3.3817.; 0.0027)",, OS.Mil : 0.0020)",,

(' (ctnilombs/g) (\ 35.702 ; 0.011 CB • • 29.869 L 0.009 (II 19) (u - 11) (\'f'n 1.19529 ! 0.00052

.Sine: uncertainties given arc standard deviations for a single measurement.

As Table XVII shows, it has been possible to prepare highly defined samples (£0.05 %) at any desired isotopic composition. 140

Table XVII : Observed and computed mass spectrometric ^/"B ratios of the synthetic blends

Mean of CompuliJ K ••- KcfX

•••II "It (It) "•!>"U IRr)

O.I 11X7 (10) O.I I 1X6 ( 7 I 0.11185 o.llisi ( si 0.249.V (ID)

0.2500(1 I 71

(1. :•!'/< l.i Ulli (101 ' MSSO

0 2455'' ( SI (1 24'Ki< (10) o.;4?7i 0.9994 24571 ( 9)

0 .544 57 (101 1). 544 l.i I 7) 0.54457 O.54:S3 0.544X1) ( 7l

IV''2724 ( l»l '12 71 s (III O.S2I-70 927'IS ( Si

1.SSI.i (II) SS2S 0.9995 | (in 3.9120 ( 9) ;..9099 (101 .1.9113 3.9095 0.9995 •v ,VI:I (IM 4 .05 IS (III 4 .11591 110) 4.0555 0.9990 •1.05,15 (111) 4.0656 (KM 4.0644 1 ('1 4.0651 4.059ft 0.9986 4 .01-54 ( X)

i; .0576 (10) 1)0) 9.0555 9.0J90 0.9984 'I.O4ft3 110) Mean: 0.99892 !: 0.00060 (sland. dev)

2.2.2.2. Solids The procedure however is not limited to the mixing of solutions. HF- levitation mixing of molten U metal components of different isotopic composition, and parallel definition of the mixture's isotopic content by mass spectrometry, also proved successfull. Thus 5 % enriched 235y samples were synthesized from naturel U and 93 % enriched ^-^U metals. 141

235 Table XVIII : Synthetic 5 U samples

Weight 235U Sample HF levitation Mass spectrome- (1) - (2) synthesis (1) tric definition(2)

I 5. 00 + 0. 02 4.989 + 0.025 o.2 °; II 4.99 + 0. 02 •i. 990 + 0. 025 o. o % 111 5.00 + 0. 02 4.997 + 0. 025 0.0 °'o

2.3. Work in progress In order to achieve still higher accuracy in the definition of samples which is needed for partical measurements (standard cross sections, half-life determination) as well as for the rapidly growing need for Reference Mate- rials, we are currently trying out new refinements of existing techniques, taking new approaches to old problems or apply techniques to sample definition which previously had not been used there. Some of these will now be indicated and possible preliminary results given.

2.3. 1 . Standard reference layers A basic difficulty in the definition of layers (evaporated or otherwise) is that one has to perform destructive assay after the nuclear measurement because the high selectivity of some destructive techniques (chemistry or isotope dilution) have higher intrinsic accuracy than physical characterization (e. g. weighing). So the layer is destroyed and does not fit a basic requirement of aReference : being stable and re-usable at all times. To circumvent this we would like to propose following approach (Fig. 7). A witness layer of evaporated material on a ring around the target is collected during the evaporation process. The difference in collected material could be determined by weighing both Ring and Target in UH-vacuum before and after the evaporation process (this difference can be arranged to be very small). Afterwards the Ring can be quantitatively assayed by isotope dilution, whereby highly precise and accurate assay of the Ring material is performed. A small correction for the mass difference between Ring and Target then allows to calculate the target layer very accurately. This method is likely to be suitable for mg and microgram layers where even the accuracy of the weighing (2:_ig) becomes a limiting factor in the definition whereas the isotope dilution accuracy essentially remains constant (0. 15 %).

2.3.2. Standard reference alloys We have been combining in the past two definition methods of known accuracy in order to make sample characterization absolute to a high degree of certainty : Al/Eu alloys were prepared by HF-levitation and defined both by the gravimetry of the preparation process as by isotope dilution. Coincidence of results was gratifying as shown in Table XIX. (27) M2 mg U on Witness Ring M] mg U on Target • WS WS 1 mg U Layer Evaporation of U

M2 Mi determined by weighing in UH. Vacuum

UNKNOWN ENRICHED SAMPLE = ISOTOPE 2333,,. M2 mg metallic U 3°n8 U layer on Witness Ring

SOLUTION UNKNOWN OF SOLUTION

Fig.7 PROPOSED NON - DESTRUCTIVE DEFINITION OF URANIUM REFERENCE LAYERS BY DOUBLE MASS SPECTROMETRIC ISOTOPE DILUTION AND WEIGHING (Expected Accuracy =0.157.) 143

Table XIX : Definition of Al Eu alloys

I-Jv c jntent (ppm) Difference after correc- 'I) (2) Diffe r e n c e tion for } . b°/o Sample Quantitative Mass Spectro- impurities in ftcparat ion metric 1 sotope (1) - (2) the Eu,determi- Di hit ion ned from sample J\.

A I 00004 100(1 "»:) 9840 _!_! 50(1 .5°-) 1 . 6 -;,, 0.0 "i°

15 1 9. 70-! 0. 20(1 19. 48-1 0.30(1. 5%) 1 .I "•„ -0.5% C 2. 00+_0. 02(1"A) 1 . 965H 0.040(2 "/<,) 1 . 8 ° <> - 0.15 °'o

Recently we achieved the definition of a U/Pu alloy in the same way :

Table XX : Definition of an Li / Pu alloy

Pu content Difference Quantitative Preparation Mass SpectromeT.ric „. , Isotope Dilution (1) - (2) ^C1«ht "] (weight %)

5. 00 j_ 0. 0? 4. 98 -_ 0. 02 0.40 % X

u c ontent

. K <) 5• ° ° 1 0. 50 9 4. 56 j_0. 40 0 46 °

Pu / U ratio

0 0 52 63 0 05266 0 0 6 % KX x Difference due to about 0. 5 °» impurities in the Uranium. These interfere with the ;;ravimetry and hence the quantitative preparation and not with the mass spectrometric isotope assay. Consequently this difference disappears when comparing the Pu/U ratio per method xx. 2.4.3. Thickness measurement Several physical methods for deterinining the Ihickness of thin layers on fla! .surfaces are used at CI3NM and have been discussed in previous papers (8) (17) (18) (19). Out of our work in progress only one example will be treated in somewhat more detail s. 144

3.3.3.1. Stylus Method The possibilities and limitations to make non destructive thickness measurements by the Stylus method has recently be evaluated when applied to evaporated films oi soft metali (20). The Stylus instrument used was a Talystep I (Rank, Taylor and Hobson Ltd. ) with a Up radius of 13 ;j; the stylus force was adjustable between 1 and 3 0 rag. The dependence of the dept of tracks on the stylus loading was measured for evaporated films of Bi, Al and Au . Tin* results oi some of the measurements are given in Fig. 8. The tracks drawn with the Talystep were inspected in a scanning electron microscope. The dept of the tracks either was measured with the Talystep itself or was obtained by measuring the width of the tracks on a profile projector. As a conclusion it can. be stated that in many cases the plastic deformation 'ntroduced by the stylus can be kept small. Even for rather soft metals films such as Al or Au the error to be expected at low stylus forces (1 - 2 nig) will be less than the resolution clouned. This is clearly shown for Au-layers by the insert in Fig. 8.

/ * Au

a 500-

•.00-

300' / /

100'

£0 60 100 \?.O KO 160 jmgj

8 : The dependence of the dept of tracks on stylus loading for Hi-, .Al- and A)i - films. -. '<. -1 . Physical behaviour of samples A final word now on stability characteristics. As almost all samples have to be used in a non destructive measurement, yet are subjected to transport, mounting, temperature differences etc. for various periods of time stability with respect to all defined parameters as well as physical behaviour is very important. Certain minimum requil ements are generally mutually agreed upon by the manufacturer and the user in order to ensure that the sample renders a useful service under tin- measurement conditions. The behaviour of the samples can be influenced by the preparation procedure. It •• s for example a well-known phenomenon that thick layers prepared by vacuum deposition tend to crack and peel of the substrate. This apears to be associated to mechanical stresses developed during the preparation of the layer. In order to get information concerning the influence of preparation parameters onto the mechanical stresses in the layers an apparatus has been constructed and the results for gold layers have been compared v. ith those from similar experiments as reported in literature.

The mechanical stresses due to a thin layer deposited on one side of a glass- or Ta strip- substrate give raise to a deflection of the substrate. The lenght of this deflection is an indication for mechanical stress in the 1 a y <:• r. Use of linear (voltage) differential transformer is made to perform the necessary high precision deflection lenght measurements (21). Fig. 9 represents some preliminary result- for Au-layers on the same glass substrate and on two different Ta - substrate s. From the scatter of the data points it can be concluded that reproducibility can only be obtained by using the same substrate.

Fig. 9 : Mechanical stresses in Au-layers on glass-(G) and Ta-substrates as a function of layer thickness. 1-46

2. 5. 5. Determination of traces of neutron poisons Because of their high neutron cross - section the elements of the rare earth group are of considerable interest in nuclear science. Their accurate determination is a problem that interests scientists doing nuclear measurements or using doped scintillation crystals. Isotopic dilution mass spectro- mdry and polarographic methods are used for the determination of europium and other rare earths (27).

A study oi the polarographic determination of europium was made, so as to allow a clear evaluation of the analytical possibilities of the method (30). One of the objectives was the determination of small quantities of europium in lithium halide monocry stals, and in view of obtaining as high an absolute europium concentration as possible, it was decided to work with a high electrolyte concentration. Differential polarographic determination of the europium content of solutions of Lil and LiCl can be made within the concentration ranges oi 50 to 500 L g Eu'u with varying precision. Precisions of the order of 0. 1 °~o or even better have been reached. 147

Acknowledgements

The authors are indebted to many of their colleages who contribute considerably to the sample preparation and definition work at C . B . N. M 148

K e i tj r e n c e s

(1) Lauer, K.F. . Verdingh, V. Preparation by Electrosprayinn of Thin Uranium, Plutonium and Boron Samples for Neutron Cross Section Measurements i?i 4 — - Geometry Nucl. Jnstr. and Methods 2J_, 161 (1963) (2) Verdingh, V., Lauer, K.F. Equipment for F.lectrospraying \ucl. Instr. and Methods 3j_. 555(1964) l.i) Verdingh, V. The Preparation of Samples by Electrochemical Me'.hods Proc. Seminar on Preparation and Standardization of Isotopic Targets and Foils, AERE - R 5097 (1965) (-1) Verdingh, V., Lauer, K.F. Equipment for Electrospraying Nucl. lnsr and Methods 9, 197(1967) (^) Verdingh, v. The Preparation of Layers by Eiectrospraying and Electrophoresis Third Symposium on Research Materials for Nuclear Measurements (Gatlinburg - Tenn.) p. 1 60 - 165, CONF - 711002 (1971) (6) Verdingh, V. , Lauer, K.F.

The Preparation of Uranium Layers b; High Voltage Electrophore sis Nucl. Instr. and Method s _o0, 125(1968) (7) Verdingh, V. The Preparation of Powder Samples Third Symposium on Research Materials for Nuclear Measurements (Gatlinburg - Tenn.) p. 77 - 81, CONF - 711002 (8) Eschbach, H. L. Preparation of Standard Layers by Vacuum Evaporation Third Symposium, on Researh Materials for Nuclear Measurements (Gatlinburg - Tenn. ) p. 122 - 133, CONF - 711002 (1971) (9) Van Audenhove, J. , Eschbach, H.L. and Moret, H. Deposition by Electron Bombardment and Weighing under Vacuum of Thin High Purity Boron Layers Nuel. Instr. and Methods 24, 465(1963) (10) Van Audenhove, J. Vacuum Evaporation of Metals by High-Frequency Levitation Heating Rev. Sci. Instr. 3_6_. 383 (1965) (11) Van Audenhove, J. . Joyeux, J. , Parengh, M. Evaporation of Metals and Semi-Conductor s in Ultra High Vacuum by Induction Heating Supplement toN". 136. Le Vide, p. 69 - 74 (1969) (12) Muschenborn, G. Die Herstellung metallischer U ranium-Filme durch UHV-Aufdamp- fung fur Kernmessungen Vakuum Technik. 20, Hft. 7, p. 197(1971) 149

ill) I,f Duigou, Y., Lauer, K.F. Lithium Content Determination in Lithium Fluoride before and after Vacuum Evaporation Nucl. Instr. and Methods 97, 199 (1971) i!4) Van Audenhove, J. , Joyeux, .1. Sample Preparation by Metallurgical Methods Nucl. Instr. and Methods _l_02, 409 - 415 (1972) (1 5) l.auer , K.F. Assay of Samples for Nuclear Measurements by counting and chemical Analysis Xucl. Instr. and Methods 102, 589 (1972)

(16) Moret, H , Verheyen, F. Assay of Samples for nuclear Measurements by Physical Methods N'ucl. Instr. and Methods J_02_, 57S - 580 (1972) (17) De Biovre, P., Lauer, K.F., Le Duigou, Y., Moret, H., Mtlschenborn, G. , Spaepen, J. , Spernol, A, Vaninbroukx, R. ,

Verdingh, V. o The Half Life of '" U, KANDC Symposium on Neutron Standards and Flux Normalization, Argonne National Laboratory, USA, October 1970 (18) Pel Bino, G. , Lauer, K. F. and Verdingh, V. The Preparation of "^'Plutonium oxide samples for neutron cros.. - section Measurements Nucl. Instr. and Methods 9_3, 205 - 209 (1971) (19) Trapani, A., Berlin, A., Eschbach, H.L , Paulsen, A.. Verheyen, F Compt. Rend. Trav. Congr. Intern, des Couches Minces, (Soc. Franc. lng('n. et Techn. du Vide, Paris) Cannes Oct. 1970 (20) Verheyen, F., Dobma.W., Eschbach, H.L. Construction and Application of a simple X Y Scanning Isodensitometer Journal of Physics E : Scient. Instr. , Vol. 4, 435 - 437 (1971) (21) Eschbach, H.L. and Verheyen, F. Possibilities and Limitations of the Stylus Method for Thin Film Thickness Measurements Thin Solid Films 21_, 237 - 243 (1974) (22) Eschbach, H.L., Lycke, W., Verheyen, F. Use of a Differential Transformer for the Measurement of Mechanical Stresses in Thin Evaporated Films. Vakuumtechnik 22, Hft. 8, 233 - 238 (1973) (2 5) Moret, K.and Louwerix, E. Microbalance for Ultrahigh-Vacuum Applications Vacuum Microbalance Techniques 5, 59 (1966) 150

[,l-\) Moret, 11.. Lomu'rix, E., S ttler, E. Comments on tin.-Applications and improv) U<>) De Bfc-vre, P. , Debus, G. 11. Absolute Isotope Ratio Determination of a Natural Boron Standard .] . Mass Spectrometry and Ion Physics 2, IS - 21 (196'.!) U7) De Bievre, P. , Del Brno, G. The Accurate Determination of Trace amounts of Europium in Aluminium Anal. Chim. SO, S26 - S50 il'»70) (i8) De Bievre, P., Lauer. K.F., LeDuigou, Y , Moret, II. Muschenborn, G. , Spaepen, .1. , Spernol, A. , Vaninbroukx, li. , V e r d i n g h, V . The Half Life of 234U British Nuclear Energy Society Meeting, Canterbury. Session 4. Paper 28, September 197! (29) Verdingh, V., Lauer. K.F. Accurate Determination of Microgram quantities of Uranium by direct and differential Polarography '/.'. Anal. Chem. 2J_5, 3-11 (1968) (30) Verdingh, V., Lauer, K.F. Studies of the precise differential Polarographic determination of Elements of Nuclear interest. Part f : Determination of Europium and Ytterbium in Lithium Chloride and Lithium Iodide Electrolytes Anal. Chim. Acta 3_3_. -169 (I 96S) f i I ) van der Eijk, W., Oldenhof, W. andZehner, W. Preparation of Thin Sources, A review Nucl. Instr. and Methods JJ_2, 34 3 - 351 (1973) (3 2) van der Eijk, W. Source Improvements in the Preparation of Thin Radioactive Sources Ph. D Thesis to be submitted at the University of Amsterdam 151

"I":i.- Hole .-'f Substrates in the Growth of Self-Supporting Thin Films

D. Ramsay

;•..'.: .iru-vnt of Physics, Stanford University, Stanford, California '34305 U.S.A.

Tin • purpose of this report is to give the specific parting agents used in •_::••

;. :vp.ir.il: i.ni of sc 1 f-;-;upporting targets of some 40 elements. I would dlso 1 ik'.• lo

iv.'.\ :ie general ;:onc] usions that could be used as guidelines .and -.vhere ;. •oss: :.]• •

• :iv" tin' Li.ita tliat led to these generalizations.

With .".ill other conditions the same in a vacuum deposition, the choice 'if the

sub:, t rate •..•an makJ the difference between depositing a uniform cohesive film <•>>-

•r-niiin:-; i !i"; no film at. all. For example, when vanadium is evaporated onto a glass

..;lui-' •.-rated wit . a 2500 angstrom (80 iJgms/cm ) layer of potassium iodide, an ov-r.

(•(•ntir.uoii.i film is formed, which can easily be floated off in water and is self-

supporting at 500 angstroms (25 ;jgms/cin ). However, if barium iodide is the? sub-

strate -outing layer, then the deposition is erratic and will not form a continuous

film even if the charge of vanadium is increased by a factor of six.

The crystal structure and lattice constant of the material, which is to be

deposited, must be considered when choosing a substrate.

The formation process of thin films (Fig. 1) begins with the arrival of a

single vapor molecule at the substrate (a). If it condenses then it can either niigr.it o across the surface or re-evaporate (b). In migrating, collision can occur

,i:i:l .-ombi nation (c). With the subsequent loss of energy, stable islands are

f'-n'H'.J ,\t preferred growth points or adsorption sites, in a process termed

iij." 1 • ^jtion (d). With the impinging of more molecules, growth occurs (e) until the a.M-Lis start to coalesce (f). This is an oversimplification of the process, but it

•.-.•ij] serve to illustrate how the substrate directly affects the formation of a

: ; Jin. The condensing molecule must have an adequate number of adsorption sites

(!•'!!!. '?) and these sites must be spaced at an interval which will allow coalesce:,^1 and inhibit re-evaporation. In epitaxial film growth, the matching of the lattice spacing:-; is critical. In pojycrystalline thin films the lattice constant nay not

)..•..- critical but it is the determing factor for both the number of adsorption sites

.uid the distance between discrete sites. The condensing coefficient and adsorption fiiomy of tiie particular compound are also contributing factors.

For instance cadmium films will grow at room temperature on a substrate of

;^inc chloride." Both are hexagonal in structure and have lattice spacinqs of 2.08 and 'J.S2 angstroms, respectively.

The critical area for causing stress and creating crystal defects is at the sul strat.e/f ilm interface. Cobalt, which in the bulk solid is normally hexagonal,

.:!i!"ortwi in part bv the National Science Foundation- 152

Fig. 2. Schematic of substrate surface.

can be induced to grow at room temperature in a cubic arrangement on a substrate with cubic crystal structure. Below 20 angstroms the lattice is strained to exactly match the substrate. Above 20 angstroms dislocations are generated to accommodate part of the difference between the cobalt and copper lattices. If a film like this were to be removed from its substrate it would disintegrate from internal stress. Ideally, in order to obtain a uniform film, substrates with the same crystalline structure as the depositing material should be used. Fortunately this is not always absolutely necessary. Many elements can be satisfactorily formed on amorphous parting agents such as teepol and fonwar. About one half of 4 the metallic elements are cubic and so are many of the water soluble salts. For some of the hexagonal elements like titanium, yttrium and thulium, there are hexagonal salts, in this case calcium iodide. Cobalt, scandium, holmium and 154

!•; is' • :x\\r:v.. um • \ r o\\: wo 1 1 on a J uin I n um oxide, J v,':i i f •"; i is . i" Li.- J in/xaqoridi . Jn pracf: _e , -3 2

i_: !• ,i ; .raisiuiii KUbstr.it.o may be as thic.-i as 2. J * 10 cm (/ nqs/cm ) and is etneu

.- :" I' i y \i\ ;i:iq in s < ••

i • :...- n! y aluminum which has been expGS"d to air ] a:w onouqh for the oxide

cat :•.!•: !o i'yrn on the surface. In this method residua] contamination eould not

: i . :• i.-.it-. ii iy iookinq for q.-iimiia rays from r rot on-induced roacti ons. This C.;YCS

:n -:r:-i-r !i::i;t of :.<. r: ::• m-i/cm" to tlie aJumimi;:! conlent •:;•:' tJie tar(jdt.

:h"ie ai"" exf [ft i'.ir.'- to tji i J a;!proiii.-h. Altliouqii barium rhlorido i:; ort::n-

r : .• ii:: i i •, j i v,T(;rks we11 '! foi" bery] i i tiin w}]i';}i is hexagonal. .^ui -.'X^ la.nation may ;.-' ( in the '•,'•-]"/ snail ijiain si;.:e (dci-'idar structure) or the very iarqe lattici.- onn-

-.;-.u!t rjf iia) i urn chloride^ ('i.V> anqstronis) . Also r. u'.-ianc -;.».:-, which is "ubi.?, con-

'•• :!•••..: v.-r 1! us a roharorct fill-, on aluminun o:dd>', v/:iich i^; hexaqona!.

.\:v \.\:Y fac"Lor in tiie pr^i'.tration of sel f-s'lpi'jort inq films is the tem; -eratar'j

. ••!" 'h- ::.:;..; t rate . If the substrate temperature if; very low, say 76 Kelvin (liquid

:. i !.!"•;:.••..•!.) i oiid,vn ; a t: on is so fast and the afci jnobi ] i tis.'S so low, that they can-

not r.i':i thr; usual positions they occupy in tli> crystal lattice. The filn is

:'•.• r, iinplelely disordered or amorphous. If the temperature is sufficiently hiqn thin i:i'il.,i l ; |-y of t!;e atoms is hiqii and sinq Le crystal films may be qrov.'n. I-Y-r

; ;.":•.'••:•',': til] im> si• 1 f-supp'Ortinq films the p.articular temperature is not as irripor-

''. in t a..; ::iru n t.ai :i i ::q the sarifj temperature durinq t'ne qrov.'tl; of tlie- film. 7 ":" you I] row a i"Ton film on a teopol-ooatea1 qlass slide, suspended by itself,

•:..• fi 1 ni will, shatter around _!U00 anqstroms (50 ..qms/cm"). If it is grown on a

; •:. :,.:_.i., . .chlfride .aitoil q],oss slide; JOB ivin tjet to about 4000 angstroms (100

• :•••••./ :n"; . if i i i :•: qrown on a thin Jayor of boron on a glass slide which lias

: • • :. •. x; . i:-.-.-d to .

':,: •-.'.••.;:; of i.i. • 10 cm (4:lJi aqms/onD. Iri th i s case, what appears to bo an

::, ' ,,n •• "f therm.il strain is not entirely true.

;:. the coins./ of depositinq a film, an increase in the temperature of the

••.•:' ! it'' comes from tiie radiation

! .'..:: !•!•• i.ondensinq atoms. If we i\ilculate the heat rise due to the latter for a

.' ;•• :.'••!•'." :,i l'/rij- fiJm urown on .) qlass slide (Fiy. 5) we find it to lie 1.5°C.

(.: 7 M C T ac; acj a-.: 1 qs •; s g s _ ^ C •- 'A C 'final

a q a1j q yy q.s

ma s s i n cj r ams

i: - silver

s ;. e ••.• i f i <_• heat

t ois;.'r; r a t u re °C

•i 1 ass slide

Fiq. 3 155

,: ! " i ::i.i>: imum increase from 15°C to 16.5°C. It is possible then to qrov: a

i :-i •..-I'iicMt introducing any large temperature changes in the substrate due to

:,.!"iisi]iq atoms.

'.'ifTO are two approaches for avoiding temperature build-up in the substrata

. •:. ; ,1.: i .it ion. Tile Eirst is to minimize the amount of heat radiated from the

:: • MM!.ion iniurci:. If the time of evaporation is short, or the vaporization

•; "r.itur" low (•' 1200°C) often no special precautions have to be taken at alJ .

:..• i i i ::is liko bismuth and lead are simply grown with the parting agent (in this

'.•••.: i mi! iodide and potassium chloride) on a glass slide suspended by itself.

:: '.-.': M- v.ipori zation temperature is very high (> 1200°C) often the area of

•.r'••:!- radiation can be reduced and effectively shielded or water cooled.

••In:- ; i I:;:; for example, with source temperatures of > 2700°C have been grown 9 : i! •:. slide coated with sodium chloride. The electrostatically focused

• -i :' .i 'iun was water-cooled and the area of exposed radiation was only 1/8 of

'.):•:, in diameter. Since the amount charged was small (~ 30 mgs) for thick i •. •.••V'MviI evaporations were made, but no curling or shattering was found even

i • :.i V.IK'SS of 4!3OO angstroms (1 mg/cm ) .

•':. • s.'ct'iid approach is to compensate for the radiant heat by removing hoat

-:i -..•• subslrat" at a rate which will keep the temperature constant. What is

••I-:.mi. is that tiie initial condensation layer must not occur at a much lower

•:•••!•.i' .:rc (".Ii-Tsi the final condensation. This would lead to excessive thermal

ii.i.-. ;u'ii;ii built into the film. Temperature gradients across a 1 mm thick i ; sli.-'ii. •.•an be rM° to 100°C, so for elements like palladium, scraetimes a

-.' ;!;:.:;t.!,:iv because of its much higher thermal conductivity is put on a

••-••.>oli'.l copper plate. In this case the cesium iodide release agent is

• !•.)•• i just prior to the palladium.

ih-- literature cites several instances where the substrate is pre-heated, 10 • illy to iOi)-400"C. This will increase the mobility of the atoms on th ;

.'• t .it-.- surface, the rate of coalescence and the grain size, but it requires

:;••• a substrate with the same coefficient of thermal expansion as the conden-

,- aifiTi.il. iJtherwise, severe strains are built up in the film when it is

11 •• i down. r'nly in cases where the film will not be removed would this .,e

•••! I.I:<1.-. The two titanium films in Fig. 4 were both grown on glass slides

. r !:n- -;,U:IP conditions but witli different substrate layers. The one marked

•;.•..! ,i f i In of aluminum 6000 angstroms thick (150 Ugms/cm ) deposited on i'.:

• • i i. i; .•!-_,' followed by the titanium of about half that thickness (3000 angstroms).

'. :••. [••>'•: .-nip leti on, tlie film sliattered leaving the pattern shown. No effort

!:-..i'i' to prevent the slide from heating up and it reached a high of ~- 7C°C.

, ,•:•::!•,in' (:1K.1 geometric pattern of fracture with slide (b) . This titanium

::i is- l.!;e same thickness and was deposited at the same evaporation rate, 4000

• -t. r. .p\:-i ;•.•!.- ninute. However, the undercoating is calcium iodide. At the; 156

(a) (b)

Titanium films of similar thickness,

bat different substrates.

(a) a 1 umi. num (b j i ale iiini iod i de

•n: U't ii,n ot the run it way about 70"C and the; film was "intact, but as it

••"'Li] l.o room temperature random cracks appeared. In tho first case the stress in fhe film was due to much more than a difference in coefficients of thermal

:•:: ansif'n. The strain was mainly because of a hexagonal structure growing on .'.

Mi'ic suiistrat!'. In the :_:ecorui case tile random Irearimj of tile film is truly diu

'.i' !;:•• difference in the thermal oxj^ansion, by a factor of two, between tile

:la-;s and titanium. Because it is relatively thin, the tensile stress over-car

';.•• ohr-sive s t i'L'nq ti;. In a much thicker film it uiay have remained intact, but i! removed woulrl curl info .i tiqht cylinder. I recommend erhausting tlie list

••'" different substrate materials, before going to a proccdu' e of growing a thin film on a deliberately heated substrate. Table 1 is a list of the partir-f sub- ti'iiL'cs usixl in produciivj sol f-supporting films for targets in tandem accel-

•ral.or ex;n'rinents at Stanford University. 157

Table 1

Substrates used in the growth of self-supporting films.

E] ernent SubsLrate Element Substrate

Aluminum Teepol Manganese Aluminum Oxide Antimony Cesium Iodide Molybdenum Sodium Chloride Beryllium Barium Chloride iJeodymium Barium Iodide Bismuth Cesium Iodide Osmium Sodium Chloride Boron Boron Oxide Palladium Cesium Iodide Cadmium Zinc Chloride Potassium Hexadecylamine b) Calcium Hexadecylamine Praseodymium Aluminum Oxide Carbon Teepol Nickel Coppe r Chromium Potassium Chloride Rhodium Potassium Iodide Cobalt Aluminum Oxide Ruthenium Aluminum Oxide Copper Tcepol Scandium Aluminum Oxide Erbium Aluminum Oxide Silicon Potassium Chloridt Germanium Barium Chloride Silver Teepol Gold Teepol Tellurium Teepol Holmium Calcium Iodide Thulium Calcium Iodide Indium Formvar Tin Teepol Iron Copper Titanium Calcium Iodide L^ad Potassium Chloride Vanadium Potassium Iodide Lithium Hexadecylamine Yttrium Calcium Iodide d) Magnesium Sap on Ytterbium Copper

Teepol 610 (sodium secondary alkyl sulphate) Shell Chemical Co. b) S. H. Maxman, Rev. Sex. Instr. 35_, 1572 (1964) Formvar 15/95E (polyvinyl formal) Monsanto Co., St. Louis, Missouri d) J. L. Gallant, Nucl. Instr. Meth. 102, 477 (1972) 158

References

1- K. D. Leaver and B. N. Chapman, Thin Films (Wykeham Publications Ltd., London, 1971) . 2. L. I. Mirkin, Handbook of X-ray Analysis of Polycrystalline Materials. 3. L. 1. Maisel and R. Glang, Handbook of Thin Film Technology (McGraw-Hill Inc., New York, 1970). '4. A. Taylor and B. J. Kagle, Crystal log raphic Data on Metal and Alloy Struc- tures (Dover Publications, New York, 1963). ''>. M. Harchol, Nuc-i. Instr. Meth. 40, 158 (1966). 6. D. N. Braski, Nucl. Instr. Meth. 1£2_, 553 (1972). 7. A. n. F. Muggleton and F. A. Howe, Nucl. Instr. Meth. ^3_, 211 (1961).

H. J. R. Erskine and D. S. Gemmell, Nucl. Instr. Meth. 24_r 397 (1963). J. R. F. Casten, J. S. Greenberg, G. A. Burginyon and D. A. Bromley, Nucl. Instr. Meth. 80_, 296 (1970). 10. S. II. Maxman, Nucl. Instr. Meth. 50_, 53 (1967). 11. G. V. Samsonov, Handbook of the Physicochemical Properties of the Elements (IFI-Plenum Data Corp., New York, 1968). 159

PREPARATION OF ISOTOP1 CAI.LY F.NRJCtiLD SAMPLES OF IRID1UM, OSMIUM, PALLADIUM, AND 1'LATINUM FOR RESEARCH USE

K. W. McDaniel, L. 0. Love, V. !'. Prater, R. L. Bailey

Oak Ridge National Laboratory Oak Ridge, Tennessee 37330

Abst me i

The .'\ ;:;'.' i fluorination technique, which is currently be ins; employed at Oak Piiire National Laboratory for enriching the stable isotopes of iridium, osmium, palladium, and platinum in electromagnetic isotope separators, is de-scribed.

Ion-source modifications and safe-handl i .nj.; procedures for usinj; chlorine trifluor- iue (the fluorinating agent), both necessary for successful separation, are given in some detail.

It is suggested that the internal fluorination method might be used in smaller laboratory separators for target making by depositional methods.

Simple chemical recovery and purification schemes are i'.iven for each of the four elements, along wi'\ some of the current uses of the nuclides in research and me d i c i n e .

-A Research sponsored by the U.S. Atomic Energy Commission under contract with the Union Carbide Corporation, Oak Ridge, Tennessee 37830. 100

!'Ki-.!'ARAno.\ OF 1 HIM <)VICAi.i.Y I.'-Kl''.'. ••'. !' :.A!'.:'i.!..; OF I "., i'.T.'-'.I 1,'M , i'Ai.i.Ai)! L'M, ,Vsi) i'i.ATlNCy. i '•; Ki:Si.Aj<:Gi i:Si.

i\a' :.:i: •.' ' ..si's Liu ft1 has r>;isU:J a need I'M' separated .iscti 'pc.-^ o'. the p S it L; ,ui"

;. - s-._ > i i •. i ••! , indents, e,pedal ly I'or tiie study o! v.ii:ii-i;: inrit.ij- ;r porties. L'J.LIJ • rr.i "I i he e I ei:-en I. . Th is i • i.n i •.;(».•, ' ..';u,:.; si/i i:ia.s:'ii i i: :•;. l.ii ni ii>; ;-n.l •!]>; ( a 1 ! y •. iiriciicd iwciii.u- • s "I iridi".:;., ;•., i ; aa i ;:;.,, ..r.ii p ! ,n : nu;:. (ii^::i- i.::. •.!.!.-. !;:••;!. . ,c ;ia r a! a ii 11 •. i i.. •, I '• 11 • 1.1 r.:>:i i!c) , ri~ii;:!n. u ::•:'! •, i;aa ; i I i t-s. "I m.ii.. •- lia] '.•.•:!;: i I-^L1 r-t aaii-av.'rave- i :.DL"; . i a purities.

. ;...•.-.' ti ;a:iicjuv du vv i u[)i.' J !"!.-'< t'I;/. !y a; i •]•:'.';., pat t''i':;i ..i a] L .. r tia. \'.\.-: -:. i .ia a ua'1 .^! i--11-:..<.•]; J a r .' i iLiorJiiL'" and iJiv.i.'.i i::-.'-- ij: v.Li ., ut i 1 i /..v; f. i i" . (..! i\ icl witii a;:d uraa^pi ia piatinui;; and p lal i luini-^râp niclais d i iv t 1', iat- tau i on i zaL Loii rL-'i'Mi •.'1 tin.- i .-.! lit. ri);: i uii suurcc. iaiD:parud Lu tin- •.- U c 1 ron-a<:%harU:-,ia: t ':; ,;a -Cfi:.peraL i:rc :::''liiod a 1 :u.'parat iai'.- p laL i nai;!-;!!*.1 La Is i L-L-L i>p<> , i.i-- -:-u r'tv-n t L--. '•; i: \/.] uu has r.. v j ] u- I i oiii Zi.'ci i.lu- saparat. i ens oi Liu- p 1 ;;t i i'.uir.-i'.v i. i! is.-Lipas (u:;pi.'c !.••'! Lv i r i d i una in ; • I'I • ail !ia t aad i s ; t • > p i c put: i Ly . (Ire a L ca" o '• • a'- : : ; 11!:> •-1 : r i a i nr , isri ur:, j\:: ; a a i m~., .•HI'.; p!ati;:nni .na- ,-•: i-r-.1 »t ly fcr SDI.II iili ;U') a.-.: i ! .I'.I I ^ it IIIV/HT ist't.pi.' parity ! :i-.iu L ia'iiâ t ! â.M ;: ]L i 'a.1-, i a ,i;:.;ri [ i ;:;L- ,I^;I> .

ia1 ; I i;a if L nat i ;a; :\L- t. aad in :a • i ii;' aSL- d in p raaiK L i L;II- t. vp./ s^: p ;rat > a ^. ; jr t iir rurov- ci'v i I /. r.in ipian t i 1-ias ui must, DI' Lin.- :su'i'[v-i; ;n)Wc\-L-;', ti'^iv appears [.;"•• bu no r ^ •.!,-•• = 3; '.vir: : iu- t t.'i:i; ni-qut' C.MJJ(] not i-u- a.-.t-d , ri :a:^a 1 1 S'.-pjr.itt.'rs v.:a.-r^ L-::;piia.sis is p:a....

V a-.t..'d !i'i tiu' ' raiisiii 1 .ss i on J t CC 1;, Lu L i: L c:iar;j.c iujLtiL' -.•.'.'is Vvp I aci-d v/iLh a nickel ! : in lur !iii-: L ranmia s>; i i n at" |_. I 1-', I're:. ,• cvlindar niuuntvd nn the ii i aa-vo] tage tcr- iiiia.ii bo:-: aiia.TLly inLo Liu- '.>.ii-k ui tin.- ar/ ciuiir.iio r wiiii-ii aoiiiaineJ the eJcncnt ( is iiiflal) ;.)i.-in>: p rinL'Ssod; iiccaasi.- <;l IIK1 1~COC L i on be LWI-IMI grap'nite ar.d Clr-j and p. ss i i.. 1<- n-ai.-t i oiis but'.-.'ei.-r. p 1 at i nun-aela J c .or:,;:aa:nds and j-,rapl> i 1 e , a. nirkel liner •-.I.-; i aii i" i ratt-d to cover tiie inside surfaces of the :n"t. chaniber; the nickel linei' •...is ]!i-r! iiiati:i! vitii l/3_!-in. iioJes every i/'-i-in. to providi- a more unilrrn f 1 ..-,<- ii[ ill. tiV'-r Liie netaJ.j tile oven heater was eiini naL.d; ti;ub the i on Lzat ion arc and i.u- < ititrorj drain supply 1 roiu tue sourco were couiitea CMI to- pro\-i(:e sufficient a---.il lor LIIL- iau-r.iicai reaction lietwcL-n tile kll-". .aid meL-.tl . ALuj\e tile vaporization Leiiiperal ure .1 tile metallic fluoride, tile rate of ia.ac.tioii vas , i.;i (. roJ 1 ed largely i.y tile : !. iv rate o! (.IF; which •.-.-, is regulated by a n.eterinj; vaivi outside the ion

Al l;:.m,;:i C i F . is toy.ii: and ii az ardous , as a resnl! ef til-.- adu-p t i i in ui a very ri^ : set oi n.uidii n;.', procedu rus , no inei dents have been experienced in tile use of tilis ^ja.s . Figure .1 siiows Liie precautions t aive n durini; rvir. tern: nat i ons , lilese procedures were instituted durin;.!, sliort du ve lopr.ent runs with tiie advice and approval • a tiie i'lant Industrial Hygieni (Iroup. As a further protective me ch,m is:?i as we]] a.-, .u; aid to reducing tank pressures, caiciun; metal was vaporized at tiie rate oi 1-3 :' Ca/br. into l.iie lar^e '.'acaiun sys tei;: vnie.n i'l}\ was beim; used.

\ . V . liueiin, !i . V.. C.avr i I ov , ai\d V. :> . Zojotarev, :•'. i •. •': i roaiu^nc t i c Separation of tiie isotopes •:•;.' tiie !' 1 at j nun-!'.!! J adi ur.i Croup, 1 soti.penpra>: i « 7(b): 23d-5 (19 71); i ii loass i ;ii:. 161

HIGH-VOLTAGE TV TERMINAL

PI METAL AND/OR SPONGE

ION CHAMBER

Fig. 1. Fluorination System for Calutron Ion Source.

Fig. 2. Precautions Taken During Run Terminations. 162

Table 1

COMPARISON OF ASSAYS USING FJJ:CTRO.\' UUMHARDMJiNT SOURCE AND FLUORINATION WITH GIF,

NA "Bombardm ent" "GIF/1 U) (%) C.)

:•)• Pt 0.0127 0.76 4 .2

1 !; i't 0.78 n.s 57.3

•'•"•l-t. 32.90 65 .0 93 .9-97.4

i ."Tt 33.80 60.1 9 7.3

: !•• 1>t 2 5 . 30 65.9 9 7.5

- l:i\'t 7.21 60.9 95 .8

XUN DATA: Time 4,769 lir 4,400 hr Recovered 64.9 a 176 .3 !•, ::.!,,. 37.3 9 4.66 9 7.1-9 8.2

i ) .'. | r (.2.7 9 8.7 96.9-99 .4

IUN DATA: Time ..,471 h r 2,600 hr Recovered 2.7 g 151 .0 g

i.cule 1 is .i Lable comparing the outputs and purities of the GIF-j system with the electron hiiir.bardment method.

Chemical Recovery and Refinement of Enriched Platinum-Metal Isotopes

Iridiuni, osmium, palladium, and platinum isotopes are all collected on fabricated, ii iyn-p-.iri Ly graphite receiver pockets (receiver is shown in Figures 3 and 4). The cnriihcd material is recovered in each case by the simple expedient of burning i mlj i' idu.'jl graphite pockets, in oxygen, after removal of any neutral or adjacent- iHoLupe material from the outer surfaces of the pockets. Iridium, palladium, and p 1 at i unuiii isotopes remain in the quartz ignition containers, along with any graph- ilc impurity, as relatively pure metals. Some oxide is usually associated with tiie iridium and palladium deposits. Osmium, however, forms a volatile oxide Vi>,[), u.p. 121.2°C) during the ignition step and consequently must be chemically trapped for further recovery. 163

Fijj;. 3. Osmium Receiver With Faceplate.

Fig. 4. Osmium Receiver Without Faceplate .16-3

1 rid i uiv

: r id i u:.. I .-. .'Hi1 i'! i;i> .:<>ie dillicult tlr-tnts t .-• reii;:<- !r ,, p,:re .-•'..;!.• ;> r i ::.•: r i ! y . ;ijr Li1 t;: i Vi r .' 11|[.-.' so! -Ui i i i t\ o i the mo L .11 (jr ' -x Kie ) in c i t_' K t :;; ii: ] v '.• r t'.ii :•:-.<; .:'-u.s. Ii: aduiLion, since there are no sclfctiw precijii l..-;i;l.s i.iiown ]i>r 1'L-;:.UV i r.i1 i r: uii n::. 1 • ~ . . 11 r i t • :p I cs i uns Lliat urc.is iona 1 1 y o.-ciir unri:'.;1 UM1 V.: i :,•_• ;irv.;:-.'.ii"i' !•.•'.•! JI.'L! KM.! Li' vaeii.iral > < in I ami naL i oi is ainJ/cr l.'.sses I.! i •;• >L up; < a 1 '. : L :.- : i. : i tii VAAI I. r i a I , v. :i i I < • I i:«• 'us inn :;;i-! hod {.ends L f i inl rudiiii' h i ;;.i :.. ^ i.: i • • ' • : .s • :l,..iii .-.alls wiiicii i nl e ]')r re with jirev. i p i I at i on ir.etnods and are . ! i • n diili.-.!'. ! • ••; j'Vi' .j'.j.i]'. t i I at i ve 1 v .

'•.-. .i les'.iit v-'i tiie associated .soluiiility res t r i r t i i. ;:s, earii ol tiu1 l:.'<. isoteje.-, !.-/••'I .jiui -I'M) wa.s iioi:;:ji;e;>i zed by liieciian i r.il ly stirring tiie sai:ipleM i ii va'.ei :'• r uj' Le 7.1 ..LMII'.I L-.uii. liiis technique, although ordinarily uncies i rab 1 e, v:as .Ki^rt l.r tvo reasons: (,1) pru11 iir.inary mass analyses ol 15 individual sanijiles ..-in;.-, t i- tul ii:,,1 e.i;:ii i..iiUi|'iv lot diilered by less liian j t;r,is:-; jjerreiit, thereby es! ab i i s.. ii:.-; .. rear. • i:a:-1 e ii;:.)t. oJ i nhoir.ogeile i t\r, and {!) .ivaiia'jle dissolutiL'n r:e liiov:^ •.eie v-'iis i de i"e;i .. iLaei" teo hazardous or :nade'[uale !"or tiie larye sariMe sizes. i -le:..! •. .i 1 IVI'i ne'..;ei'.i. v:as ae coinp 1 i ;;ii ed by renioviiiK soluble in.purities by K-av.-., i;i>;

1 e, iiiii'jne.s r.iLinv tnan !)V Liie usual prut, ip i LaL ion methods. Foliuwirn; a ny drr>'.'-T. re due t K-n at til)0"L to entail", tiie most reacLive ion:! el as many i;..pur i. Lv ele::;ents pessiijle, eaca HI Lue i so Lopi ca 1 Ly enriched samples uas LreaLed first wit'n iiv! 1 iH •iil'.ile ( i : 1.) nitric acid and then i-.'ith Jilulcu hydroch 11 •y i c acid; t.ie n;e t a i was .i.isaed al U !' each step- wiLli a 1,'. solution o I" M-.,(;] . '1 ne liquid was separated iro iae ii. a. ned i^e L a ! by cen t r i t '.!j.;at ion , and the irceLa! v:as air-dried a i I ..• r vasi: l i-.i.1, .. i t a ,ii-i I one and etner. It vas Cinally re-:educed in hydrogen at 'JUl'cC to obtain :''ri>.at, si^inv, i.ieLaliie i r i d i urn. i he' metallic sai::pJe.s wei'e ccoleii in a 1 1 ew oi .ii";.o:; L.; j-re'.ef.'t ;>..' ->.s j ;:• J e adsorption ol hydrogen.

Osmium

i he chemistry ol osi'.u um is dominated bv one compound, tile Letroxide, Ost)^ , bec:iust.' il is both volatile and toxic - properties that have to be considered carefully in tiie recovery and refinement of the isotopes in order to prevent losses by volatility and avoid exposure of personnel. The oxide, for example, can be formed even at ! ov (ei;;per;)t ure when finely divided metal is exposed to oxygen; it is also forme J .-.'hen Liie nie t a 1 or one of its compounds is boiled in nitric acid. Osmium tetroxide : as an appreciable vapur pressure even at room temperature and lias to be handled v. itii i . • prolerabjy in a closed system. Its vapors arc poisonous (maxiiTiim per- ::;i;Ud at inospiie r i c level 2 x l(_Tf:> g/niJ) . It attacks the eyes, causing blurring ui vision and, in severe cases, temporary blindness; it atso attacks and irritates tiie nose and throat linings and may aggravate bronchial conditions.

As ::,eiitioned earlier, osmium isotopes are recovered initially from graphite collector pockeLs lv ignition in oxygen, which oxidizes osmium to volatile 0s0L and tile graphite, to O), . The quartz combustion train (Figure 5), fitted with ground joints and .leaied wilii Kilieone grease (which does no_l_ reduce OsO ), includes three sequen- tial i.-jiei.i i ral traps which are partially filled, respectively, with, concentrated iiir + Nl!..!;r (u g/g of Us estimated), 1:1 lilir, and an alcoholic solution of KOH.

An oxygen flow is begun and heat applied to the quartz Lube by means oi an electric lube lurnace. An appreciable quantity of 0s0:. appears just under 300°C as noted by i he dee.p ridden i ng of the first trap of concentrated liBr. 'Ihe ignition is continued until all she ;:raph.ite is consumed at ' 800 °C. 1G1

Fig. 5. Combustion Train foi Recovering Osmium.

•-..• !M',r I rap scluLiuns are conbined and the solution is evaporated to near dryness •I; a steai:i ii.u;]. iJilute (1:1) 11C1 is added to ttie moist salts and the solution is )"'-ev.:;>i'r.iled to near dryness, (This conversion to chloride is made because Osti it". |'.i:'ed I )"ur,; liiloride solution is easier to wash and reduces to a cleaner metal.) .';•; ,n::,::i!ii i u!;; s;ij t i .s added during evaporaLion to )irevent loss o 1" osmium.

i ...- a :;i::'>n i ui:; ch 1 oro-usma t e salts are dissolved in distilled water and the resulting .w'.mie;: is ireated dropwise with saturated NaliCO; to pli .'*, at which point a black jTri.-ip i lai e oi iiydrated OsO-. appears. The solution and precipitate are heated to '.)• • i 1 . u;.\ I., ensure complete precipitation.

iin. ..it.-:-: i t], • is separated by cen t r i fuga t Lon, washed free from sodium ion with a 12 -.olui. ion ..I Mi.,Cl, and finally allowed to air-dry. The dry OsO- powder is trans- ;i-rr..-;i I" a quarts bo;:L covered with Nil.,t'. 1 and moistened, reduced with 11 at 900°C, a:: ••: tin:; a 1 I ov.''.d to cool in an atmosphere of CO • which appears to protect the metal i i\.<:. it ..;•: ic.it i"!i .

1'al ladi ura

-.ii- JVI •••.'<. ry .i:d re) i ncr.cnt of palladium isotopes are primarily dependent on the ease- wit;: wiiuh i-Ki:iental palladium is reduced from solution with forraic acid.

I'iu asa resultii!;-, froin the ignition of the pal ladi urn-bear ing graphite collector pui.i-.ets is moistened with formic acid and waraied to destroy the surface coating of ;ii'o\;:i palladium oxide. (The absence of 1M0 appears to enhance the solubility of palladium.) the residue is boiled with concentrated H.\'O-, until solution is essentially complete. iiie solution is then evaporated to near drynoss and taken up in y> v..\ •'! aqua rei'i.i. Alternate additions of concent rated lid and evaporations. 106

v.'itli formic acid, na::pk'tw Lin.-

V 1 at i nun

. .• :T,-;',vrv and re 1 i neir.en t of. platinum isotopes, like (hose ul pa 1 I .'.di uv., are ,- j . •:,( •t-.w! .wj the rt our ih i 1 i tv ot the element fro::i solution with forn.ie ,'ici-:.

..::<• r iv;.,i'va 1 iM so[u;>k' impurities Lrom platinum residues wilh diiute nvdr. <•:>!'::-'.• : • : i;i[)"ii- .ji:ids, wiiii appropriate water extractions alter each (. re a 1. me nt , v i .it ; •: u- i.. dissolved liy boiling with aqua rogia. The UNO-, is eliminated by success i \ * .••. : •i.;I ioiis \-.ril!i concentrated UC1, iollowed by small additions ol formic acid. i:n' :,. :i!f!i:v, sn] ill iun is neutralized (p[l - ' 4) with Nai)ll and platinum is ji ret' i \< \ i .:( -.- : . •. .nidiiif: lonaic acid and boiling- (Nalill is used instead oi Ml,,()l! in order [•- .i.-..ill pr.T i p i tuL I on of ammonium cliJ uropl at inat e .)

: I at; mi::, ;:»•(.•]) is separated by ceil t ri fugat ion, is washed with a 2/ solution . .' '.::,i.'i [.i i'ei:i<)ve traces ol' sodium, is air-dried in the centrifuge tube, then re:.- .- - :•• -i -'.u-.i Lz oi" Vvcor disli wiie re the sample is i inal ly heated to 8')n°C, a». i-.h ic:i ,- int. n! uiiium assumes the characteristic metallic gray appearance.

L'.si-s of liii.' iM at inum-Metal Isotopes in Research

"-.,[..-ii.;i: .ill el I'uc isotopes ol the polynuclidic platinum-metal eler'-jnts .ire des i r-.

in .-.I.:", ';u.ml ilies lor studies of miclear structure by the (n, :)reaction, os;ji:i::: .•-u! o'.ie:; are currently drawing the highest interest due to their invv 1 wr.ic-nt in .tiicj.-ar ..i i rono 1 o^.y studies proposed by U. 1). Clayton in Astrophys. J. 13^', b^7 I. >ii,.,j ,md Nature J24, 56 (196'J). Specifically there is an urgent need lor care- i • i i !y Ei-.isurcJ '"'Us and '" 7Os neutron-capture cross sections in the keV lo sev- • r.i i -liunureii-ke V range .

:., .:.i-.i i L i on to nuc I car-meas uremeiit studies, highly enriched '*Vl is currL-iuiv i i:u .,'.,. u.-,e a.-, Larg.et material for the production of iJ-mPt which is being used in ::u:-le.ir i'.;eri i c i ne . The 4 .0-day 1''' ITil>L lias gamma emissions of 99 keV (11/'.) and 1J> • ••''.' (l.ti, ) that make it suitable Cor whole-body counting. Following Professor .'.. ,'Ul :•'•;; irradiations, and systhesize and provide the radi opJ at inum-1 abe 1 ed cor.'pou.u

i •-!.'•!-'•') 7 7 , "Nucle/ir Medicine and liiomedical Kadioisotope lecimo ] ogv ml iiil v -l.'ecci;iber 1973," i . K. I'oggenburg, May 1'I74. 107

EVAPORATED SELF-SUPPORTING FOILS OF ENRICHED

IRON FROM THE IN-SITU REDUCTION OF Fe20-,. W. D. Riol Department of Physics State University of New York Stony Brook, New York 1197*3

ABSTRACT: This paper deer i !M?::. the reduction of i sotopi rally enriched Fe0CQ usinc a controlled thermite reaction. The method will produce strain free y evaporated films of 100-300 ;,y/cm which can be floated and mounted, using standard techniques, to make strong self-supporting targets. Producing self-supporting targets of enriched Fe foils from the oxide, which is the most inexpensive form available, is a difficult procedure. The general approach is to reduce the oxide either electrolyticly or with hydrogen and subsequently evaporate the won metal from tungston boats or filaments. The two greatest difficulties encounteres in evaporations of iron are its high solubility in almost any boat material available and the extremely brittle foils produced which must be annealed at temperatures of 450°C or better thus assuring the destruction of the releasing agent. Peck roiling the won iron is a suitable method but requires that excess metal available for trimming during the initial passes through the mill. It requires an experienced hand indeed to roll iron foils -200 wgs/cm2. A method was devised which uses the least amount of an isotope in its conmon form, the oxide, and simultaneously anneals the iron film allowing it to be floated off on water and picked up in the usual fashion. The method developed uses aluminum powder as the reducing agent. This is also the basis of the commercial thermite process used to weld on a large scale. The reaction is rapid and extremely exothermic yielding 185 K calories per mole and must be controlled. After much investigation the best means of impending the reaction was found to be mixing an inert metallic powder into the batch. 200 mesh molybdenum powder is ideal. The ratio of its volume to that of the reactants is used to regulate the temperature at which the reduction takes place. The reduction can then be sustained at a temperature chosen to provide the correct amount of radiant heat to anneal the iron as it is being deposited. The procedure used for an evaporation is as fellows: 10 nigs of enriched Fe20, is ground together with 4 mgs of aluminum powder in a spot plate. 1.5 times the volume of this mixture is added as 200 mesh molybdenum powder and they are ground together thoroughly. This is then loaded into a boat made from a piece of 1/2" tantalum tubing 6 cm. long, pinched off at its end and a 2 mm hole drilled at its center. A tantalum mask with a 3/8" 168

hole punched in it is mounted 2 mm above the boat in the evaporator. A 3" x 4" piece of stainless steel ferrotype plate (the kind used ti: <*.ry photographic prints on) is next mounted on a vacuum manipulator i.5 cm above the mask. The manipulator allows the Substrate to be swung over a second tantalum ribbon boat charged with Bal? releasing agent. The system is evacuated and "-40 pgs/cm of BaI„ is deposited on the substrate. The otier boat is then heated slowly to 1000'C as observed with an optical pyrometer and the stainless substrate swung into position over this boat. This will avoid the snnll grains of powder which inevitably blow out of the boat's mouth on initial heat up from aligning on the substrate to cause pin holes in the foil. The temperature can then be raised rapidly to 1600°C. Five minute;, at this temperature will produce foils of --100 ugs/cm^. Lowering the temperature rapidly to 1000°C will stop the deposition from which point at least ten more minutes must be alloted to turning the power completely off. The bell jar is then vented to ••-1/3 atmostphere of argon. This is allowed thirty minutes to cool the substrate by convection. The bell jar is then vented slowly to air. This procedure minimizes thermal shock and the resultant strains in the iron. It will also provide an adsorbed layer of argon which helps considerably to inhibit the corrosive effects of the floating off process. Should foils show signs of cracking at any stage of mounting or in storage it is a result of strains and the evaporation temperature must be raised and the cooling down times increased. While target thickness is critical the stainless substrate is provided with a 1/2" hole through which a quartz crystal ii'onitor sees the evaporant at a distance suitable to lower the temperature on the unit. It can be calibrated gravametricly in evaporations using natural iron. Upon completion of the evaporation the iron on the substrate is scored into appropriate squares and floated off in water and picked up on target frames having 1/2" holes. No special precautions other than placing them in a vacuum desiccator to dry as quickly as possible are necessary. The foils can be stored for long periods in an argon flushed desiccator. 169

SPKCIAL TAKCKT i'RHPARAT I UN TKCHN l< il/l-.H FOR CHALK K1VICK NTU.KAR PHYSICS

I. I.. Cal l.-mt

Atomic F.nergy of Canada l.irai to.i >i'S Division, Chalk River Xnclear (..ibora Chalk Kiver, Ontario, Canada Kfl.l l.!O

AliSTKAC

ihe techniques and apparatus described in this paper were developed tv fulfill the needs of Nuclear Physics Research at Chalk Rlwr. Thcv include: sale apparatus lor the preparation of adsorbed tritium targets, .1 t echni<;ue :\ tiie I'aiiri i\it ion o! temperature resistant tar'gets of i sotopi <•;: ] 1 v enriched seU-niu:!i, an apparatus tor the p re[7a ra tion oJ reat'tive nit-taL targets lor scattering exper int'n C;;, and a !ec!in.ique I or the labrication .i<\d ::u>un t i n ; i>: u 1 ! r.-i-i'h i a, J ..s/riif carbon .stripper films.

In t'ne 1 lelci ot Nuclear Phvs'c i-ii'sea reii, success t>! ten (iej^ends on liio qu.ililv oi the target material. It cinst meet the highesl siandard oi purity, th (iiickue.ss fitus i ol ten be measured with i'.rea! ai'curacv1, and the target- must he c;;c::! i ca 1 ! v stable and iiea t re.sistant. Many techniques suitable tor one type e. I nuclear reaction may not be appropriate to another. Some special preparation methods developed for Ml' tandem accelerator experiments have been published previcrasiy . In this paper, other techniques used at Chalk River lor the nuclear physics program will be discussed. They are: 1. A safe apparatus lor the preparation of adsorbed tritium targets. .'.. A technique for the fabrication of temperature resistant targets of iLotopically enriched selenium. i. An apparatus for the preparation of reactive metal targets. • 1. A technique for the fabrication and mounting of ultra-thin 2 ..g/cii!*", carbon stripper Jot Is.

1. A Sale_ A£>par_a_t_u_s_ _f_t>j;^jjjc _ J!r_£P_'-Lr_i1 •- L°iL °'L •\l'ii!irJ";L4 lr i t iuni Targets Solid targets containing tritium are very valuable for the study of nuclear reactions using heavy ion beams, such as T( C,p) C (ref. 2). The apparatus to be described has been used to produce titanium tritide targets on 170

liiii-k b.ick i nj;s and on thin copper foils. The latter targets were stretched to

;M-,ivi.ie flat tartiets suitable for the measurement of nuclear lifetimes by the

iv.c. i I distance technique.

'i lie tritium target preparation apparatus is illustrated in Fig. 1. All

L'IC pn is are stainless steel and consist of a reaction chamber, a 1'irani type

;-i"!• .:-;ij-1• gauge, a calibrated chamber, an 8 1/s ion pump, a cryopurap and a small

(if CI:I I ch.inber containing about 25 grams of uranium turnings used to purify

a!-, i . iorr tritium. The reaction chamber and the uranium chamber have internal

i i on-const an tan thermocouples to monitor temperatures. The volume of the

;>!'>• .sure gauge and the associated chamber have been accurately measured so ihat

'.'.!<• total volume of tritium used can be determined. This stainless steel systen

i- ! i::: i nat es the. dangers of tritium release from breakage of a glass svstem.

M i-iuii.'ii tritium is contained within the system by adsorption in the cellular

•;;iiuU' id the ion pump .

I'rior to use with tritium, the uranium oven must be activated with

;.-.•..' :'<>::<-ii. Ihis hvdrogen c'nemically reacts with the uranium to lorni hydrides.

.-.'I'.-L: I iu- hydride is subsequently decomposed, the uranium becomes a finely divide.! powder which will readsorb hydrogen readily . The procedure or this

."ict ivation is as follows. Uranium turnings, normally stored i" oil, are

.v.'.Teased in triihlorethylene and rinsed in acetone and distilled water. The timings are then immersed in 8 N nitric acid to dissolve the surface oxides.

: :.e uranium now exhibits a silver metallic lustre. The turnings are rinsed

>:.. wra 1 times in distilled water, in alcohol and then transferred rapidly to

; '.••• own.

1'iie i.H'1'ii is evacuated and baked at approximately 600°C for ten rainutes,

•:-.-..!i the fe;:ijHT,-ittire is reduced to approximately 120°C. Hydrogen is introduced

;!i'.idil;, at a pressure of 13 kl'a (100 mm). After the absorption lias stopped,

(lie temperature is raised above 300^'C and the hydrogen is released. This

•u .•••••diire is repeated several titr^s because the initial impregnation of hydrogen

;:: uranium is slow. When the uranium lias become sufficiently activated, the

1 i-;:!jic[\i tuiv is increased to 500°C to remove all lydrogen and the u "nniutii oven i-

•..lived ul'f.

To charge the uranium with tritium, the uranium is heated to 120°I.', and i- ::own quantities of tritium are repeatedly absorbed from tile calibrated chamber. i :.. ai'Si)i";H i on jirocess will, .apidly reduce the tritium pressure from 13 ki'a

\ iii'.i :".;"! o! llg) to 133 tn Pa (1 micron) in the chamber. The target foil tritiatiin

;-:oi edure is as follows. The foils to be tritiated are mounted between two

•uainless steel discs (Fig. 2) several foils are mounted between spacers on j

:-,t .i inl L-s.-i steel rack (Fig. 3) which is then inserted into the reaction chamber

.:: ! .i-.-hi-ii !o the tritium apparatus and evacuated to J 30 ,.Pa (10 tori). The h.n tiv'n cha;:iiii'!' is then heaceti to 500°C until the contents havt been outgassed 171

and then the temperature is dropped to 32O°C. The pressure should now ruaJ about 1 j . ra (10 ' Lorr). The uranium oven is heated to 22O°C. Ine reaction I'.nanb.-r and ion pump are vaived oi 1 and the uranium oven valve is opened. Tritium is introduced into the calibration ehambe to a pressure oi 27 kia (200 mm of llg) , the valve is reciosed and the uranium oven temperature is reduced to 12U'JC. ihc- valve is opened to the reaction chamber unj the pressure drops steadily as tritium is absorbed into the titanium. When the pressure readies 1. 'i kl'a (JO ::v:, ni ;<;;"> tlie valve to the uranium oven is opened and the remaining tritium is readsorbcd by the uranium. The ca librati on and reaction chambers are a;sain evacuated to 13 ..I'a (10 torr) and the tritiation cycle is repeated. With this technique the amount of tritium absorbed by the titaniu.,1 cm be monitored. It has K'jeu possible to produce targets containing one ato:s ••.'! tritium per atom of titanium. After the initiation is complete, the- temperature o: Lhe reaction chamber is allowed to drop slowly to room temperature.

A i'echiriqjie__!_or_ the Fab/Jcaj.i__on_ oj l'_einjTei"_a_t_utry _K_e.si '-XanJ^_J_;y",y_e_t_s_ c-J [ so Lop i ca i l_y tJTrJ ehed _Se. i eji_i_ura Selenium nictal has a nieltin.q point of 217 = i^ am! vapour pressure of 1 ; .i'a • it J2r)°i:. Targets oi metal lie .selenium deteriorate rapidlv u-ider hemhaiiimer, t. i wo p I'ru-L'du res were:1 used at Chalk River to prepare isoropica! Iv enri.iied s^ 1 ^> il: • n. l.iiTi.'ts for tin- foil Luwi'i^ nuclear physics exper in)en lr..

>•'i l:ns ill 11 uiii i niuni ( 70 .:i;/om ) are mounted on suitable frames aaci pla--e n a v.K'in.n systei". Molybdenum and lungs (.or. boats contain respectively Se an^. ure alimiini nr. lor evaporation. The selemun is sublimed onto the aluminium to

iie rc'[uireci thickness and a film of a 1 uminium ( 70 ..g/cm") cv,i-,.'jr,iLed ovc" the

•I h'l.i '• ,

':-r..i selei.ide has a i-.ii.-l tin,, ,"inL of 106b' C. iliis "asily evapora;ed

ie; : •] .in be used to fabricate target; of isc topi ca 1 ly Ciii~ic!-.ed .,olenium.

ejiar.!' > on ol Si. len i do

Lead selenide i .L: prepared hv melting s toi chionio t r i i- quant t L i es of oiopic SL-leniur.i anil lead . i'he melting is carried out in an evacuated quart;-; anbi-i at ll!l)(l"i:. The grey metallic lead sclonide is evapcr.i i eil from a I v! .; i •iium boar on a suitable ;;ubs t ra te, such as a ibin ^ i h'n o, carbon or "i he targets produced by these two methods were exposed to a 33 >ieV He

•:•.<•::. •.••••• a i ;i ii-fs i I, v '; . :• .A for many lioars with no sign of deterioration.

.'. Mi liind f 11r_ the_ I'rej^ar

';'.. i r; •.!_•(. s '-

.'• :'i:;i-!]fs such as tiie recent studies of Li breakup . These- metals will

:••:•: '.. i i !i air and deter iorate rapidly. As several targets are needed for one

• -;.i ri.'i.ent , the i hamber described below is used. The apparatus consists of a

•• .: i: 11 • i -i pyre:-: cross with tile following attachments:

1. A 9 inch adapter plate, with an O-ring to fit over a conventional

7 inch pumpiiv, system (Fig. 4).

J.. '•'•:M: horizontal outlet with an adaptor flange to fit an 8 inch

'."Lingo mi tile Ortec scattering chamber (Fig. 5).

Another horizontal outlet equipped with a flange containing

two high current electrical feedthroughs (Fig. 6).

i he remaining vertical outlet has a rotatable feedthrough Tor

a thickness monitor sensor head (Fig. 7).

'i h i s apparatus permits lithium or calcium targets to be produced by i".'.ijMi";i! \ en on tiiln carbon films preraounted en target frames. The auxiliary

;.")i :.•.<.•! . j-,. .T;;J) <_• r allows up to eight targets to be produced and transferred under vii.iiiK'.: or in an atmosphere of inert gas to the Ortec 19 inch diameter scattering

>•: .a:-i;c r . I'iir apparatus is easily installed on a conventional pumping system and ;•.. IT i I .-, an accurate monitoring of target thickness.

A lechnique for the Fabrication and Mounting of Ultra-Thin 2 eg/cm" '• '"'bon Stripper Foils Thin carbon-foil strippers provide significantly higher average states !.u.!ii ..as strippers for heavy ion beams. However, for stripping in the terminals <>: tandem accelerators they suffer from two disadvantages compared to gas strippers. liiey have a finite lifetime and they produce larger multiple scat tor ing since self-supporting foil thicknesses are generally much thicker I IMP. necessary to produce an equilibrium charge state distributior, at tandem energies. Therefore il is very desirable to produce the thinnest possible f.!r!«m films, providing that the reduction in thickness does not decrease their l.iutiir.e under bombardment. :«• have developed a technique for producing ultra-thin (2 ;jg/cm*~) carbon :: iis ;iv supporting them on a 5 iig/cm film of cellulose nitrate. The cellulose :i;;r.iU' "J.ikes them easy to mount on frames and transport to the accelerator •A I ''• -t breakage, and it quickly evaporates under bombardment. These foils are 1 I J

:>•• ..•!' . ,IVI:!';I;>' lifetimes s i m i J ;J r to 5 or 10 ..g/ci;r" foils, with improved

1 r,insn o! the lower multiple scattering.

i": n • ;:ftiii!ii oi '-.reparation is s tr;::' ght forward. 2 ;.. g/cin" of carbon is

' '• iei onto .i ;:i i rroscope slide which has been coated with a suitable parting

. i i!t• carbon loaled glass is dipped in a dilute solution of cellulose i'- ( 7(! ml o! col Indian in one litre of iso-amyl acetate) and allowed to dry

iit'Ju in a vertical posiLion. The film is scribed and individual foils are

eil on warm distilled water aru' mounted on metal frames with an opening of

ii [.iiueLei- (!•" i g. 8). Tlie '.'. ..;;/cm" foils with cellulose nitrate layer could

.mt.eil on these I'rames with a success rate greater than 80%.

i he 12 nun diameter of the self supporting foil is somewhat larger than

ri,-in.il (8 mm) size provided for the National Electrostatics Corp. foil

er in order to prevent the beam striking the frame, (Fig. 9). in this way

'i.yed to avoid activation of the frames and foil breakage at the frame

'I lie large opening size is useful in compensating for the difficulty in

luc ib ];•. positioning the foil holdtrs behind the beam-defining aperture,

iiform darkening and eventual cracking in the bombarded region which is

Mt in I'ig. 8 is typical o! the failures observed in stripper /oils.

ihe author wishes to thank N. Burn and Drs. T.K. Alexander,

\.:(. '-icDona hi and II. Schnieing for llieir invaluable assistance.

i-U-; <.• re!]'..'!_•»

!. I.I., liallant, Nuclear Instr. and Meth. 102 (1972) 477.

;.K. Alexander, C..I. Costa, J.S. Forster, 0. HSusser, A.B. McDonald, A. Olin

m.i •.-.'. Witthuhn, Phys. Rev. C9 (1974) 1748.

'. K.S. Kocl)lin, Rev. Sci. Jnstr. 22. (1952) 100.

Selenium and Selenides, D.M. Chizhikov and V.P. Shchastlivyi, 1968,

Academy ol Science U.S.S.R.

'-.. H.i;. Mak, Cl.T. Kwjn, II. C. Evans, A.B. McDonald and T.K. Alexander,

Atomic Ijiergv of Canada Limited report AECL-4841, p. 10, 1974. 17-1 V i \\ • ire J : I'r i t i a L ed I c i 1 numn L rd he t fwc stainless steel discs.

Figure J: Stainless steel, rack supporLing several Lritiated foils. for the. preparation of reactive metal Largclt

i"i' '): Klan):,e run t ,1 i ni nj; the Ortec si'a L t IT i ny, chamber. 'Sir

!•' 1. J nii.L- con l.;i i ni ny, two high current f 1 fc t r i i\i I 1 I'i'il (iir oui'.ii.

7: K1 .inju1 I'onl.'i ininj; lliirkncss monitor srnsor ho.i.l. ivurv 8: Stripper foil (2 |jg/cm ) mounted over a 12 mm opening

2 \ y,\\Vi' 9 : S t r ippur foi 1 (2 jjj;/cm") mnnntctJ over n 8 HUM opcMi i n.^. 179

'I'!;!' I'HF OF LAHF.RH FOR TI!K SEPARATION OF ISOTOPES Robert D. i^cAlpine Physical Chemistry Branch Chalk River Nuclear Laboratories Atomic Energy of Canada Limited Cli.ilk River, Ontario KOJ 1J0

! .:';'R

GVr.l-'.RAL FFATURES OF LASFR ISOTOPE SEPARATION HCHEMFS Laser Isotope Separation (hereafter abbreviated L1S) is still in its early development stages, and specific problems have yet to be solved in most cases. However certain inherent features of these techniques make them attractive for development work. Some of these are: ; :'•.".,• '•.'.'>: ••: i::; / • - •• t i::. .;,•;•;.".-;.' . >'. ;'; <- .'.. .•/•.•:. Separa- tion factors comparable to those obtained by mass spec- tronetric techniques are expected. It is conceivable that lor many elements, isotope separation may be carried out in a single step.

.: • '.;•••:; ';..'/,• /V;,1 ••';. ;',,, •.'::.' .••• /\: !•••'•:,•'::.:. Conven- tional techniques treat the bulk of material, including the isotopes not wanted. However LTS selectively treats iiniv those isotopes that arc requj rcO. The physical prir.- (•Ljiics lend to an cnorny roqui.rci^ont of the order of a :<•:•: i.'V por separate',! ato:':. Allowing for laser and process i:.- efficiencies, the overall energy requirement miqht be o f the order of several koV or less, per separatee' atop. '} h i .'•; eornaros with an ennrciy requirement for uranii::r y<'"y~ a rat ion by qas diffusion of the order of MeV per separ.it> d uranium atom.

s< '-pa ra t i on factors and low onorqy costs mi ant lend to relatively 'ow operating! and capital costs. For isotopes such as ! ' 1'', conventional techniques renui re larqe ivanv- <:t.-tae plants. It is conceivable that LJS processes would require fewer stages and hence smaller plants. .' • .•.•.'•• ' ''.-. While it is true that particular M.c i (•(•hni'iues would, have to be developed for each elenetif (and possibly oven for each isotope in some cases) it siHfs likely that some form of T.TR could be found to lie appropriate for each element. n ieneral terms, the follov/incr conditions must he net for a ;••) rent i al !,rs scheme: M) The spectral lines due to the most abundant isotope (s'l rv;st show a qap in a spectral region where there i.s a .-.irenc line due to the molecule (atom) containinq the desired isotope. The initial isotopic selectivity will be det.ermned by the relative absorption intensity of the Iine due to the desired isotope and the baekaround due to el.her isotopes. At the initial abundance levels, this intensity ratio should be as larqe as possible'. 12) A high power source; must be available with a spectral resolution great enough to attain (.1) and which can be i uried to the absorption line due to the desired isotope, f ') A fast process(chemica1 or physical) must be developed r.ueli that the. selectively excited entity, from (1) can be irreversibly and selectively "captured" before it lias a rh.i^ee either to lose its excitation eneniy LO transla- tion.)) rnc;rciy, or worse, to transfer its excitation to a molecule containing an undesired isotope. he emrplex process of transforminu a convenient source of a ^articular isotope into r. usable form, enriched in the desired

:;r.tope, nny )•„. considered accordinq to the five process steps 181

mnu

.". "" • I / ?•'

shown in Fig. 1. The process steps displayed in Ficj. 1 are:

.' .'. -1 .'>'.. ; .•-; ••< • i\- : ! • ' i.'! /:,';; .'•-;.•' .:.,'. •••'•• .•;'•;. This in- volves the selective formation of an excited state of molecules containing tho desired isotope without causing the; formation of excited states involving other isotooes.

'••',' .-'.•.•"•!.•.'.•'..>; -•.:' :;•;•,•.' ••;.";..• ;:;•. .•.'.•.;. In general, the secondary species will consist of radicals, ions or stable products.

: •• •' ••' '../.•:.•' • • ;' ; ."•'.•. ::.;• .;. These include dissociation or ionization brought about lay the absorption of a second photon. They also include selective processes which occur as a natural consequence of '.".', such as photodeflection by momentum transfer and predissociation (or pre-ionization) processes. Certain discrete vibrational levels, of a particular electron state, which / .••• .-• .••• (in ten..- of energy) the dissociation limit of that state, may be weakly coupled t:o a diiiioncratc

of tin; molecule into this ; .•••' •• • '.• ••'.••" - (or • ••• -

• ':: •• ;' •. ) level will lead to a '.-..: •'.-•' •. (or an • '.-:;• ' •. ,• of the mokculo without causinq an ap- j.'i'cc- i able broadoninu of the spectral linowidth. I (iissi ciation (or pre-ionization) processes iray be assisted by collisions. In addition, external ma;;- netic fields may lie used to induce the weak couplin in certain cases.

'•'.• •: ' • •• .' / •-. -• .:.•:• ••. A scavenger is used to selectively react with the excited state formed by :'.'.'. The secondary species are a result of this reaction, find consist of ions, radicals or stable

"'lie secondary species resulting from (;'•) consist of ions, free radicals or stable molecules. These secondary species must be captured and removed from the reaction x.nne before isotope scrambling reaction can take place. Th-.• iipp.cral strategy is to quickly react the secondary :;:•'-<• i o.'5 to ' orm stable products (if they are not already i:-. Hi is form) and to remove these stable products. In a.idilion, ions can be removed by electromagnetic means; lifiv.'ever , 'diarae oxchanqo processes between the ions con- i.iii! in'l Mie desired inutupc and other compounds will cause sci .ipibl i ny of isotopes if romoviil is not done quickly '••noiuili. Free radicals may be involved in chain reactions that result in isotope mixing, hence one must carefully :;•'!('("!• scavengers for these as well. Fvon stable mole- •iie;-., under certain conditions, will undorqo isotope ex- c'viri'tc reactions, alhhouqh LIS processes which lead to :;iab]e rnlecnles for the secondary species have a definite r '•-.-lnt ,i';e. In many resjjects, process (!) is the most dif- : i •" • 111 I st-fp tin achieve for LTS. Many proposed L T S schemes !'. i"'1 ,!''hicve<] ureal selectivity in the formation of the .••• •• •< >!!']<[ ry spoci.es, only to lose this selectivity before •ii;lnii' ,md rcTioval cr>u).d lie carried out. 1.3'i

'.'• '•• I C I ^ I (^ V\FT'R ! SOTOP!~! K ^:'P •V-'.A'T If V .' PR^C! T ;-' i' S

In all of the processes which are surnariz'"! in F:::. 2,

tl'.r initial sclf-ction of the required isotope is .-ichi >-vod by

the sf 1 pet i •..-(•• formation of an. o:-:ci ted f?l ect roni c or vi bra t ion "i i

stato of or.lv those ;-o tecu ] os tor atons) which contain the- '••'•>-

%\TQ<.\ LSO tope '''!•." various I,IP schor.os vary accordi n:c.i tcii state is tr.msfor:".-.:

into a stable product (procosr, .stops .' and •' of Fiu. I). I:. nany cases, tiiero are analoqous mol'.-culnr fM) and at'o:;ic- ('•.) processes. Jn all cases relaxation and enerqy transfer processes from the excited state compete with I,IS. Some of the L1S processes are briefly described below:

(1) Selective two step processes. Those processes include A-l, M-l and M-2 (Fig. 2). Two photons are used t<-. give a selective step-wise dissociation or ionization. The energies of the two photons (h'.i and h . .- respectively) are restricted by:

hv, + h... -• Fi H )

h. , , hv, •• Ki (2)

where F.]_ is the low energy limit of dissociation or ionization. Photon h,: selectively populates a sharp vibrational or electronic level which shows an isotope shift. Photon h. .- provides the photophysical (process •••"; Fig. 1) formation of the secondary species. (2) Fxcited state chemical reactions. These processes include h-2 and M-3. A laser is used to selectively form an excited v.ibrational or electronic state of the molecule containinq the desired isotope. A scavenger is sought such that it reacts at virtually every collision with the excited state, hut not with the ground state containing the i: ">fiesi red isotope (s). This scavenger must successfully conpote with cnerqy transfer and relaxation processes from the excited state. The scavenger provides a choi" i (\i I formation of the secondary species (process .'".•', Fiq. 1) and takes the place of the second photon in the nrovious scheme. 134

~<»\l -^OL H 0 L I i ':• L I l> P P Q C t S r: L S

A" • e »

/v,v...

'' C

A - 3 DEFLECTED

t ! SELECTIVE T»0 STEP I ON IZAT I ON «-l SELECTIVE T*O STEP DISSOCIATION

.1. ELECTRIC FIELD ASSiSTED i.i; USING AN INTERMEDIATE VIBRATIONAL STATt (In USING AN INTERMEDIATE ELECTRONIC STATE A ? CHEMICAL SCAVENGING OF EXCITED ATOM »i-2 SELECTIVE TWO STEP DISSOCIATIVE IGKi2ST;0N . .'.- [.ECThON TRANSFER REACTIONS (.1) USING AH INTERMEDIATE VIBRAliONAL STATE

* :' PHr.TDCIfLFCTICN BY MOMENTUM TRANSFER 1 ill) USING A ! INTERMEDIATE EUCTRONIC STATE

M-3 CHEMICAL REACTIONS IN EXCITED STATES.

(ai EXClTED VIBRA1IONAL STATE

It) FXC:TED ELECTRONIC STATE

III-A SELECT I V. PREDI SSOCI AT! ON

(.11 SPONTANEOUS (111 COLLISIONALLY ASSISTED Id MAGNETIC FIELU INUUCED

M-5 SELECTIVE PRE - IONIZATION

is) SPONTANEOUS

{til COUIS'ONAHY ASSISTED

ic> FIELD INDUCED

".'•: J '. i" ' .' is:

(3) Selective predissociation or pre-ionization. Pro- cesses M-4 and M-5. The laser selectivelv populates a relatively sharp predissociation (or pre-ionization) level of the molecule containing the required isotope. Second- ary species are formed as a result of the predissociation (or pre-ionization) from this level.

(4) Photodeflection by momentum transfer. Process h-'i. The transfer of momentum during the selective absorption and subsequent emission of light, is used to deflect the desired isotope; out of an atomic beam. Work on this process has only recently(14-16) commenced.

IV REVIEW OF STUDIES ON VARIOUS SPECIFIC ISO^TPES It is useful to briefly list some cf the. isotopes that have been considered to date for LIS. (1) U :' "' ' The scientific feasibility of separation by process A-l has been shov.'n on a very small scale' '' but no commercially viable process has yet been revealed.

(2) ir_ Compared to other isotopes, conventional methods for deuterium recovery are very cheap. This is a difficult challenge for I.IS. Work has been done on selective predissociation ^-9) (M-4b) and photochemical separation'20) (M-3a). (3) Hg_ (Various Isotopes) A commercially viable process has been v/orked out'^) based on the use of single isotope mercury resonance lamps as a source (process A-2). Single stage separation factors better than 3.3 have been attained, and certain isotopes can be prepared to greater than 90% isotopic purity in a single step. Lasers might be used to improve this process for certain isotopes.

(4) Br7q and Bre' A selective predissociation of Br? (M-4a) has been achieved. The resulting isotopically specific bromine atoms are scavenged by HI to produce isotopically specific HBr (v = 0, 1, 2). The isotopic selectivity has been confirmed by observing IR fluorescence from HBr; but no real attempt has yet been made to remove this HBr from the 186

reaction mixture(24) _

(5) C1 \ C1", Q1 7 and o' B The feasibility of LIS for the ahove isotopes has been demonstrated by selective predissociation '-'-- '22) (process M-4b).

(6) Rb Isotopes Separation is feasible by selective two step ionina- tion (A-l) (23) and by photodeflection '16'.

(7 ) N i i, Separation is feasible by selective two step dissociation '2A) (M-2a) and selective predissociation >") (M-4

(8 ) Ca^ Isotopes Preliminary studies on selective two step ionization (A-l) have been reported (27)_

'9) B Isotopes Preliminary studies on selective two step dissociation (M-la) have been reported^ J.

CONCLUSION Laser methods offer the potential of cheap large scale separation techniques for many different isotopes. These techniques are still in their development stages, and many problems have yet to be worked out before this potential can be realized. Processes such as M-la, M-2a and M-3a suffer from the current lack of high power tunable lasers in the IR region, 2-5:. Improvements in the power and efficiencies of dye lasers in the visible region are needed for processes A-l, A-2 and A-3, M-lb, M-2b and M-3b. In addition, new high power UV lasers are required to produce the second photon for selective two step processes. A good deal of fundamental work is needed to under- stand the reactions between scavengers and molecules with excited electronic and vibrational states. As well, the various relaxation and energy transfer processes which compete with the "capture and. removal" .stages of LIS need more fundamental study. In spite of these problems, there is already at least one element, mercury, for which photochemical isotope separation is commercially feasible. 1S7

At this time, deuterium and uraniurn-2 35 are the only two isotopes which are user5 on a larae scale commercially. However, the availability of large quantities of different isotopes will lead to nary new application for pure isotopes in chemistry, physics and medicine.

VI REFFRFNCFS

1. T.R. Merton and !!. Hartley, Nature 105, 104 (1920).

2. H. Hartley, A.O. Ponder, r.J. Bowen and T.R. Merton, Phil. Mag. XLTII, 430 f1Q22).

3. S. Mrozowski, Bull. Acad. Poland I_, 464 (1930).

4. S. .Mrozowski, z. Physik. 7_3, 826 (1932).

5. K. Zuber, Helv. Phys. Acta, 8_, 488 (1935).

6. K. Zuber, -lature 1_3_B, 796 (1935).

7. K. Zuber, Helv. Phys. Acta 9_, 285 (1936).

8. H. Kuhn, H. Martin, K.H. Eldau, Z. Phvs. Chen. Abt. B21, 93 (1933).

9. W. Kuhn, H. Martin, K.H. Fldau, ?,. Phys. Chen. Abt. B5 0, 213 (1941).

10. G. Liuti, P. Dondes, P. Harteck, J. Chen. Phvs. £4_, 4052 (1966). 11. H.E. Gunning and 0. Straus?, Adv. Phys. Chen. 1., 209 (196 3) and rp.ferences therein.

12. R.L. Farrar Jr. and D.F. Smith, "Photochemical Isotope Separation as Applied to Uranium", report K-L-3054, Oak Ridge, Tenn. (1972).

13. V.S. Letkhov, Science 180, 451 (1973).

14. A. Ashkin, Phys. Rev. Lett. 2_5_r 1321 (1970); A. Ashkin, Scientific American 226, P3 (1972).

15. I. Nebenzahl and A. S^oke, Appl. Phys. Lett. 2_5, 327 (1974).

16. A.F. Bernhardt, D.E. Duerre, J.R. Simpson and L.L. Wood, Appl. Phys. Lett. Z5' 617 (1974).

17. R.H. Levy and G.S. Janes, U.S. Patent 3,772,519 (1Q73).

18. S.A. Tuccio, J.W- Dubrin. O.G. Peterson and B.B. Snavely, "Two Step Selective Photoionization of U235 in Uranium Vapor", paper presented to VIII International Conference on Quantum Electronics, San Francisco (1974).

lg. F.S. Yeung and C.B. Moore, J. Chen. Phys. 5_8, 3988 (1973). 188

20. S.W. Mayer, M.A. Kwok, R.W.F. Gross and P.J. Soencer, Appl. Phys. Lett. 17_, 516 (1970).

21. S.R. Lfinnc and C.B. 'loore, Phys. Rev. Lett. 22' 269 (1974).

22. E.S. Yeunq and C.B. Moore, Appl. Phys. Lett. 2_1, 109 (1972). 23. R.V. Ambartzumian, V.N. Kalinin and V.S. Letokhov, Sov. Phys. JFTP Lett. 1J3, 21? (1971).

24. R.V. Anbartzumian, V.S. Letokhov, G.N. Makarov and A. Puretskii, Sov. Phys. JFTP Lett. 1_5, 501 (1972).

25. R.A. Back, private discussj.ons.

20. ::.G. Basov, E.M. Belenov, L.K. Gavrilina, V.A. Isakov, K.P. Markin, A.M. Oraevskii, V.l. Romanenko and ?-:.B. Ferapontov, Sov. Phys. JETP Lett. 1_9_, 190 (1974).

27. U. Brinkmann, W. Hartig, H. Telle and H. Walter, "Separa- tion of Calcium Isotopes Using Stepwise Fxcitation by Lasers", paper presented to VIII International Conference on Quantum Electronics, San Francisco (1974).

28. S.D. Rockwood and W.W. Rabideau, "Boron iFotope Separation by Laser Induced Photochemistry", paper presented to VIII International Conference on Quantum. Fiectronics, San Francisco (1974). LIST OF DELEGATES

ALEXANDER, T.K.. Canada

ALLEN, K.W. Oxford, Enjlcr-id

BALL, <-.C.

BEAl'OIA.MP, R. •'. L.nMi, Canada

BECKER, R.J. j. t 01' :'; ;. •;, :,..:. , ,V-\M YCJ-'K, U.S.A.

BEER, O.A. i'iH-''<-:>:'i t.i, of V '.c',oi'ia.

• •- : * -.,/,-, T , T : SOUCHARD, C.A. • ..'; •. ~'ij: <-• • 1.-7 ijC<. .'L>. i

Qucbco, P.O., r.,KUJc RROI/DE, C. W- ir.raun Ini-iiiutc

BURN, N. h-i'.k River \Utr-lcLir Ui: oratories Ccmada

CANCE, '•'.onivoiiqe, Fravj^.

CARDINAL, C. f

CHUNG, A.H. i'>'ivci"'ity of Pofonto 'Toronto, Oi:i<.i2>'io, CiVia'^c

CONNER, W.V. Pov C'ncT'.ioa'i Co. Golden, Colorado, U.S.A.

DAVIES, W. Chalk River .htelcav Laboratories

DIAMOND, W.T. Chalk River Nuclear Laboratories Canada

DRESSER, J.A. Research Chemicals; Phoenix, Arizona, U.S.A.

EVANS, J.E. Lam-cnce LiVermore Laboratory Livermore, California, U.S.A.

FIEGENBAUI'i, I L. Upton, L.J., JVVJU York, U.S.A.

EORSTER, J.S. Chalk River Nuclear Laboratories Canada FOULGER, P.H. Vayian Associate;: of Cmjucju. Ltd. Georgetown, Or-.iario , Canada

GALBRA1TH, D. Oak Ridge National Laboratory Oak Fidae, Tennessee, U.S.A.

GALLANT, J.L. Chalk River iluc'car Laboratories Canada

GE1CER, J.S. Ch.alh Rive/' U\Anlcrw Lu:.^r^rorlec Canada

GRAHAM, R.L. Chalk Bioei' Uu\:i<:.ar Lo}^rat,ri>.a Canada

GRENIER, Co''3rris(;ana!- a 1 'Er.ci'jt'i Aronioiiv Mo n tpou

HARDY, J.C. Chalk Rivet1 Hncl^w L,;.i oraiovici Canada

HAUSSER, 0. Clialk River Nuclear uxi^oratorie;: Canaan

HEAGNEY, J.M. liisivo !'atie~r Ccr^anu Seattle* ISasttinjiCK, .','.."..;.

HEAGNEY, J.S. Kioro !JattsP Coritpami Seattle, 'dashiivyton, U.C.A. HIGH, M. Laoal Ur.i'JePsity Quebec City, P.Q., Canals:

JOHNSON, P. Laurence Liverwort? Laboratory Livermore, California, U.S.A.

KARASLK, F.J. Ai-gonne Natiorjal La: oratory Avgonr.e, Illinois, L.L.A.

KAST, J. John Hoykins University Baltimore, rid., U.S.A.

KI.ESER, I. University of Toronto Toronto, Ontario

KIL1US, L. University of Toronto Toronto, Ontario, Canada

KLEIN, R. Rutgers Unioersi tu Hew brunsuic-k, !\!.J., U.S.A.

KORISK, E.H. Oak Sidue national Laboratory Oak Eidje, Tennessee, U.S.A.

KORN, G. University of Rochester Roches ter, /'/. ¥., IJ. S.A.

KWXNTA, J. Comnissariat a I'Energie Atom.que Paris, France LAURIER, R. Cana..:iar: Sari-KCi.:

LEITE, R.J. .-.'."-•• r^.r.icK IK-. AKK Ar: or, /-.'v /•. ' :an, '

LOUGHEED, R. '..-.i.-f.-y-.. L'-.\:r::.:•'•' LacOi-t

LOVE, L.O. ••'•. .••-.'•;.:• f: .\':2i\cn.;l :.•:: oy:

LOZOWSKI, B. //...••.•^•'1J i';::' :tv..-v .^r-

MAIER-KOMOR, P. I"...,-•.-.. V;.,;.'.-.".•'i.>: U^h-ht

MacDOMALD, J. :'•.•: I < r, I >.:)' .V,; -lc::.r ',:i.:.. ?

McALPINE, R.O. ^:;;;:' :-::•,:/' U..-JI, •::.:• :,..-l::

McDANIEL, E. ..•;-: r.idjc U.ii^Kji Lai a-

MCDONALD, A.B. S--JZ> .HJ^I- ;;^;>Q>- uai^z

MERRITT, J.S. C.cAk Siva- .W-.^-rar Said Canadu

MITCHELL, I.V. C'laV-: l-.ivev UUCUCJJ- Ualoi

OLMSTEAD, W.J. "'..a I'- Ris;y .V,-.;•.",•.•"." l-i:. .:•

PECK, A.A. iVfi's 'Jrdvcr-cA r •

PERRY, W.L. S>.alk Finer i-i^l--j_i- Lalui-

POVELITES, J.G. Lea ALunos Scientific Labcrarcy:

PRIEBE, A. Aflee Engineer: >!J limized

RABY, B. AiUCL, Canr.eroiai Lvodiici,-

RAMSAY, D. Stanford Uiiiverriiv

RIEL, W.D. Sz~ie Univeraiij of i.'m? York StOKU 3i'CC\, L.I., ,':'. Y. , U.S. RITTER, E. Brookhaven national Laboratory Upton, L.I., lieu York, U.S.A.

SAETTEL, M.A. Laboratoire de C^ectrc.ieirie Uu.aleait'e Strashoiwg, France

SANTRY, D.C. Chalk River Hucleat' Laboi'uioriec Canada

SCHMEING, H. Chalk Rive'/- .'raolsar LaLcsruiorics Canada

SiUN, V.M. University of CaakatS-ievuK Saskatoon, Sisk., Cunado

SLETTEN, C. liiels BchP Inatituct Ristf, Benmar-k

SLOBODIAN, J.T. A.ECL, Cowr.erciai Pi'o'ujia Ottawa, Canada

ST1NSON, J.D. national Research. Council Ottawa, Canada

STOREY, R. national Research Council Ottauxz, Canada

THOMAS, G.E. Argonne National Laboratory Argonne, Illinois, U.S.A.

THOMPSON, D.M. University of Minnesota. Minneapolis, Minnesota, U.S.A.

TOONE, R.J. Chalk River Nuclear Laboratories Canada

VAX AUDENHOVE, J. Bureau Central dc Mesufee Huoleaives Geel, Belgium

WALKER, R.B. Chalk River Nuclear Laboratories Canada

UUSHAAR, M. Laboratoire de Physique iiuc-leaire Strasbourg, Prance > X

c c X > -^

The International Standard Serial Number

ISSN 0067-0367

has been assigned to this series of reports.

To identify individual documents in the series we have assigned an AECL—number.

Please refer to the AECL-number when requesting additional copies of this document from

Scientific Document Distribution Office Atomic Energy of Canada Limited Chalk River, Ontario, Canada

KOJ 1J0

Price S9.00 per copy

642-76