INFORMATION TO USERS

This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.

The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.

4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.

Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 P':* tevi 1 74-11,071 WIELAND, Bruce Wendell, 1937- DEVELOPMENT AND EVALUATION OF FACILITIES FOR THE EFFICIENT PRODUCTION OF COMPOUNDS LABELED WITH CARBON-11 AND OXYGEN-15 AT THE WASHINGTON UNIVERSITY MEDICAL CYCLOTRON.

The Ohio State University, Ph.D., 1973 Engineering, nuclear

i. University Microfilms, A XEROX Company, Ann Arbor, Michigan

1

THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. DEVELOPMENT AND EVALUATION OF FACILITIES FOR THE

EFFICIENT PRODUCTION OF COMPOUNDS LABELED WITH CARBON-11 AND

OXYGEN-15 AT THE WASHINGTON UNIVERSITY MEDICAL CYCLOTRON

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Bruce W. Wieland, B.S.

The Ohio State University 1973

Reading Committee:

Professor Donald D. Glower, Chairman Professor Robert F. Redmond Professor Walter E. Carey

Approved by

Advisor, Department of Nuclear Engineering TO THE MEMORY OF MY FATHER

IN RECOGNITION OF HIS CONTRIBUTIONS AS AN EDUCATOR

ii ACKNOWLEDGMENTS

My parents, Stephen and Mathilda Wieland, have provided me with a continuing respect and personal motivation for the education process through the example of their two lifetimes of productive service in the Iowa public school system. My pretty wife, Edna, while pursuing her own career in counseling and teaching, has helped me through several troubled times during the course of my dqctoral program. She has also applied her expertise in English to the careful reading and revision of this dissertation. Dr. Michel Ter-Pogossian of Washington University supplied much appreciated guidance, patience, and financial support while directing the dissertation research project in his Division of Radiation Sciences at the Edward Mallinckrodt Institute of Radiology. Dr. Donald Glower of The Ohio State University has been unwavering in his encouragement and support as advisor and as sponsor of my NIH Special Research Fellowship throughout the entire program. Dr. Robert Redmond of The Ohio State University has provided thoughtful assistance during the coursework portion of the program, and has coordinated the dissertation phase. Dr. William Hunter of The Ohio State University supplied recommen­ dations and assistance with regard to physiology, nuclear medicine, and medical cyclotron applications during the coursework phase. He was also helpful in applying for and maintaining NIH support. Dr. Michael Welch of Washington University and Dr. Walter Carey of The Ohio State University made valuable suggestions after a careful reading of the dissertation. John Hood at the Washington University Physics Cyclotron was a source of frequent and friendly technical advice on medical cyclotron problems. Lee Troutt and William Margenau, operators of the Washington Uni­ versity Medical Cyclotron, were a source of congenial and capable assis­ tance in performing experiments, and made numerous helpful suggestions during the research project. Julius Hecht of the Division of Radiation Sciences deserves thanks for technical assistance and careful preparation of the dissertation figures. Susan Hartner of the Division of Radiation Sciences accomplished the typing and proofreading of the draft and final copies of the dissertation in an extremely capable, pleasant, and much appreciated manner.

ill I am grateful to Dr. John Eichling, Dr. Michael Phelps, Dr. Michael Welch, Dr. Marcus Raichle, Dr. Robert Grubb, Dr. Michael Loberg, Dr. Kenneth Larson, Dr. Edward Hoffman, Dr. Judith Metzger, Maria Straatmann, Katherine Anderson, Nizar Mullani, Carol Coble, and Ruby Hicks, for their friendship, advice, and technical assistance during my stay in the Division of Radiation Sciences. The National Institutes of Health provided financial support for this program with Special Research Fellowship GM-41730 from September 18, 1968, through October 24, 1971, and from December 6 , 1971, through March 28, 1973. Partial support was also provided by NIH Grant No. 5 P01 HL13851 from May 1, 1971 through August 31, 1973.

iv VITA

15 April 1937 Birthplace - Carroll, Iowa

1955 - 1959 Engineering Trainee, John Deere Waterloo Tractor Works, Waterloo, Iowa

1960 B.S., Mechanical Engineering, Iowa State University, Ames, Iowa

1960 - 1966 Engineer and Nuclear Analyst, Oak Ridge National Laboratory, Oak Ridge, Tennessee

1961 Instructor (2nd Lt.)* U.S. Army Nuclear Power Field Office, Fort Belvoir, Virginia

1965 Postgraduate Certificate, Oak Ridge School of Reactor Technology, Oak Ridge, Tennessee

1967 - 1968 Research Engineer, Battelle Memorial Institute, Columbus, Ohio

1968 - 1973 NIH Special Research Fellow, Department of Nuclear Engineering, The Ohio State University, Columbus, Ohio

1971 - 1973 Research Associate, Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, Missouri

v PUBLICATIONS

Production of short-lived isotopes by charged particle accelerators, M.E. Phelps and B.W. Wieland, Physics in Medicine and Biology, Vol. 18, No. 2, 284-286, March, 1973.

Techniques for irradiating high temperature materials in a steep flux gradient, G.W. Keilholtz, R.E. Moore, M.F. Osborne, B.W. Wieland, and A.F. Zulliger. Proceedings of the May 1966 International Symposium on Capsule Irradiation Experiments, Pleasanton, California.

Analysis of the High Flux Isotope Reactor fuel element shipping cask (Shielding analysis, Provisions for criticality control, Provisions for ruptured fuel elements), B.W. Wieland. ORNL-TM-959, January 1965.

Temperature distributions in fuel samples irradiated in Engineering Test Reactor lattice position J-12, 6 .H. Llewellyn and B.W. Wieland. ORNL-CF-64-5-60, May 1964.

Review of stress calculations for Tower Shielding Reactor II (Shield and supports, Vertical scanning components), B.W. Wieland. ORNL-CF- 63-5-30, May 1963.

vi FIELDS OF STUDY

Major Field: Nuclear Engineering

Studies in Nuclear Physics: Professor H.J. Hausman

Studies in Nuclear Detection Instrumentation: Professor R.A. Krakowski

Studies in Pulse and Digital Electronics: Professor D.W. Miller

Studies in Mathematical Statistics and Biomedical Computer Applications: Professor T.A. Willke

Studies in Physiology and Nuclear Medicine: Dr. W.W. Hunter, Jr.

vii TABLE OF CONTENTS

Page DEDICATION ii

ACKNOWLEDGMENTS iii

VITA v

PUBLICATIONS vi

FIELDS OF STUDY vii

LIST OF TABLES x

LIST OF FIGURES xli

Chapter

1. INTRODUCTION 1

1.1 Scope of Research Objectives 1 1.2 Utilization of the Washington University 3 Medical Cyclotron 1.3 In Vivo Studies of Cerebral Circulation n and Metabolism 1.4 Assessment of the Requirements for Labeled 16 Compounds

2. CYCLOTRON PERFORMANCE AND RELIABILITY 21

2.1 Characterization of the Cyclotron Beam 21 2.2 Development of a Gas-Cooled Beam Window 26 Assembly 2.3 Modifications to Cyclotron Components 32 2.4 Future Work 37

3. TARGET AND LABELING SYSTEMS 42

3.1 Development of a New Target for the 42 Production of H e

viii Page 3.2 Investigation of a Photosynthetic 58 Process for Labeling Glucose with C 3.3 Design of a Hc/15o Tandem Target Assembly 70 and Associated Labeling Systems 3.4 Future Work 88

4. CONCLUSIONS 97

4.1 Evaluation of Equipment and Procedures 97 Implemented to Achieve Efficient Production 4.2 Applications to Other Cyclotrons 100 4.3 Future Work 104

APPENDICES

A. Procedures for In Vivo Studies Utilizing Compounds 107 Labeled with and ^ C

B. Beam Energy Measurements 111

C. Beam Mapping Measurements 114

D. Prototype Beam Window Development 130

E. Parameter Study of Uptake of ^ C 02 by Swiss Chard 141 Leaves

F. Design of Gas Chamber for the Production of *^0 161

G. Range and Energy Loss for Deuterons in Various 166 Materials

H. Yield Calculations for ^®B(d,n)^C and ^N(d,n)^0 170

LIST OF REFERENCES 174

ix LIST OF TABLES

Table Page

1-1 Research Activities of the Division of Radiation 2 Sciences

1-2 Physical Properties of ^C, ^ 0 , an(j 18p 5

1-3 Specifications of the Washington University 7 Medical Cyclotron

1-4 Time Distribution of Current Medical Cyclotron 10 Activities

1-5 Labeled Compounds for In Vivo Studies 13

1-6 In Vivo Studies with ^ 0 14

1-7 Labeled Compounds Developed for Proposed 17 In Vive Studies

1-8 Criteria for the Efficient Production of 19 Compounds Labeled with -^C and ^ 0

2-1 Comparison of Several Methods for Determining 23 Deuteron Beam Energy

2-2 Criteria for Developing a New Beam Window 28 Assembly

2-3 Attenuation of 6 MeV Deuterons by Various Materials 29

3-1 Properties of Boron Oxide 44

3-2 Evaluation of Screen Material and Sizes 45

3-3 H-C-Glucose Production Parameters 63

3-4 Current ^C-Glucose Production Data 64

3-5 ^C-Glucose Production Data Exclusive of 65 Radioactive Decay

x Table Page 3-6 Detector Systems Available for ^C-Glucose 67 Studies

3-7 Predicted Count Rates for ^C-Glucose Studies 68

3-8 Target and Labeling System Volumes 79

4-1 Summary of Methods Implemented to Meet Requirements 98 for Efficient Production of Compounds Labeled with n C and 150

4-2 Activity, Time of Production, and Reliability for 99 Compounds Labeled with ^ C and -^O

4-3 Summary of Proposed Future Work 106

C—1 Films and Developers for Quantitative 119 Autoradiography

D-l Data for Havar Alloy 132

D-2 Comparison of Silicone Rubber Bonding Materials 134

D-3 Comparison of Silicone Rubber Dlgestors 138

xi LIST OF FIGURES

Figure Page

1-1 Floor Plan of the Cyclotron Facility 8

1-2 Path of the Deflected and Extracted Beam 9

2-1 Mean Range of Deuteron Beam in Aluminum 22

2-2 Beam Density Contour Map 24

2-3 Details of Current Beam Window Assembly 30

3-1 Prototype ^ C Screen Target 47

3-2 -^C Section of Tandem Target System 49

3-3 Current ^ C Screen Target 51

3-4 Irradiated Target Rod 53

3-5 Excitation Function and Yield for l®B(d,n).^C 56

3-6 Production of ^C-Glucose 61

3-7 Side View of Tandem Target Assembly 73

3-8 Top View of Tandem Target Assembly 74

3-9 Front Exploded View of Tandem Target Assembly 75

3-10 Piping Diagram for Target and Labeling Systems 80

3-11 Shielded Labeling Cave Equipment 83

3-12 Tandem Target System Installed - Front Section 85

3-13 Tandem Target System Installed - Rear Section 86

3-14 Glass Target Chamber 91

C—1 Beam Mapping Techniques 115

C-2 Film Calibration Curves for Quantitative 120 Autoradlography xii Figure Page

C-3 Densitometer and Beam Density Curves for a Vertical 122 Profile

C-4 Evaluation of Beam Patterns on White Paper 123

C-5 Evaluation of Beam Patterns on Polaroid Prints 124

C-6 Evaluation of Beam Patterns on Blue Cellophane 125

C-7 Evaluation of ^®A1 Autoradiograph Beam Patterns 126

C-8 Evaluation of Autoradiograph Beam Patterns 127

D-l Original, Prototype, and Current Beam Window 131 Assemblies

E-l Hot Lab Area Used for Prototype Development 142

E-2 Emission Spectra for Various Light Sources Compared 147 to Relative Effectiveness in Photosynthesis

E-3 Swiss Chard Plant 157

E-4 Uptake of by Swiss Chard Leaves 158

F-l 150 Yields from Irradiation of NH4NO3 Cylinders 163 in Air

xiii CHAPTER X

INTRODUCTION

Section 1.1 Scope of Research Objectives

This dissertation presents the results of theoretical and experi­ mental studies performed in conjunction with the design, fabrication,

and operation of medical cyclotron facilities for the efficient pro­

duction of compounds labeled with ^ C and ^O. These studies were

performed at the Washington University Medical Cyclotron, which is

operated in support of research projects carried out by the Division

of Radiation Sciences at the Edward Malllnckrodt Institute of Radiology.

Table 1-1 depicts the organizational relationship of this group to

Washington University, and indicates the research activities in which

it is currently engaged.

An investigation of the requirements for short-lived isotopes was

completed, and modes of production adequate to satisfy present needs

for ^ C and ^ 0 were developed and implemented. Efforts were made to

design new systems with sufficient flexibility to adapt economically

to predicted future needs. Performance of the new equipment in routine

use was evaluated and applications to other cyclotron facilities were

examined. A continuing program of improvements in production and re­

liability is suggested in anticipation of continually growing require­

ments. . 2

Table 1-1

RESEARCH ACTIVITIES OF THE DIVISION OF RADIATION SCIENCES

WASHINGTON UNIVERSITY - St. Louis, Missouri SCHOOL OF MEDICINE MALLINGKRODT INSTITUTE OF RADIOLOGY DIAGNOSTIC RADIOLOGY NUCLEAR MEDICINE RADIATION ONCOLOGY (Therapy, Cancer Biology) RADIATION SCIENCES In vivo isotope studies of regional cerebral metabolism and hemodynamics in animals and patients utilizing H2^ 0 . 15o-hemoglobin, C ^ 0-hemoglobin, Hc-glucose, and 1%-ammonia In vivo studies of regional cerebral blood volume by stimulated x-ray fluorescence Nuclear instrumentation development Laboratory computer and interfacing development Mathematical modeling of cerebral metabolism and hemodynamics Research chemistry Production and analysis chemistry Cyclotron, targetry, and hot lab operation and development 3

The remainder of Chapter 1 will be devoted to a discussion of

the importance of short-lived cyclotron produced isotopes in medical

research, a description of the facilities and research projects at

Washington University, and a summary of the criteria to be satisfied

in order to accomplish adequate and efficient production of compounds

labeled with and ^0. Chapter 2 covers measurements of the cyclo­

tron beam characteristics used to design targetry and evaluate cyclotron performance, and efforts devoted to achieving a reliable extracted beam with increased current and energy. Chapter 3 describes the development

and implementation of new target and labeling systems for the produc­

tion of compounds labeled with ^ C and It also outlines methods

to extend these systems to meet future needs and accommodate the pro­

duction of additional labeled compounds. Chapter 4 is an evaluation

of the equipment and procedures implemented, a discussion of applica­

tions to other accelerators, and a summary of the recommendations for

future improvements at the Washington University Medical Cyclotron.

Chapters 2 and 3 are intended to convey sufficient background and detail

to permit the future work suggested in Chapter 4 to be carried out

efficiently.

Section 1.2 Utilization of the Washington University Medical Cyclotron

The cyclotron-produced short-lived nuclides 11C, 13N, 150, and 18F

are of particular interest in medical research because they can provide

photon-emitting labels for molecules containing carbon, nitrogen, oxygen, 4

and hydrogen (*®F and groups containing can be chemically substitu­ ted for hydrogen in some cases). This offers the potential of tracer molecules which can be detected externally and used for in vivo study of the wide variety of metabolic processes which consist of chemical 11 13 reactions involving these elements. Physical properties of C, N,

150 , and are given in Table 1-2. There are no longer-lived iso­ topes of these elements with the exception of -^C, which does not decay by a process involving photon emission. Advantages of compounds labeled with these short-lived isotopes include high initial counting rates for in vivo studies coupled with low total radiation dose deliv­ ered to the subject, frequent repetition of measurements if desired, possibility of coincidence detection of the annihilation radiation, and the possibility of in vivo subtraction techniques by using differ­ ent combinations of labeled compounds which all emit the same photon energy (identical detection efficiency if labeling the same physiologi­ cal space). Disadvantages include the inability to study metabolic processes occurring over long time spans, the requirement for rapid synthesis of the label into the desired tracer compound, and the logistical problems in producing and delivering the labeled compound.

A recent publication by Wolf et al (1) treats the synthesis of labeled compounds utilizing short-lived isotopes and references recent reviews and symposia dealing wholly or in part with cyclotrons and their applications in medicine (2»^»j4,_5,_6,_7,^). An early article by

Ter-Fogossian (9) gives a good general survey of considerations 5

Table 1-2

PHYSICAL PROPERTIES3 OF 1:LC, 13N, 150, AND 18F

Property 13n 150 18f

Half life, min 20.3 10.0 2.067 109.7

Positron emission: % disintegration 99.8 100 100 97 Endpoint energy, MeV 0.980 1.19 1.70 0.633 Mean energy, MeV 0.3942 0.4880 0.7206 0.2496

Electron capture: % disintegration 0.19 0 0 3

Annihilation radiation: Energy per photon, MeV 0.511 0.511 0.511 0.511

Journal of Nuclear Medicine Supplement Number4, Pamphlet 6 , March, 1970, Volume 11, and Journal of Nuclear Medicine Supplement Number 2, Pamphlet 4, March 1969, Volume 10. 6

encountered in selecting a medical cyclotron. The dozen or so hospital installations in the world termed "medical cyclotron facilities" in the above publications are supplemented by an undetermined number of accel­ erators which were initially designed for nuclear physics applications, and which are now used in part or full-time isotope production for medical applications. Choosing an optimum accelerator for applications in biology and medicine is not an easy task. Wolf reports that an ar­ ticle dealing with these considerations is in progress and and will be published soon by the IAEA (10).

Basic operational parameters for the Washington University Medical

Cyclotron are given in Table 1-3. Figure 1-1 Illustrates the equipment arrangement in the space occupied by the cyclotron facility and Figure

1-2 shows the details of the beam path. This cyclotron was the second one in the world to be installed in a hospital, the first being a large machine at Hammersmith Hospital in London. An internal beam was achieved in August 1964. A deflected beam of 6-8 pA was extracted through a 25.4 p (0.001 in.) aluminum window in June 1965. The produc­ tion of short-lived isotopes for in vivo studies has continued uninter­ rupted since the availability of the external beam. Current routine in vivo isotope studies include the measurement of cerebral blood flow

(CBF), cerebral blood volume (CBV), and cerebral oxygen metabolism

(CMO2) with l^O-labeled compounds. ^C-glucose and ^%-ammonia are also used extensively for the measurement of cerebral metabolism.

Table 1-4 presents the allocation of cyclotron time with regard to 7

Table 1-3

SPECIFICATIONS OF THE WASHINGTON UNIVERSITY MEDICAL CYCLOTRON&

Type: Fixed frequency, uniform magnetic field Overall Dimensions: 1.83 m by 2.74 m by 4.88 m Space for Accelerator and Power Supplies: 7.32 m by 4.27 m by 2.74 m Weight: Cyclotron and oscillator: 25,538 kg Power supplies: 4,445 kg Control console: 680 kg Utilities: Power: 150 kVA, 440 V . 3 phase 60 Hz Cooling water: 230 liters/mln Magnet: Pole diameter: 91.44 cm Pole tip diameter: 83.82 cm Mean field at exit radius: 14 kilogauss Maximum current in conductor: 650 A dc power required: 40 kW Current regulation accuracy: 1/10 Oscillator System: Type: self-excited Operating frequency: 11.7 MHz Number of dees: 1 (180 degrees) Dee to ground peak voltage: 50 kV Available dc power at rectifier: 50 kW Deflector maximum voltage: 75 kV Ion Source: Type: hooded, hot cathode Maximum filament current: 200 A Filament power supply maximum voltage: 6 V Performance: Deuteron energy: 6.2 MeV Deuteron external beam stable operation: 40 pA Maximum measured: 80 pA Beam size at exit port: 2 mm x 40 mm (approx.)

aManufactured and installed by Allis-Chalmers Manufacturing Company, Milwaukee, Wisconsin. 8

MAGNET TORAGE POWER WORK BENCH SUPPLY

SHOP

OSCILLATOR WALL POWER SAFE TV MON, SUPPLY

CIRCUIT CONTROL BREAKERS C O N 80LE STAIRS 0 0 WN

SUPPLY CYCLO. CABINET VAULT DESK VACUUM PUMP8 non WATER WINDOW CYCLOTRON CONTROL ROOM .ENTRANCE. TV CAMERA STORAGE OSCILLATOR STORAGE TARGET STORAGE AREA TOILET RABBIT LINE AND CONTROLS OO rrxr SINK VENTED GAS BOTTLES' HOOD FOR TARGETRY, w e l l 'T- COUNTER SCALE

HOT LAB VENTED I LABELING CAVE I J CHEMISTRY WORK BENCH

Figure 1-1 Floor Plan of the Cyclotron Facility

\ \ VACUUM CHAMBER

ION SOURCE

itM*

BE*** VM** b e ***

Figure 1-2 Path of the Deflected and Extracted Beam 10

Table 1-4

TIME DISTRIBUTION OF MEDICAL CYCLOTRON ACTIVITIES FOR THE YEAR FROM DEC. 1, 1971 TO NOV. 30, 1972

150 for cerebral blood flow, blood volume, and oxygen metabolism studies 13% 11 i s C and J0 for cerebral glucose metabolism studies 27%

^ N and -^O for ^NH- metabolism studies (first animal study 9/25/72) 3%

^C, ^N» and ^®F for research chemistry 23%

^^Na for renal studies 2%

Cyclotron development and maintenance (targetry, hot lab labeling systems, machine improvements) .32%

The first three groups listed above (total of 43%) represent studies of 30 patients, 80 monkeys, and 2 dogs. These studies average approximately two per week. 11

isotopes produced during the year ending November 30, 1972. Recent changes not reflected in this table include an increased emphasis on production of for in vivo studies and preliminary testing of targetry for the production of by the ^S(d,a) reaction. The heavy utilization of compounds labeled with and -^0 has occurred for several years and is expected to continue indefinitely, thus dic­ tating a high priority for the development of efficient target and labeling systems for these isotopes.

Section 1.3 In Vivo Studies of Cerebral Circulation and Metabolism

An understanding of basic life-sustaining physiological processes is a major gcal of the investigations conducted in the Division of

Radiation Sciences. Compounds labeled with the isotopes discussed in the previous section are utilized for the in vivo study of major metabolic pathways which sustain the function of vital organs. It is hoped that a natural outgrowth of this research will be the development of new and non-invasive tests for the detection of disease that can be employed routinely in the practice of nuclear medicine. In vivo stud­ ies involving experimental animals and patients have concentrated on the brain as uhe major organ of interest since its geometry offers easy accessibility for external recording of radioactivity by multiple scintillation probes collimated toward specific regions. An equally important motivation is the strong dependence of the welfare of this organ on an adequate continual external supply of oxygen and glucose. 12

A description of labeled compounds utilized for studies of meta­ bolism and hemodynamics Is given In Table 1-5, along with an indica­ tion of the modes of production. The abbreviated information presented here is only to supplement a cursory description of in vivo studies.

Details of production and labeling techniques will be treated exhaus­

tively in later chapters.

In vivo measurement of regional cerebral blood flow and regional

cerebral oxygen metabolism in animals and patients by means of blood

labeled with -^0 has been reported in several publications by Washing­

ton University authors (11,12,13,14,15). This work represents a coop­

erative interdisciplinary effort involving physicists, chemists, math­

ematicians, engineers, physiologists, and physicians. Additional

papers being prepared for publication will present a method for deter­ mining regional cerebral blood volume, including results of animal and

patient studies. Table 1-6 gives a brief summary of in vivo measure­ ments with -^0 and some currently pursued applications. Patient studies

are typically performed on selected hospitalized neurological patients

who have undergone carotid catheterization for diagnostic arteriography.

A description of these studies is presented in Appendix A.

^C-glucose, synthesized by the technique indicated in Table 1-5,

is utilized in a unique method developed for the in vivo measurement

of regional cerebral metabolic rate for glucose (rCMRG) in man. The

method has been validated in the rhesus monkey, and a current series of

patient studieo is in progress to evaluate its potential as a research Table 1-5

LABELED COMPOUNDS FOR IN VIVO STUDIES

Vascular Tracer Red blood cells labeled with C^o-hemoglobin (recirculate N£ + 3% CO target gas containing cl5o through venous blood)

Freely Diffusible Flow-Limited Tracer Blood labeled with H2 ^ 0 (recirculate N 2 + 3% O2 target gas containing C-^02 through blood)

Metabolic Substrates Red blood cells labeled with l^o-hemoglobin (recirculate N2 + 3% O2 target gas containing ol^o through venous blood)

Glucose labeled with ^ C (recirculate He + 1% O2 sweep gas containing H C O 2 past illuminated plant leaves, extract and purify phctosynthesized ^e-glucose)

Ammonia labeled with (recirculate bombarded CH, target gas through acid solution, or produce from Al ^ N in AI4C3 powder target by dissolving in HC1 and distilling from a NaOH solution) Table 1-6

IN VIVO STUDIES WITH 150

Regional cerebral blood flow determined from washout of (Internal carotid Injection)

Regional cerebral oxygen metabolism determined from measurements with l^O-Hgb (internal carotid injection)

Regional cerebral blood volume determined from combining measure­ ments of: (1) Mean transit time of Ho ^0 (internal carotid injection) (2) Mean transit time of C^O-Hgb (serial paired injections into internal carotid and venous circulation) (3) Cerebral hematocrit determined from (a) mean transit time of RBC tracer (same as (2) above) (b) mean transit time of plasma tracer (serial paired injections of -^I-RISA into internal carotid and venous circulation) (c) large vessel hematocrit

Flow index for regional cerebral blood flow (13 and 75 second data from washout of H2^ 0)

Investigation of permeability limitations of water as a physiolog­ ical tracer Correlation of fast compartment rCBF with pathology

Correlation of rCBF, rCBV, and regional oxygen metabolism in patients with dementia, head trauma, and pseudotumors 15

and diagnostic tool in humans. See Appendix A for a description of the method.

In vivo investigations of the cerebral metabolism of (syn­ thesized as indicated in Table 1-5) are in early stages of development and were not started until the ^ C / ^ O target and labeling systems described in this dissertation were nearing completion. Preliminary results with are encouraging and do not preclude the possibility of an eventual method for studying metabolism and hemodynamics by employing non-invasive intravenous injection techniques. These studies are mentioned here because of the important consideration of Integra-

lO ting NH^ production facilities with those for compounds labeled with

1 1 • • C and in order to perform future experiments which may require combinations of compounds Incorporating all three isotopes.

The interdisciplinary aspects of the effort supporting the in vivo studies described in this section were indicated in Table 1-1. Recent instrumentation advances have added the capability of a 26-probe sys­ tem for regional cerebral studies to the previous 6-probe and single­ probe systems. A dual processor computer system employing Interdata

Model 70 and Model 80 units is nearing completion, and will greatly add to the current capability offered by a "classic LINC" system for data collection and processing. Support offered by these recent additions will allow the Division of Radiation Sciences to greatly Increase the

quality and number of in vivo studies performed. 11 13 Further experience with methods for using J"LC-glucose and NHg

is expected to result in additional types of studies. An interesting 16

example is a proposed In vivo study of cerebral circulation and metabo­ lism in REM sleep (human volunteers) utilizing washout of inhaled ®^Kr for blood flow measurement and ^C-glucose for metabolism measurement

(16). In addition to the currently used labeled compounds listed in

Table 1-5 and discussed in this section, research chemistry efforts have resulted in the availability of the labeling capabilities de­ scribed in Table 1-7. 11 13 A recent paper on the importance of cyclotron produced C, N, and (17) underscores the usefulness of these isotopes by indicating the extensive program undertaken at the Washington University School of

Medicine for their use in the systematic in vivo study of a number of metabolic processes. It is evident that the disadvantages of these short-lived isotopes pointed out in Section 1.2 have certainly been overcome in many important applications, and that an effort to achieve efficient target and labeling systems in support of them is easily justified.

Section 1.4 Assessment of the Requirements for Labeled Compounds

In vivo studies described in Section 1.3 and Appendix A require rapid sequences of labeled compounds with adequate activity to permit good counting statistics. Production methods must also emphasize reliability and radiochemical purity. Significant loss of experimental data due to inefficient target and labeling facilities cannot be toler­ ated where a large interdisciplinary team of professionals is involved in conducting a complicated in vivo study. Table 1-7

LABELED COMPOUNDS DEVELOPED FOR PROPOSED IN VIVO STUDIES

Amino acids l%_glutamlne and glutamic acid for cerebral metabolism studies

Long chain fatty acids ilc-palmitic acid for regional cardiac metabolism studies

Ketone bodies llc-acetoacetate for cerebral fatty acid metabolism studies of ketogenlc diets for epileptic seizure control

Iodination of compounds with ^ ^ 1 18 FluorinatIon of compounds with F

30p-phosphate for cerebral metabolism studies and applications In nuclear medicine 18

Prior to the start of this dissertation project In April 1971, requirements for the production and delivery of compounds labeled with

and 15o had outgrown the capabilities of the original modes of production. It was evident that the higher activity and more rapid production for in vivo studies could be accomplished through increases in beam energy and extracted beam current, more efficient targetry, and more efficient labeling procedures. Rapid target changes were needed for serial measurements and concurrent experiments, and had to be accomplished without an increase in radiation dose to cyclotron operators. Table 1-8 summarizes criteria for the efficient produc­ tion of compounds labeled with and ^-*0 and indicates the basic approaches selected.

Interrelated cyclotron, targetry, and labeling considerations that affect the efficient delivery of labeled compounds were involved in developing new systems to satisfy the criteria of Table 1-8. Al­ though it was possible to delineate individual components or systems and assign a priority to their improvement, it was necessary to develop these three major areas concurrently in order to achieve an effective end result. This caused frequent reassessment of prototype equipment in terms of potential for increased performance, difficulty of achieving the proposed improvement, and cost of the final permanent equipment. It was also necessary to review the compatibility of the and 150 targetry designs with requirements for their use in conjunction 13 with production for other isotope applications, such as the NH^ 19

Table 1-8

CRITERIA FOR THE EFFICIENT PRODUCTION OF COMPOUNDS LABELED

WITH 11C AND 150

Cyclotron Performance and Reliability (1) High stable beam current of high energy under reliable operating conditions (a) vacuum isolation window permitting minimum energy attenuation of intense beam (b) improvements to cyclotron ion source, dee, septum, deflector, magnetic channel, and vacuum system

Target and Labeling Systems (1) High total and specific activity (a) improved target efficiency and geometry (b) minimum volume gas recirculation systems for labeling (c) improved labeling techniques (2) Reliability of production and purity of labeled compounds (s.) thorough routine leak checking procedures (b) independent labeling systems to eliminate cross contamination of shared system (3) Rapid target changes (a) tandem targets (b) beam switching (c) rabbit targets (d) independent labeling systems with automatic pumpdown when not in use (4) Low personnel radiation dose to operators and chemists (a) remote target changes (b) improved labeling system manipulations (c) shielding 20

studies which evolved during the development of and ^ 0 prototype equipment.

This section has indicated, in very general terms, the require­ ments for labeled compounds which motivated the research described in this dissertation. Translation of these requirements into detailed criteria for the development of individual components and systems will be covered in subsequent chapters, culminating in a description and evaluation of the permanent improved facilities for the production of compounds labeled with and l-*0 . CHAPTER 2

CYCLOTRON PERFORMANCE AND RELIABILITY

Section 2.1 Characteristies of the Cyclotron Beam

This section presents the techniques developed for measuring the energy, density, and spatial location of the deuteron beam. These parameters are basic information needed for the design of targetry, evaluation of changes made to increase cyclotron performance, and the monitoring of beam characteristics to assess shifts or malfunctions.

Evaluating deuteron beam energy by measuring the mean range in aluminum is described and discussed in Appendix B. Typical data ob­ tained using this technique are shown in Figure 2-1, which indicates a mean energy of 6.13 MeV. Table 2-1 compares this method with three others which are also described In detail in Appendix B. Although the crudeness of the total range in air and the two reaction measurements make them inferior to the measurement of beam energy by transmission in aluminum, the correlations are reasonable and serve to confirm the validity of the use of the reaction measurements in evaluating the targetry design presented in Chapter 3.

Figure 2-2 is a contour map of the beam constructed using a quan­ titative autoradiographic technique described in Appendix C. Iso­ activity lines correspond to beam densities ranging from the maximum

21 TRANSMITTED BEAM 0.0 0.5 1.0 0 A.RNE 51. gc* (6.63MeV) mg/cm* .5 1 MeV) -5 (6.13 RANGE MAX. mg/cm* >45.2 RANGE MEAN 10 jtk ALUMINUM ABSORBERS, ABSORBERS, ALUMINUM N EM VALUE BEAM ON 10 Mean Mean Range of Deuteron Beam in Aluminum 20 ETRN NRY MeV ENERGY, DEUTERON Figure 2-1 0 5 0 4 0 3

mg/cm 8 6.5

7.0 0 6 Table 2-1

COMPARISON OF SEVERAL METHODS FOR DETERMINING

DEUTERON BEAM ENERGY

Method Mean Deuteron Energy (MeV)

Maximum range In air 6.22

12C (d, n) threshold 6.23

•^N(d,n)^"*0 threshold 6.08

Transmission in aluminum 6.13 DISTANCE FROM BASE OF WINDOW, mm o m m 0 « » I I ■ « » I » I I «■ I » « « « I « » t - l - l » I i - t - i L - « « I ■ « « « I « I » « I « ■ « I I » I » » I I « » ■ « I ■ I I « I I » I « »

O O DT PIT FR S-CIIY LINES FOR ISO-ACTIVITY POINTS DATA o ETRN EM CIAIG RPIE LT WS TEUTD BY ATTENUATED WAS PLATE GRAPHITE ACTIVATING BEAM DEUTERON WN AA WNO WT 62m AR A, OLWD BY FOLLOWED GAP, AIR 6.25mm WITH WINDOW HAVAR ^ 4 5 . 2 TWIN 16m F OM AIR. ROOM OF 41.6mm OAIN F m OE I GRAPHITE IN Imm PLATE HOLES LOCATION OF 5

10

15 ITNE RM DE F IDW mm WINDOW, OF EDGE FROM DISTANCE

50%' % 0 •5 10 i%‘ - % 20 Beam DensiCy Contour Map

Figure 2-2 25

30

35

0 2 40 %

45

L 50 1 . 1 I . I . . I 1 .1 1 1 I

55

fO 25

(100%) down to 1% of the maximum. This contour map was used to eval­ uate fine qualitative methods of beam mapping (paper, blue cellophane,

Polaroid prints, aluminum plate autoradiographs, and graphite plate

autoradiographs). The results of this study (described in Appendix C)

support the conclusion that beam patterns made by thermal and chemical

effects induced in paper, blue cellophane, and Polaroid prints have

limited usefulness for all but the crudest measurements. Polaroid

autoradiographs made with graphite plates containing 1 mm marker holes were found to be the most satisfactory method of beam mapping for most

applications. Figure C-8 is an example. As a rule of thumb, a 10 yA

irradiation for 2 sec with a 10 mln decay period will result in a

plate reading of about 300 mr/hr y on contact, and a 1 min exposure on

Type 57 film developed for 15 sec will produce a pattern representing

an envelope where the beam density has dropped off to approximately

10% of its maximum value. If the exposure is low enough to correspond

to an envelope of around 50%, the 1 mm marker holes will not be clearly

resolved. If the exposure is high enough to correspond to an envelope

approaching or exceeding 1%, the 1 mm marker holes will start to fill

in and become blurred. This suggests a technique of making several

increasing exposures on the same sheet of film and selecting the auto­

radiograph with clearly resolved marker holes bracketed by the less

clearly resolved 1% and 50% conditions.

The evaluation of five qualitative methods of beam mapping using

one quantitative method was done at a position normal to a deuteron

beam attenuated by one twin window assembly and 4.13 cm of room air. 26

Until more experience is acquired with graphite plate autoradiographs at a variety of locations, it will be necessary to continue to use the quantitative contour map technique to evaluate the autoradiographs.

This is particularly desirable in an application such as the design of a new gas target that will be used frequently for a long period of time.

Time required to activate a plate, make twelve exposures on one sheet of film, and develop and dry the film, is about four hours. Densi­ tometer analysis, film calibration, and contour map construction takes about four hours. Several plates can be activated and processed at the same time and it should be possible to produce a sufficient number of contour maps in one day to design a gas target.

This section has discussed methods for characterizing the energy, density, and spatial location of the deuteron beam. These measurements were utilized in the development of a gas-cooled beam window assembly

(Section 2.2), improvement of cyclotron components (Section 2.3), and the design of the target and labeling systems presented in Chapter 3.

Section 2.2 Development of a Gas-Cooled Beam Window Assembly

The requirements presented in Section 1.4 establish the need for stable high external beam current of highest possible energy under reliable operating conditions. Potential for increasing the usable external beam current was great since the stable deflected beam inci­ dent on the beam valve in April 1971 at the start of this research was 30-40 yA, while the characteristics of the vacuum-isolation window limited the extracted beam to 15-18 yA. 27

Criteria for developing an improved beam window assembly are given

in Table 2-2. The approach taken was to develop a double foil geometry of minimum foil thickness with provision for high velocity gas cooling between the foils. Goals established were energy attenuation less than

the original design, beam current capability equal to the maximum de­ flected beam, and the ability to withstand two to three atmospheres of differential pressure. Havar foil of 2.54 y thickness (see Table D-l

for properties) was selected for initial tests because of its high

strength and because it was the only material that could be purchased

in this thickness without pinhole defects.

Energy attenuation for various window geometries is given in

Table 2-3 from the results of calculations presented in Appendix G.

A double window constructed with 2.54 y Havar and employing a 0.635 cm

air cooling space between the foils (last geometry in the table) meets

the energy attenuation criteria. Appendix D describes the development

of a prototype assembly of this configuration constructed by bonding

the foils between aluminum plates with a silicone rubber compound. The

search for and testing of suitable adhesive/sealant materials, and de­ veloping techniques for assembling and disassembling window hardware were major problems.

Figure 2-3 shows the window assembly design for the new tandem

target system described in Section 3.3. Details of this design are

discussed in the last paragraph of Appendix D.

Routine utilization of the new window assembly began in November

1972 with the complete installation of the tandem target system, 28

Table 2-2

CRITERIA FOR DEVELOPING A NEW BEAM WINDOW ASSEMBLY

Criteria Remarks

Minimum energy attenuation B(d,n)^C yield drops rapidly with energy in the 5-7 MeV region (see Figure 3.1-5). Beam path in July 1968 window design is 27% longer than if beam were perpendicular to window.

Maximum beam current Gas cooling of foil in July 1968 window design limited to recircu­ lation of target or sweep gas where used (flow rates around one liter per minute). Desirable to have high flow rate supply independent of tar­ get or sweep gas.

Mechanical strength Elastomer "0M-ring mounting of July 1968 window design causes wrinkling and pulling of foil, deterioration of "0"-rings in radiation field, and limited ability to withstand differential pressure pulses or gas target pressures above atmospheric.

Geometry July 1968 window design not normal to beam. Window opening should be large enough to accommodate shifts in beam location but no larger than necessary in order to maximize ability to with­ stand differential pressure. Table 2-3

ATTENUATION OF 6 MeV DEUTERONS BY VARIOUS MATERIALS4

Energy loss in keV at 2 beam angles of: Attenuating Material mg/cm 52° 90°

25.4 y A1 6.86 771 602

6.35 y Havar 5.27 466 364

2.54 y Havar 2.11 182 144

0.635 cm Air (1 atm) 0.75 95 75

0.635 cm He (1 atm) 0.07 17 13

5.08 y Havar and 4.97 478 376 0.635 cm Air a See Appendix A for method of calculation. Figure 2-3 Details of Current Beam Window Assembly 31

although it was tested and used for beam mapping determinations a month earlier. Ability of this design to withstand differential pressure was tested by pressurizing the cooling space until failure occurred at approximately 4 atm, which satisfies the criterion of Table 2-2. Calcu­ lated energy loss in this window is reduced 102 keV compared to the prototype since the tandem target system places the window normal to the beam (see Table 2-3), again satisfying the criterion of Table 2-2.

Individual assemblies utilized starting November 1972 have operated from one to two months each before developing pinhole leaks in the foil surface on the target side. Two assemblies failed violently at

30 yA currents because the air cooling supply was not activated (opera­ tor error). The only other failures encountered were a series of ten assemblies inadvertently put together with RTV-602 silicone rubber that had been stored longer than the recommended shelf life of six months.

These failed immediately when one side was pumped down due to the much reduced strength of the rubber. Uncured rubber and catalyst are now discarded after three months of use.

Cooling gas for all prototype and tandem target twin-window

assemblies has been room air supplied by an air compressor connected

to the inlet passage. The outlet is connected to the same discharge vent as the chemistry hood in the cyclotron hot lab. Flow was measured by attaching the outlet to a collapsed 6 liter anesthesia bag and mea­

suring the inflation time. This was done for a single assembly and for

two assemblies in parallel, resulting in a flow of 1.2 liters/sec in 32

both cases. Pressure between the foils during operation was measured by attaching a pressure gauge to the tapped hole in the center plate which is normally blocked by a handling rod (see Figure 2-3). It was

■ found to be 1.6 atm for a single assembly and 1.3 atm for the normal case of two assemblies in parallel. Calculated average air velocity in the space between the foils is 10 m/sec for two assemblies in parallel. This level of air cooling has proved adequate for beam currents up to 50 yA. Potential for use of this type of window assem­ bly at higher beam currents will be discussed in Section 2.4. The way in which the November 1972 window assembly is incorporated in the tan­ dem target system will be described in Section 3.3.

Section 2.3 Modifications to Cyclotron Components

Higher external beam currents permitted by the air-cooled twin- window provided the immediate potential for a factor of three increase in the activity of -^0 and ^ C produced. Since the cyclotron had been routinely operated from September 1968 to September 1971 at steady external beam currents no higher than 15-18 yA, a cautious approach was taken to increasing the beam current used for steady running.

The higher internal beams associated with operation at 30 yA external beam3 caused visible glowing of the dee edge, ion source pullers, and ion source chimney. Careful and more frequent position­ ing of the ion source and dee were necessary to minimize the additional heating effects. In March 1972, after several long production runs at

35 yA, the copper trailing section of the two-piece tungsten/copper 33

septum strip sustained local melting. The edge of the copper section which butted against the tungsten front section bowed out into the path of the beam sufficiently to cause local melting and intercept the beam. The entire septum strip was then replaced by one fabricated from a single continuous piece of tungsten.

Glowing of the Elkonite alloy knife-edge pullers continued to be a problem, and they were replaced in June 1972 by a tungsten finger- type puller. The heating of the dee edge was also eliminated at this time by the installation of a graphite collimator block with a 0.952 cm slot which shielded the dee from the stray beam. Another area suffering melting damage due to stray beam at high beam currents was the ion source support tube. This problem was similarly solved by installing a graphite shielding block of the same height as the top of the ion source tube.

Higher peak accelerating voltages necessary to insert larger ion bunches into orbit (and thus achieve higher internal beam currents) caused the expected problems with high-voitage sparking between the dee and the grounded vacuum chamber components. Such spark discharges decreased somewhat with conditioning from operating at higher beam currents, but with the limitation of the beam window removed they became the factor controlling the maximum activity that could be produced.

Several approaches were taken to raise the ultimate limit on dee voltage due to sparking.

Cyclotron vacuum was improved by the installation of new water- cooled baffles on the diffusion pumps. The change was not noticed on 34

the vacuum gauge, but did result In more stable operation at high beam currents and considerably less frequent changing of the vacuum chamber port plates due to oil vapor deposits.

It was necessary to go down to air several times to remove

"needles" built up from sparking in the areas of the septum and de­ flector structures. This was remedied by rounding some sharp edges and by installing copper sheets over some exposed areas of the iron chamber lids.

Ion source performance was improved by modifying the connection between the chimney and support tube in order to facilitate cooling.

Also, a continuing effort is being made to alter the geometry and size of the emission hole in order to increase efficiency and reduce the amount of deuterium gas contaminating the vacuum at high beam currents.

The foregoing considerations are efforts to increase the external beam current by improving the internal beam. Extraction efficiency enhancement is the other approach undertaken. Careful and frequent tuning of the septum and deflector was necessary in order to maintain high stable beam currents, but no way was available to evaluate the position of the magnetic channel to assure that most of the deflected beam was being delivered through the window to the target area. With the machine down to air, a Polaroid sheet film autoradiograph was made of the channel face at which the beam enters. This showed a radioactive area on the median plane extending approximately 0.6 cm from the outside edge of the channel opening, indicating that possibly one third of the deflected beam was being intercepted by the structure of the magnetic 35

channel. Channel width was increased in May 1973 from 2.54 cm to 3.81

cm by extending the outer edge of the aperture 1.27 cm further to the

outside. This change was evaluated with the aid of the hardware used

to determine beam energy by transmission in aluminum (see Appendix B).

Since the collimator hole in this hardware is near the inside edge of

the beam, the ratio of beam current on the beam valve to current trans­ mitted through the hole should increase if the beam striking the beam valve is increased by removing the channel blockage on the outside edge.

From Figure 2-1, on 1/9/73 before the channel was widened, a 10 yA

current on the closed beam valve corresponded to a transmitted current

of 1.45 yA with no attenuators in the collimator hole. On 6/21/73,

after the channel was widened, a 15 yA current on the beam valve corre­

sponded to a current of 1.55 yA under the same conditions. The ratio

of beam valve to transmitted current increased from 6.9 to 9.7, indi­

cating a 40% increase in beam current due to the modification of the

magnetic channel.

In May 1973 the dee stem and dee structure were removed from the

cyclotron for structural and water cooling channel defects. These had

resulted in low beam current and erratic running due to excessive move­

ment of the dee during operation. When the repaired components were

replaced and the machine tuned, the leading edge of the septum was

further out. Beam energy measured on 6/21/73 by transmission in alumi­

num was 6.4 MeV, about 300 keV higher than measurements made in January

1973 before the dee and stem defects became significant problems. Since

the oscillator frequency and septum position are not routinely measured, 36

it is not possible to determine how much of the observed beam energy increase might be due to a change in frequency caused by vertical repo­ sitioning of the dee after the repairs. An attempt will be made in the future to periodically measure and correlate frequency, exit radius, and beam energy in order to assess slow drifts in machine geometry and condition.

A significant difference exists in the beam energy measurements presented in Section 2.1 and the beam energy calculated assuming the mechanical center of the pole tips is the center of the last orbit.

On 1/9/73, the date that the mean deuteron energy of 6.13 MeV was measured by transmission in aluminum, the frequency was measured as

11.76 MHz. The exit radius was also established at this time as

34.76 cm by measuring the relationship of the leading edge of the septum to the perimeter of the pole shoe. Using the relationship

T = (1.063 x 10“^)f^R^, where T is in MeV, f is in MHz, and R is in cm (18), a calculated energy of 7.0 MeV results. Assuming the 6.13

MeV energy and the 11.76 MHz frequency are correct, the calculated exit radius is 32.54 cm, or 2.22 cm less than the measured radius. An inves­ tigation leading to an understanding of the reason for this discrepancy might also suggest a possibility of increasing deuteron energy by re­ positioning the dee and ion source.

This section has presented problems encountered since September

1971 in the course of increasing routine production run capability from

the region of 15-18 yA to the vicinity of 35-40 yA. Considerable poten­

tial remains to achieve higher extracted beam currents and higher 37

deuteron energy. Beam currents as high as 80 yA have been measured through a twin-window assembly during runs of a few minutes duration.

It is hoped that additional improvements to cyclotron performance and reliability as proposed in Section 2.4 will result in routine produc­ tion runs with beam currents up to 80 yA and deuteron energies up to

7 MeV.

Section 2.4 Future Work

Section 2.1 of this chapter establishes convenient and reliable methods for accurately determining the energy and spatial location of the deuteron beam. Energy measurements have been made months apart, and have indicated significant variations in energy. It would be desirable to measure the beam energy and oscillator frequency on a regular basis, perhaps once a week, until the extent and nature of the variations are determined. The technique described for measuring deuteron energy takes about four hours if sufficient points are acquired to assess the effects of energy spread and straggling, but an accurate determination of the half-maximum of transmitted current to determine a mean range can be done in an hour. For convenience, the determination could be made before or after a regular run which utilizes the required tandem target hardware. Continued use and comparison of the qualitative and quantitative autoradiographic beam mapping with activated graphite plates should result in increasing reliance on the less time-consuming

Polaroid autoradiographs. Experience with the contour map procedure may reduce the time required, but it will still be very long compared 38

with the Polaroid method.

The air-cooled beam window assembly described in Section 2.2 has proven very adequate for the beam currents of up to 50 yA currently available. Crude heat transfer estimates based on forced convection only have indicated that the service limits of 400°C for the Havar foil and 260°C for the silicone rubber bonding may be approached if routine runs above 50 yA become possible. Window height and width are both optimum for the deflected beam available since the magnetic channel was modified, so future gains in window performance related to window geometry will have to come from narrowing the space between foils to increase cooling gas velocity. It would also be possible to achieve higher velocity by using a larger compressor to boost the inlet pressure to the window assembly. The pressure between the foils is presently

1.3 atm for two windows in parallel and 1.6 atm for one window, so a substantial increase in flow could probably be achieved before stress on the foils becomes the limiting factor. Another approach to increas­ ing the surface coefficient of forced convection would be the substitu­ tion of helium for air as the cooling gas. This would require a re­ circulating system for economy reasons. Using a relationship developed by Brown and Marco (19) for turbulent flow parallel to smooth plane surfaces, the surface coefficient is predicted to increase from 9.6 x

10~3 Watts/(cm2)(°C) for air (17 Btu/(hr)(ft2)(°F)) to 13.6 x 10“3

Watts/(cm2)(°C) for helium (24 Btu/(hr)(ft2)(°F)) if the velocity re­ mains at its present value of 10 m/sec (two window assemblies in 39

parallel). The above considerations are presented as possible means of extending the capabilities of the present window assembly only if it does not perform acceptably at beam currents above 50 pA. Since the present design may have considerable reserve, it should be periodi­ cally tested under the most exhaustive conditions possible as machine improvements make higher beam currents available. Successful short­ term testing with an intermittent beam of 85 yA for approximately five minutes has been accomplished.

The problems encountered and solved in the process of increasing the beam current capability for routine production rune were discussed in Section 2.3. This work suggests that a continuing effort to re­ design and upgrade cyclotron components will lead to significant further improvements in performance and reliability. An area which should be given high priority is the process of tuning, which has be­ come more critical and frequent as the beam currents for routine produc­ tion have gone up. Adjustments to position the deflector, septum, and dee are now made manually at the surface of the machine, and the mag­ netic channel is not adjustable. These components could be redesigned to provide motorized remote adjustment capabilities while the machine is being operated. At the same time provisions could be made to insulate the septum and channel so that intercepted beam can be moni­ tored while tuning. This will require appropriate circuitry in the case of the septum to protect the microammeter from occasional deflector- to-septum arcing. A means of analyzing the deflector leakage current for intercepted beam current superimposed on it would also be desirable. 40

The provisions to motorize, insulate, and instrument the components just described would be a time-consuming design project and involve considerable fabrication expense. The decision to undertake this effort should be preceded by a concentrated attempt to tune for abso­ lute maximum beam current using the present manual adjustments. Since tuning is such a frustrating and time-consuming task, and the demand for production runs in support of research studies is usually pressing, tuning efforts in the past have stopped when an acceptable beam current for normal requirements (30-35 yA) is reached. A full week or more spent on nothing but testing above 40 yA should indicate what is possi­ ble with the present mode of tuning and suggest stability and relia­ bility behavior. If a stable maximum around 80 yA could be achieved, there would be little reason for the remote controls and instrumenta­ tion suggested in the preceding paragraph.

As discussed in Section 2.3, it would be useful to measure oscil­ lator frequency and beam energy frequently to assess the extent of variations and attempt to determine their cause, which may suggest improvements resulting in higher beam energy. Acquisition of a meter for daily frequency measurement would be a good investment, and the hour or two required for a weekly beam energy measurement is not un­ reasonable. An increased priority (and allocation of time) for tuning considerations, with routine frequency and energy measurements to evaluate the results, are certain to lead to immediate improvements in performance and reliability, and generate the basis for further modi­ fications. In order to maintain reliability at present beam current 41

levels and the higher ones anticipated in the future, it will also be necessary to place increased emphasis on Improved, formalized, and

scheduled preventative maintenance procedures. Reliable operation at

80 pA of 7 MeV deuterons is felt to be a feasible goal and would in­

crease the currently available and ^"*0 activity by approximately

a factor of three. It is felt that achieving these conditions would

approach the performance limits of the Washington University Medical

Cyclotron and allow the production of optimum quantities of and

■^0 compounds for the type of research studies to which they are

applied. CHAPTER 3

TARGET AND LABELING SYSTEMS

11 Section 3.1 Development of a New Target for the Production of C

Design of an optimum target system to satisfy the criteria given

In Section 1.4 was severely limited by the state-of-the-art targetry available for the production of by the ^®B(d,n)^^C reaction. This section presents the development of a new type of target whose yield, size, and life characteristics permitted the design of the tandem target system to be described in Section 3.3.

■^B(d,n)^C targetry in use at this and other medical cyclotrons at the start of this research project (20,21) consisted of a large 9 10 area (typically 20 cm ) of boron oxide, isotopically enriched in B, applied in a layer about five millimeters thick on a sloping surface

(typically 30°) which was sometimes serrated to retard run-off. This type of target was first described by Kamen in 1947 (22). The beam melts the B2^3 * formed by the ^B(d,n)^^C reaction reacts with 11 11 oxygen in the target material, and the resulting CO and C02 gases are released. In most labeling applications at this cyclotron, a helium sweep gas containing 3% oxygen is passed over the target and the

1:LC0 undergoes radiolytic oxidation to **C02, which is the desired pre­ cursor compound. The target geometry just described as used at this

42 43

cyclotron was bulky, released only about 40% of the activity pro­ duced, and continuously migrated away from the beam strike under the influence of gravity. Remelting or replacement of the B2O3 was necessary after approximately two hours of bombardment to restore the target material to its initial geometry.

Research on high-temperature lithium-bonded reactor fuel pins (23) suggested that a refractory metal screen might offer the possibility of supporting the molten B2O3 in a thin layer by surface tension in the screen pores. Some properties of boron oxide relevant to the inves­ tigation of this idea are given in Table 3-1, and will be referred to in the following discussion. Seven different types of screen in a prototype target geometry were evaluated for adequacy by attempting to fabricate a workable target. Natural boric acid powder (HgBOg) was sprinkled onto a horizontal screen surface and placed in a 700°C oven until boric oxide glass (B2O3) had formed and coated the screen, which took about ten seconds. This process had to be repeated four to six times before enough B2O3 glass had been accumulated to fill in the screen pores and produce a smooth surface. The water lost in the con­ version of H3BO3 to caused violent sputtering and fuming, and the large reduction of volume required the several repeated applications.

A description of the screens tested and the results of the tests are given in Table 3-2. The geometry used is footnoted in the table and illustrated in Figure 3-1, which shows a 100-mesh molybdenum screen with 0.069 mm wire size coated with B2O3 . Molybdenum was the only 44

Table 3-1

PROPERTIES OF BORON OXIDE®

Chemical formula: B2O3 , 69.64 molecular weight, 31.07% B, 68.93% 0 .

Names: Boron oxide, boric anhydride, boron trloxide, boric oxide, boron sesquioxide, Improperly called anhydrous boric acid or fused boric acid.

General description: Colorless, semitransparent, brittle, vitreous, hygroscopic.

Physical properties: 450°C melting point, 1860°C boiling point, 1.85 g/cm* density.

Related compounds: Boric acid (boron hydroxide) forms are orthoboric acid (H3BO3 , white powder) which loses H2O at 100- 105°C to form metaboric acid (HBO2) which loses H2O at 140-160°C to form pyroboric acid (H2B4O7) which loses H2O above its melting point of 160°C to form boron oxide (B2O3).

Decomposition: B2O3 glass left in room air hydrates to the H3BO3 boric acid form, the white powder becomes visible on the glass surface in a few hours, and a one millimeter thickness of glass is completely converted in about a week.

Natural abundance: 19.78% ^B, 80.22%

Enriched material: (20th Century Electronics, Surrey, England, 3-4-69) 1^ 0 3 , B is 90.4 atom percent 10jj, approximately $10 per gram 10b, 1 - 6 .8 , X - 13.6384. aHandbook of Physics and Chemistry, 50th Edition (1969-1970) and The Merck Index, 8th Edition (1968). ( ( (

Table 3-2

EVALUATION OF SCREEN MATERIALS AND SIZES

Pore Melting Wire Open Structural Oxidation Overall diameter size B2°3 h Material point Mesh (mm ) (mm ) area stiffness3 retention0 resistance0 evaluation W 3410°C 16 0.203 1.384 76% Good Poor Poor Unacceptable

W 3410°C 25 0.127 0.889 76% Good Poor Poor Unacceptable

W 3410°C 50 0.076 0.432 72% Adequate Poor Poor Unacceptable

Mo 2625°C 70 0.102 0.262 52% Adequate Fair Good Fair

Ta 2996°C 100 0.076 0.178 49% Adequate Good Fair Fair

SST 1500°C 100 0.025 0.229 81% Unacceptable Untried Untried Unacceptable

Mo 2625°C 100 0.069 0.185 53% Adequate Good Good Good**

Prototype screen target geometry of 1.27 by 4.45 cm target surface and edge extensions bent at 90°. ^Before irradiation. °Two minutes in air at 700°C. ^Newark Wire Cloth Co., Newark, N.J. Filmed as received without page(s) 46

UNIVERSITY MICROFILMS. Figure 3-1

Prototype Screen Target jj> 48

original target geometry.

Prototype screen targets were used for production and test runs 11 requiring C activity until November 1972. Prior to this time the only significant change made was in the method for coating the B2O3 on the screen. Rather than using the fine HgBOg powder, large pieces of enriched B2O3 glass were crushed to coarse grains using a steel piston hammered in a cylindrical hole in a steel block. The crushed glass requires only one oven application to coat the screen, and less is lost in handling than when using several applications of the finer powder.

In September 1972 the screen target geometry was adapted to the requirements of the tandem target system to be fully described in

Section 3.3. Figure 3-2 is a photograph of the section of this system used for -^C production. The rectangular aluminum target body on the left has an opening in the face (surrounded by an "©"-ring groove) through which the beam enters. The beam opening goes completely through the block. A round aluminum target mounting rod with a rec­ tangular end is shown in line with the hole in the block into which it fits. Guide rods projecting from the block are used to align and restrain the target rod. The left end of the rod has a surface in­ clined at 45° to the beam containing a 2.222 by 5.715 cm recess 0.238 cm deep to retain the screen target shown directly in front of it. A bare screen is shown just in front of the coated screen, and a teflon form used to bend tl.e edges of the screen under is in front of the bare *eC 50

screen. The screen target is held in place by the ,,U"-shaped 1 tnm stainless steel wire clip which passes through and into holes in the target rod. By inserting and withdrawing the target rod, the screen target can be alternately placed in the path of the beam or completely removed, allowing the beam to pass through the target block. Sweep gas enters and leaves the target block via the two small holes in the top closest to the front edge. The third small hole is a vacuum pumpout.

Sweep gas flows directly down on one end of the screen target, across its surface, and exits directly above the other end. Large holes on the tops of the target block and target rod are water-cooling inlets for passages surrounding the beam opening in the block and running under the screen target area in the rod. Water exits are on the bottom of the block and rod opposite the inlets. Other parts shown in Figure

3-2 are "0"-rings for sealing the target block against a beam window on both sides, '‘0"-rings for sealing the sliding target rod and the dummy plug shown above it, nuts near the end of the guide rods for limiting the outer movement of the target rod, and nylon insulating sleeves for the two large mounting stud holes in the face of the target block, above and below the beam opening.

Figure 3-3 is a photograph of a bare screen, a coated screen before bombardment, and a screen target used for about 25 hours at 30-35 yA.

Approximately one gram of crushed B£0^ glass is sprinkled onto the screen and melted by placing it in a 700°C oven for one minute. Target thick­ ness averages 3.9 mm before use, and 0.6 mm in the area of the beam (

Figure 3-3 Current n C Screen Target ^ i- 1 52

strike after several hours of use. Excess B2O3 tends to migrate to the lower surface of the screen at the bottom edge of the beam strike, but does not attain sufficient thickness to touch the surface of the aluminum rod below it, which would provide an unwanted conduction path and reduce the temperature of the target material. This was a problem with the prototype screen target whose short 1.27 cm height and verti­ cal mounting caused excess B2O3 to collect at the base of the screen on the floor of the target block. Activity trapped in the current de­ sign was measured at the end of three 30 yA runs of 30 minute duration in the same manner as previously described for the original and proto­ type systems, The values of 11-13% for activity remaining in the target indicate the importance of maintaining the mounting surfaces of the screen target in poor thermal contact with the target rod. Figure

3-4 is a photograph of the end of the target rod after about ten hours of operation at 35 yA.

Mean ranges for 6 MeV and 7 MeV deuterons in B2O3 are approxi­ mately 35 mg/cm^ and 45 mg/cm^, corresponding to 0.178 mm and 0.254 mm.

Ranges for attenuation from 6 MeV and from 7 MeV down to the threshold 2 2 energy of about 2 MeV are approximately 30 mg/cm and 40 mg/cm , corres­ ponding to 0.152 mm and 0.203 mm. Since the target thickness in the area of the beam strike never becomes less than about 0.5 mm, all the beam energy (current 6 MeV or maximum anticipated 7 MeV) will be de­ posited in the target. Also, all of the -^C activity will be generated in the B2O3 layer above the screen, and the wire does not reduce the C« M

Figure 3-4 Irradiated Target Rod 54

area of target material available for the nuclear reaction.

The current target is kept under vacuum when installed on the cyclotron and not in use. This prevents decomposition of the hygro­ scopic ^2^2 to H3B03 Pow^er * which becomes significant during over­ night exposure to moist room air. When not installed, the target rod is removed and kept in a pumped down bell jar. A spare screen target is kept in a jar containing Drierite (CaSO^, anhydrous calcium sulfate).

A search of the literature pertinent to excitation function and yield data for the ^B(d,n)*^C reaction was made in order to calculate correlations with the measured thick target yield for the screen tar­ get. Of the papers reviewed (24,25,26,27,28) only one presented the excitation function desired (24). Figure 3-5 shows this excitation function along with an average cross section introduced by Ricci and

Hahn (29) for considerations involving the activation of thick targets by charged-particle bombardments. Thick-target saturation yield, cal­ culated using the average cross section as described in Appendix H, is also shown in Figure 3-5. The experimental points shown at 5.71 MeV were obtained from four ^C-glucose production runs made during Decem­ ber 1972 and January 1973, during which time the cyclotron septum position was not changed from the condition of the 1/9/73 beam energy measurement presented in Section 2.1. Energy of the beam striking the target during these production runs was assumed to be 6.13 MeV (measured on 1/9/73) less the calculated attenuation of 376 keV (Table 2-3) for an air-cooled twin-window assembly, and less the calculated attenuation 55

of 44 keV for the average of 2.2 cm of helium sweep gas between the window assembly and the screen target. This establishes a deuteron energy of 5.71 for the four experimental yield points shown in Figure

3-5. These runs were all made at 30 yA for bombardment times varying from 30 to 40 minutes, and the yields adjusted to the saturation and beam current units used in the figure. The method for obtaining the total activity produced during a ^C-glucose production run will be described in Section 3.3.

Although no experimental yield data was found in the literature for ^®B(d,n)^C, the search did lend some insight as to where to watch for future publications. Papers originating from organizations with medical cyclotrons frequently mention activity (20,21,30), but in terms of specific activity in the form of a labeled gas without refer­ ence to total system volume or activity trapped in the target. Quite often, activity is given only for a final labeled compound after syn­ thesis and decay losses have occurred. Russian and American publica­ tions on charged particle activation analysis (25,31) come closer to the kind of information desired, but usually treat natural boron or a compound other than boron oxide in energy ranges where the ^ B ( d ,2n) reaction is contributing to the activity (above 6 MeV). The suggestions of Ricci and Hahn (29) outlined in Appendix H will be useful in adjust­ ing yields to compensate for differences in target materials, if suitable experimental yield data becomes available in the future. A paper by

Svoboda (32) on experimental and theoretical considerations regarding 200

150 CALCULATED SATURATION YIELD (THICK TARGET)

JO

,0B(d,n)"C EXCITATION FUNCTION (BRILL AND SUMIN) / ZlOO RICCI AND HAHN

CO

CO 0 50

• ■ EXPERIMENTAL YIELDS AT 5.71 M«V ON TARSET (S.IS M«V OEFLECTEOl

DEUTERON ENERGY, MeV Figure 3-5 Excitation Function and Yield for ^B(d,n)^C 57

the terms used in the production of radioisotopes by charged-particle bombardment was also helpful in interpreting the often confusing and

ill-defined units for yields reported in the literature.

The large change in yield with deuteron energy in the 5-7

MeV range (see Figure 3-5) deserves special emphasis. Consideration of

this effect was the primary motivation for the priority given to the

development of the air-cooled beam window (described in Section 2.2) with the stringent requirements for minimum energy attenuation. It is also a strong motivation for the effort to increase beam energy in the

future as proposed in Section 2.4. An increase in deflected beam energy

from 6.00 to 7.00 MeV (5.58 to 6.58 MeV on screen target) corresponds

to an increase of 54% in yield.

This section has described the work resulting in the current

targetry shown in Figure 3-2. ^®B-enriched B2O3 required for a target

has been reduced from 20 grams to 1 gram, corresponding to costs of

about $50.00 and $2.50. Volume of the target chamber has been reduced 3 3 from 320 cm to 21 cm , and the distance traveled by the beam in the

sweep gas before striking the target has been reduced from 5.7 cm to

2.2 cm. Target life has been increased from about two hours to over

twenty hours at beam currents of 30-40 yA. The usable fraction of the

total activity generated (labeled gas released by the target) has been

increased from about 40% to about 88%. Small size and the ability of

the screen target to maintain its geometry and performance at high

beam currents for long bombardments have allowed the design of the

remotely operated H c / ^ 0 tandem targetry described in Section 3.3. 58

The section of this assembly has performed successfully with 50 yA of 6.1 MeV deuterons, corresponding to 305 watts of power generated in the target. No problems are anticipated in approaching 80 yA of 7.0

MeV deuterons (560 watts) which is the future potential for cyclotron performance proposed in Section 2.4. It is felt that this design adequately meets the criteria for improved target efficiency and geom­ etry specified in Section 1.4. The way in which the targetry is utilized for production of labeled glucose will be presented in Section

3.2, and its relationship to other parts of the tandem target and associated labeling systems will be described in Section 3.3.

Section 3.2 Investigation of a Photosynthetic Process for Labeling

Glucose with

Biosynthetic preparation of 13-C-glucose utilizing uptake by

illuminated Swiss chard leaves was developed by Lifton and Welch (33) in

1970 and continues to be used frequently at this cyclotron facility. 11 The original method of exposing the leaves to the C0£ was adequate for

^C-glucose production until late in 197.1, when the available

activity was increased by approximately a factor of four due to routine

use of the prototype twin windows (Section 2.2) and prototype screen

targets (Section 3.1). Radiationdose to chemists, and a demand for even

higher activity for research studies, motivated a study of the parameters

affecting leaf uptake of and the procedure used to accomplish it.

This section describes the investigation and resulting modifications

tested during the period February through June 1972, culminating in an 59

improved system for accomplishing the photosynthetic portion of the labeling process.

In the original method (33), helium sweep gas containing 3% O2 re­ circulated past the B2O3 target for a 40-minute bombardment at 18 pA.

1 1 The He-02 mixture containing carrier-free COj was then pumped out of the target system through a liquid nitrogen trap (77°K) which collected

11 ^ the CO^. A 105 cnr glass chamber containing 4-7 g of light-starved

(48 hr) Swiss chard leaves cut into 2 cm squares was evacuated, valved off, and connected to the trap containing the After warming the 11 trap in hot water, the CO2 was pulled into the glass chamber by flush­ ing helium through the trap until atmospheric pressure was reached.

0.2 ml of water was added to the glass chamber, and it was illuminated by a 650-watt tungsten iodide lamp at 61 cm and a W31 warm-white fluores­ cent lamp at 8 cm (Mic-O-Lite II, 3200 lumens over 15 cm circle, 30 pA at 650V power supply, from Aristo Grid Lamp Products, 65 Harbor Road,

Port Washington, L.I., N.Y.). The chamber was illuminated for 20 minutes and was agitated every two to three minutes to facilitate uniform exposure of all the leaf pieces.

A recirculating system in which the leaves could be illuminated while the leaf chamber was on line with the target was proposed as a possible improvement. ^ C 02 could be supplied to the leaves as it was produced, avoiding decay losses due to the separate bombardment period and the trapping process. Total production time would be re­ duced and radiation dose to chemists during the trapping and flushing 60

processes would be eliminated. For these reasons a simultaneous bombard/circulate/illuminate process was Investigated and a study made of the parameters affecting leaf uptake of "^COg. The alcohol extrac­

tion and acid hydrolysis steps immediately following the illumination were also evaluated, and the activity losses due to decay and process­

ing were determined from the target to the final -^C-glucose solution ready for physiological injection. These studies resulted in changes

in the type and geometry of the light sources used to stimulate photo­

synthesis, in the geometry and quantity of leaves used, and in the

routine use of the on-line process.

Figure 3-6 illustrates the basic steps in the total ^C-glucose

production process using the recirculating method. A running total

of the time xequired is indicated at each step. Details of those steps

evaluated and/or changed from the original method are presented in

Appendix E. Parameters assessed are listed in Table 3-3.

The original preparation of "^C-glucose used approximately 5 grams

of leaves and contained about 5 mg of non-radioactive glucose in the

final solution ready for injection. Using the improved illumination

conditions for the on-line process described in Appendix E, it was found

that a 5 cm diameter piece of leaf weighing about 0.5 gram was adequate

to achieve uptakes as high as 85% of the ^ C 02 present at the end of a

20-minute illumination period. The 0.5 gram leaf is now used routinely,

resulting in about 0.5 mg of non-radioactive glucose in the final injec-

tate. This is desirable since a minimum perturbation of the blood PRODUCTION OP 11C-GLUCOSE

["P* time] [20 minutes]

Glucose starved Photosynthesis , Fructose 11CO, + Swiss Chard Sucrose leaves Starch

alcohol extraction

Glucose Fructose Sucrose

acid hydrolysis liquid Glucose chromatography Glucose neutralize Glucosei + + « * - + 10% Fructose on Magnesium Fructose + concentrate Fructose Silicate

[75 minutes] [40 minutes] [30 minutes]

Figure 3-6 PRODUCTION OF n C-GLUC0SE 62

glucose level in an animal or patient study is wanted. In the commonly studied rhesus monkey with a normal blood glucose concentration of about

1 mg/ml and a blood volume of about 500 ml, the amount of glucose in­ jected would be reduced from 1 .0% to 0 .1% of the blood glucose present.

Plant proteins which are removed in the liquid chromatographic purifi­ cation on the magnesium silicate column (Figure 3-6) are also decreased by the reduction in leaf weight. This lessens column decomposition and reduces the frequency of changing the column packing which is now done about every ten runs instead of every two or three.

Current production data on ^C-glucose for a 30 yA bombardment is given in Tables 3-4 and 3-5. This type of data was used to evaluate and develop steps in the on-line process and hot lab chemistry. Carbon-11 activity at most of the key points indicated in the tables was recorded for 23 of the routine production runs made during the period June 1972 through October 1972 using the prototype on-line system. In November

1972 the tandem target system became available and the data was recorded for eight of the ^C-glucose runs made through January 1973. Four of these were the basis for the experimental points shown on the -^C yield curve of Figure 3-5 discussed earlier in Section 3.1 of this chapter.

The current capability of 3.5 mCi of ^C-glucose ready for physiological

injection as indicated in Table 3-4 is more than adequate for frequently performed studies of non-regional cerebral metabolism in rhesus monkeys but is less than the activity desired for future studies planned to

investigate regional cerebral metabolism in patients. This will be

discussed in the following paragraph. 63

Table 3-3

n C-GLUCOSE PRODUCTION PARAMETERS

Illumination during photosynthesis:

Wavelength Intensity Geometry Duration

Swiss chard leaves:

Structural relationship to illumination Structural relationship to atmosphere Temperature during photosynthesis Preparation Growing conditions

On-line target gas system:

Pressure effects Humidity Purity 64

Table 3-4

CURRENT 1;lc-glucose PRODUCTION DATA

Time Measured Conditions (min) activity (mCi) 0 0 Start 30 yA on target with leaf chamber purged and valved off 10 Lights on and leaves on-line for 20 min illumination 30 15 Target material at end of run (calculated from well counter measurement at 70 min) 25 IICO2 8as *-n circulating system at end of bombardment (calculated from 50 cm^aliquot in well counter) 100 Leaf uptake (leaf in purged 125 ml leaf flask measured in well counter) 40 71 Leaf flask containing leaf and solution after alcohol extraction and acid hydrolysis 16 Leaf flask after decanting (23% of previous measure­ ment remains in leaf residue and flask) 55 Solution decanted (into 125 ml flask for well counter measurement, usually decanted into 250 ml rotary evaporation flask) (34) Rotary evaporator flask placed on top of well counter (poor geometry causes falsely low activity reading which is about equal to actual activity at the end of hot lab chemistry procedures, providing easy check point because other measurements above are not rou­ tinely made, in order to reduce radiation dose during handling) 47 35 5 cm^of neutralized, concentrated, and millipore- filtered solution ready to leave cyclotron hot lab to be injected on magnesium silicate liquid chromatograph column 85 3.5 5 cm of solution ready for physiological injection after column elution, rotary evaporation, and milli- pore filtration

NOTE: 84% of the 140 mCi total activity generated in 30 minutes is lost in targetry and chemistry exclusive of radioactive decay, and the remaining 16% is reduced to the yield of 2.5% (3.5 mCi final product) by 55 minutes of radioactive decay (e’’^t =* 0.153). 65

Table 3-5

i;lC-GLUCOSE PRODUCTION DATA EXCLUSIVE OF RADIOACTIVE DECAY

Category Fractional breakdown Fractional losses of categories based on activity present at end of illumination

Target H e retained in screen 0.11 0.11 target

•^COj available to leaf 0.89 — 1.00

Leaf Uptake of 1*C02 ^ C 02 not used by leaf 0.20 0.18

1:LC activity in leaf 0.80 — 1.00

Chemistry Activity remaining in leaf 0.23 0.16 flask (primarily leaf residue after decanting following ethanol extraction and acid hydrolysis

Losses in rotary evaporation, 0.15 0.11 pH neutralization, and milli- pore filtration (cyclotron hot lab)

Column elution for separation 0.40 0.28 of compounds other than glucose, final rotary evaporation,and millipore filtration (chemistry lab)

Final product ready for 0.22 physiological studies 1.00 0.84 66

Table 3-6 gives a comparison of the detector systems used in terms of experimentally determined relative efficiencies (34) determined with a

®^Sr flood source. Sixty-four day ®-*Sr decays by electron capture (100%) and has a 514 keV gamma (100%) which is a convenient substitute for the

511 keV annihilation radiation of ^C. Two preliminary patient studies using an IV bolus injection of ^C-glucose have been done with the 6- probe system, and none have been done on the 26-probe system, which has just been constructed and is still undergoing testing and interfacing.

For this reason, data based on non-regional cerebral metabolism studies on rhesus monkeys was extrapolated with the aid of the relative effi­ ciencies to predict requirements for ^C-glucose activity. Table 3-7 is based on studies achieving an average count rate of about 10,000 counts/sec per mCi at two minutes after IV bolus injection in rhesus monkeys using the 1-probe system (see footnote "b" in the table). In extending these results to patient studies, it must also be assumed that the cerebral geometry and head-to-body ratio differences between monkeys and patients do not have a drastic effect on the activity require­ ments. The only available correlation with patient data (given in foot­ note "c" of Table 3-7) shows an experimental count rate of 150 counts/ sec per mCi, which is 60% of the 250 counts/sec per mCi predicted by the extrapolation. The activity requirements shown in Table 3-7 for the

5000 count/sec criterion (explained in footnote "a") indicate that the presently available 3.5 mCi of ^C-glucose is adequate for all probe systems for a non-regional measurement obtained by summing probes. The Table 3-6

DETECTOR SYSTEMS AVAILABLE FOR 1:LC-GLUCOSE STUDIES

Probes- in Nal crystal size Individual Relative Relative Typical ^C-Glucose system Diameter Depth probe efficiency efficiency study ( cm ) collimators per probe for all probes summed

1 7.62 5.08 6 cm ^40 ^40 rhesus monkey (non-regional) straight hole

6 7.62 5.08 19-hole 1.0 6.0 patient (regional or non- convening regional)

26 5.08 5.08 7-hole 0.5 13.0 patient (regional with high converging activity, or non-regional)

26 5.08 5.08 1-hole 1.7 44.2 patient (regional with low 1-annulus activity, or non-regional) converging ( ( (

Table 3-7

PREDICTED COUNT RATES FOR n C-GLUC0SE STUDIES

a Probe Collimator cps/mCi 2 min cps from 3.5 mCi mCi required for 5000 cps system after IV bolus 2 min after IV 2 min after IV bolus injection bolus injection inj ection (per probe) (per probe) (per probe) (probes summed)

1 6 cm VL0,000b 35000 0.5 --- straight

6 19-hole 250c 880 20.0 3.3

26 7-hole 125 440 40.0 1.5

26 annular 425 1500 11.8 0.5

a5000 cps produces an acceptable standard deviation of about 4% for the present method of determining the cerebral metabolic rate for glucose. bBased on three rhesus monkey studies. Activity before injection was measured with a ringstand calibrator where the syringe containing ^C-glucose is 56 cm above a 5.08 cm Nal crystal which is 7.62 cm deep in a lead sleeve with a 8.26 cm hole. This geometry produces 7500 cps/mCi. c5/24/73 study of dementia patient with 6-probe system produced about 150 cps/mCi for each probe at 2 min after an IV bolus injection of about 2 mCi.

CO 69

probe system predicted to produce the best regional measurement statis­

tics is the 26-probe arrangement with annular collimators which requires

11.8 mCi for the 5000 count/sec criterion. Even though it is possible

that this predicted requirement will be reduced in the future after more

experience has been gained with the processing and analysis of regional

cerebral glucose metabolism data from patient studies, a factor of two

increase in the maximum ^C-glucose activity currently available would

seem to be a worthy goal. It should be pointed out that although the

5000 count/sec criterion is desirable, useful data has been accumulated

at much lower count rates, and it will be appropriate to evaluate all

the probe systems using the presently available activity level of 3.5 mCi.

This section has described an investigation of the ^C-glucose

production process which resulted in a change from a batch method to an

on-line method. Improvements in the wavelength, intensity, and geometry

of the illumination used to stimulate photosynthesis were evaluated in

a prototype system which was operated concurrently with the prototype

testing of the twin window assembly described in Section 2.2 and the

■^C-Bcreen target described in Section 3.1. The way in which the on- 11 15 line process is integrated with a C/ 0 tandem target and associated

labeling systems will be covered in Section 3.3, and suggestions for

improving ^C-glucose production further will be discussed in Section

3.4, along with related proposals for future work. 70

Section 3.3 Design of a ^ C / ^ 0 Tandem Target Assembly and Associated

Labeling Systems

This section begins with an examination of some alternate approaches to the design of target and labeling systems that will meet the criteria given in Section 1.4 (high activity, rapid target changes, reliability, purity of labeled compound, and low radiation dose to cyclotron oper­ ators) . It then describes the way in which the selected approach was combined with the prototype hardware described earlier to design new targetry and labeling systems. Experience with the measurement tech­ niques and prototypes discussed in previous sections was essential to the successful evolution of the new target and labeling systems.

The Washington University Medical Cyclotron is dedicated to the

I support of cerebral metabolism and blood flow investigators whose jin vivo studies require rapid sequential changes between various compounds labeled with -^C and ^--*0. Time for the original single-purpose target hardware to undergo radioactive decay following bombardment was re­ quired in order to permit the safe disconnection of water and gas lines and removal of the target assembly. Thus manual target changes imposed severe restrictions on the desired protocol for physiological experi­ ments. Another factor introducing delay in sequences of the three compounds labeled with ^ 0 was the use of a single common target gas circulating loop with a motor-driven manual valve array to select various furnaces and filters piped in a parallel arrangement. This required a few minutes to evacuate a gas mixture, switch to and pump 71

down the appropriate furnace, and fill the system with the next gas mixture required. Several methods of eliminating the type of delays described above were considered.

Magnetic switching of the beam, to parallel target stations was con­ sidered as a means of achieving a remotely operated change capability.

A rough cost estimate of $50,000 minimum was made (35), including

$12,000 for a bending magnet, $20,000 for a quadrapole pair, and the remainder for power supplies and other equipment. In addition to this cost, a concrete structural and shielding wall 3 feet thick would have to be removed in order to obtain space for the beam switching equipment and target stations. Space now occupied by a cafeteria in the adjacent hospital would have to be secured and extensive excavation and con­ struction carried out to build a shielded target room. The beam switch­ ing approach was quickly abandoned due to the prohibitive financial and space considerations.

A series target system was conceptually designed where self-con­ tained "rabbit" target assemblies could be inserted in the path of the beam preceding a conventional window and gas target. The round "rabbit"

targets were to be inserted and withdrawn pneumatically through a tube and valved inlet part, with rotational alignment maintained by an integral permanent magnet affected by the magnetic field present at the perimeter

of the cyclotron. These assemblies (appropriate for solid, liquid, or

powder target materials) were to incorporate their own vacuum isolation

window provisions and nipple connectors that would "plug in" to sweep 72

gas and cooling gas supplies as they came to rest after pneumatic in­ sertion. This concept was rejected because many difficult mechanical and vacuum problems were anticipated, which resulted in a prediction of an unacceptable reliability for routine operation.

The design approach selected was a tandem target system consisting of a H e section in series with an ^ 0 section. A ^®B(d,n)*'*'C produc­ tion section incorporating a B2O3 screen target on a movable target rod is preceded and followed by twin-window assemblies. For ^ 0 production, the screen target is withdrawn and the space between the window assem­ blies is evacuated. This permits the beam to pass through both window assemblies into the ^N(d,n)^^0 gas target. A communicating tandem assembly in which the second window is eliminated and the inserted target rod seals off the rear portion of the gas target was also con­ sidered. This would have the disadvantage of cross-contamination problems, and it was decided to pursue this design only if the beam energy degradation by the second window assembly could not be tolerated.

Figures 3-7, 3-8, and 3-9 illustrate the implementation of the

^ C / ^ 0 tandem target systems. The aluminum transition piece on the left bolts to the cyclotron vacuum chamber with an n0 "-ring seal and provides a mounting surface with "(V'-ring seal on the other end which is perpendicular to the beam. Water passages are provided for "(V-ring cooling. 3/8-16 brass studs above and below the beam opening are used 11 15 to locate and support C/ 0 targetry. The transition piece is also intended to provide a convenient mounting geometry for all future target ( ( (

Figure 3-7

Side View of Tandem Target Assembly Figure 3-8

Top View of Tandem Target Assembly Figure 3-9 Front Exploded View of Tandem Target Assembly 76

designs. Two twin-window assemblies shown to the right of the transi­ tion piece on either side of the hardware were described in detail by Section 2.2 and Figure 2-3. Their Delrin frames slip over the brass studs and have a thickness which provides proper nO"-ring squeeze on the window plate seals when the frames are clamped tight between adjacent surfaces. The hardware shown was described in detail by

Section 3.1 and Figure 3-2. Following the second window assembly is the 150 gas target consisting of an aluminum flange and body which bolt together with an "©"-ring seal. These two parts can be disassembled for cleaning, and for the purpose of modifying the body or completely replacing it without the expense of machining another flange. Design considerations for this target are presented in Appendix F.

Transition piece, windows, and the two targets are all separately elec­ trically insulated by nylon bushings in stud holes or by the Delrin frames. This permits beam current incident on targets or lost to beam passage openings to be monitored. Water cooling is provided in the vicinity of all "O^-rings and to beam strike surfaces of the targetry.

Aluminum was used for all components because it is non-magnetic and because activity induced by the beam is low and decays quickly follow­ ing bombardment (primarily 2.3-min ^®A1). Non-magnetic stainless steel or brass was used for threaded fasteners. A parts list and a set of 11 is 29 engineering drawings used to fabricate the C/ 0 tandem targetry are on file in the cyclotron facility control room. The way in which 77

this hardware is installed on the cyclotron and connected to associated labeling systems will be described later in this section.

In conjunction with the tandem targetry, multiple gas recircula­ tion and labeling systems terminating in a shielded labeling cave were designed and installed. A schematic of these systems is shown in

Figure 3-10. Four independent recirculating loops for three "^0- labeled gasej and terminate in the shielded labeling cave, and are used for the routine production of and *^0 compounds for patient 1 ft 11 and animal studies. The F and CO2 loops terminating in the chemis­ try hood are utilized for research on new labeled compounds. The glass target vessel in the lower left is connected to a short circulating loop for the production of for patient and animal studies. This loop is confined to the target area of the cyclotron vault, and is filled with CH^ target gas by a single line from the gas bottle mani­ fold in the hot lab. and ^ 0 loops were fabricated using 0.476 cm o.d. by 0.089 cm wall pure nickel tubing in 6 m lengths with silver- soldered joints. Leak problems have been noticeably less frequent than those encountered using the previous system of a single shared stainless steel loop with compression fitting connections. Soda lime and Drierite filters were relocated from the hot lab labeling area to the cyclotron vault and piped in with compression fittings instead of quick-disconnect fittings. This eliminated a source of leaks and reduced the radiation dose to personnel in the hot lab. Solenoid valves operated from switches activated by the four-position manual selector valve in the 78

labeling cave are arranged to keep all loops not In operation pumped down. This allows rapid remote changes to a different target gas to initiate production of a required labeled compound. Minimum volume to maximize specific activity was a consideration in the design of all components of the tandem targetry and associated labeling systems.

Table 3-8 gives the volume of each circulating loop and the various components comprising it.

The procedure for producing 11C-glucose described in Section 3.2 11 applies directly to the recirculating system for CO2 shown in Figure 15 15 3-10. The procedures for producing blood labeled with O-Hgb, C 0-

Hgb, and using the systems shown in Figure 3-10 are given below for the case of a typical patient study application.

•^0-Hgb: Pump out recirculating system with blood bubbler bag bypassed. Add N2 to reach 7.0 psia, add O2 to 7.5 psia, add N2 to 16.0 psia. This procedure results in a 97% N2/ 3% O2 mixture and insures the 02 is swept out of the gas manifold into the recirculating loop. The line from the target to the bubbler inlet side of the bypass valve contains a cold charcoal filter to remove ozone and a soda lime trap to remove contaminant CO2 . The line from the bubbler outlet side of ^.he bypass valve contains a Drierite trap to remove water vapor introduced by the bubbler which would otherwise be collected in the charcoal trap and made the loop diffi­ cult to pump down for leak checking purposes. Bombard tar­ get gas mixture at 30 yA for 4 minutes and then valve on-line a bubbler bag containing 4 cm3 of venous blood from patient. Continue to bombard while bubbling target gas containing 0150 through blood for 3 minutes. Shut down cyclotron and withdraw 2 cm^ of blood labeled with -'■^o-Hgb in a 3 cm^ syringe. Place syringe in calibrated well counter to measure activity of about 2 mCi. Place syringe in plastic "rabbit" container and send in pneumatic tube from cyclotron hot lab in basement to site of patient srudy on third or sixth floor. 79

Table 3-8

TARGET AND LABELING SYSTEM VOLUMES

3 Component Volumes in cm for systems

U C02 h2150 0150 C150

Target 21 132 132 132

Gas lines, valves, fittlrgs 248 321 339 338

Peristaltic pump tubing 36 36 36 36

1000°C charcoal furnace 133

400°C charcoal furnace 16

Charcoal filter 16

Soda lime filter 16 31

Drierite filter _ 16 -

Subtotal 305 505 555 670

Leaf flask 135 Blood bubbler bag 40 40 40

TOTAL 440 545 595 710

NOTE: Volume of nickel gas line (0.476 cm o.d., 0.089 cm wall) Is 6.92 cm-Vm. Most direct path length from cyclotron target area to shielded labeling cave is approximately 15 m. The gas bottle manifold volume is 48 cnw, and the gas line volume from the manifold to the five-way valve in the label­ ing cave is 107 cm-*, giving a total fill system volume of 155 cm3. HD 400° CHARCOAL FURNACE |—5~1 COLD CHARCOAL r-SHIELDED, VENTED ------I F* I TO REMOVE 0 , 0j + 2C-2C 0 LABELING CAVE

| F3 | 1000* CHARCOAL FURNACE Lfs_HlE}^ ? — N2 + 3 % CO Hgb DRIER1TE TO ls00 REMOVE H20 f T H s l t= g | V P 1 VACUUM PUMP s — N2 + 3 % 0 2 ,#0 - H ; b r m SODA '10 l50 s + C - C‘s0 , L r iJ REMOVE C 02 e (Q) MANUAL VALVES s — N2 + 3 % 0 2 Blood © SOLENOID VALVE s II « AIR COOLED TV/IN WINDOW $ ( 2 .5 4 ft HAVAR FOILS) He + I % 0 2 0*1000 SWEEP GAS o-iooox "C -G lu © 0 VENTED

r-VENTED CHEMISTRY HOOD 1 6.2 Mt V «°B(d nrq,;«4N((J d BEAU Si2a ‘ — H e + 1 % 0 SWEEP GAS DRY "CO, C02-* 6.2 Me V t0Ntid,«y*F d BEAM — Ne

CYCLOTRON VAULT HOT LAB

Figure 3-10 00 Piping Diagram for Target and Labeling Systems o 81

C O-Hgb; Pump out recirculating system with blood bubbler bag bypassed. Add N2 to reach 7.0 psia, add CO to 7.5 psia, add N2 to 16.0 psia. This procedure results in a 97% N2/ 3% CO mixture. The line from the target to the bubbler inlet side of the bypass valve contains a 1000°C charcoal furnace to convert ol5o to C ^ O and a soda lime trap to remove con­ taminant CO2 . Bombard target gas mixture at 30 pA for 5 minutes and then valve on line a bubbler bag containing 4 cm^ of venous blood from patient. Continue to bombard while bubbling target gas containing C ^ o through blood for 1 minute. Withdraw 2 cm^ of blood labeled with about 1 mCi of C^O-Hgb and process as described under -^O-Hgb.

l^^O: Pump out recirculating system with blood bubbler bag bypassed. Add N2 to reach 7.0 psia, add 0^ to 7.5 psia, and N2 to 16.0 psia. This procedure results in a 97% N2/3% O2 mixture. The line from the target to the bubbler inlet side of the bypass valve contains a 400°C charcoal furnace to convert O-^o to C-^Og. Bombard target gas mixture at 30 pA for 30 seconds and then valve on line a bubbler bag containing 4 cm^ of blood from patient. Continue to bombard while bubbling target gas containing 0^502 through blood for 15 seconds. Withdraw 2 cm^ of blood labeled with about 10 mCi of H2 ^ 0 and process as described under l^O-Hgb.

Problems due to contamination in the -^C-glucose production process were discussed in Section 3.2, where leaks admitting atmospheric air with its normcl concentration of CO^ resulted in competition with the carrier-free in the photosynthetic process. Such leaks are also an important factor in the production of C^O-Hgb and ^^O-Hgb, where the introduction of contaminant CO2 will produce appreciable and result 15 15 in an unusable mixture of blood labeled with H 2 0 and the desired 0- 15 15 Hgb or C O-Hgb. Another source of unwanted CO2 in the 0 targetry is caused by the formation of a yellowish-brown carbon-oxygen polymer which 15 is thought to form primarily during C 0 production runs (49). High temperatures caused by the beam stopping in this deposit on the interior

surface of the gas target also generates an unwanted supply of CK^. 82

15 T c Soda lime traps in the C 0 and 0A-'0 loops are Intended to remove CC^ contaminant, but are not adequate to cope with conditions where a significant leak or a dirty target chamber persists for several runs.

Routine leak checking prior to runs, frequent replacement of the soda lime filter, and scheduled cleaning of the target and lines are all necessary to insure adequate purity and reliability in the production of C150-Hgb and 150-Hgb.

Figure 3-11 is a photograph of the top shelf inside the shielded labeling cave. ^C-glucose labeling equipment is on the left. The selector valve and solenoid switches for choosing one of the four circulating loops is in the center, and a blood bubbler bag on the right is in one of three possible positions (C^O, or 0 ^ 0 loops). A shelf below the one shown holds the oil bath and nitrogen cover gas supply used for the ^C-glucose extraction and hydrolysis steps.

Fluorescent lamps illuminate the cave shelves and a blower and ducting are used to vent away any radioactive gases released. Valves and elec­ trical switches are conveniently grouped and labeled to minimize time spent in the radiation field. As seen in Figure 1-1, the cave is in the corner of the hot lab with the open working side closely opposed to a wall, which permits most of the room to be used during a run without high radiation exposure.

Solid blocks of ordinary concrete (19.4 by 19.4 by 39.5 cm) were on hand and available for use in constructing the cave. The density was 3 measured at 2.02 g/cm for use in shielding calculations. A proposed Figure 3-11 Shielded Labeling Cave Equipment ^ w 84

design consisting of one course of block lined with up to 2.54 cm of lead was evaluated by assuming a maximum anticipated activity of 100 mCi of positron emitter in point source geometry on the inside surface, re­ sulting in a predicted dose rate on the outside surface of 8 mr/hr.

Shielding for gas lines running one foot above the drop ceiling tiles and down into the cave was also evaluated. Assuming a maximum antici­ pated circulating gas activity of 100 mCi resulted in a predicted 1.27 cm of lead to reduce the dose rate directly under the gas lines and

30 cm below the ceiling to 8 mr/hr. Based on these predictions the cave was built without lead lining with the intention of adding it as necessary. So far, 0.95 cm of lead lining the back wall of the cave and 0.63 cm of lead wrapped around the circulating lines has been found

satisfactory to keep the dose in the center of the hot lab below about

2 mr/hr during production of ^C-glucose and blood labeled with ^-*0.

It will be a simple matter to add thin sheet lead to both cave and lines

if dictated by future production requirements.

The U C/150 tandem target and associated labeling systems have been fully operational since early November 1972. Figures 3-12 and 3-13

are photographs showing the target hardware installed on the cyclotron with connections for water cooling, target and sweep gas supplies, air

supplies for cooling the twin window assemblies, and vacuum pumpouts.

The entire assembly can be removed intact from the cyclotron by unbolt­

ing the transition piece, opening four quick-release toggle clamps

(two for window air, one for sweep gas, and one for ^ 0 target gas), Figure 3-12

Tandem Target System Installed - Front Section oo Figure 3-13 Tandem Target System Installed Rear Section 87

breaking two quick-release hose couplings for cooling water inlet and outlet, and pulling two vacuum hoses off the target pumpout nipples.

The Delrin connector blocks held in place by the toggle clamps are split so that the upper half remains attached to the cyclotron and the lower half remains attached to the targetry tubing. Each connector assembly joins a pair of lines whose inlets and outlets are attached to compression fittings threaded into the Delrin. "On-ring seals are com­ pressed by the toggle clamp when the two halves of the block are joined.

An air cylinder for retracting and inserting the target rod is shown most clearly in Figure 3-13 (target rod withdrawn). The double acting cylinder is a Schrader #100-0024 Gold Medallion Miniature with 1.91 cm bore and 6.35 cm stroke, operated by a #41997-1116 four-way single­ solenoid operated valve with #4502 muffler speed control (Scovill Fluid

Power Division, Wake Forest, N.C.).

Labor and material cost for the targetry and labeling systems described in this section was approximately $5,000, which includes machine shop fabrication of all targetry (including ten extra window assemblies), purchase of valves and fittings, and purchase of 150 meters of nickel tubing. This total does not reflect the considerable install­ ation effort accomplished by the two cyclotron operators, which included fabrication of special piping and electrical hardware for the shielded labeling cave and other portions of the gas recirculating systems.

This section has described the implementation of new medical cyclo­ tron systems for the efficient production of compounds labeled with 88

1C and J0 , and represents the culmination of the investigations and pro­ totype development described in the preceding sections of this disser­ tation. It is felt that this design adequately meets the criteria for improved target and labeling system efficiency that were specified in

Section 1.4, and that it will function without major modification at the upper beam current limit of 80 yA which is the future potential for cyclotron performance proposed in Section 2.4. Section 4.1 will present a comparison of the performance of the current and original systems, and indicate the capacity of the current system to adapt to future needs.

Section 3.4 Future Work

Modifications to the basic tandem system for the production of compounds labeled with and -^O are expected to be minor in approaching a future goal of 80 yA of 7 MeV deuterons proposed in

Section 2.4. They include the possibility of helium cooling for the twin window assemblies (discussed in Section 2.4) and a change in ^ 0 gas chamber geometry (discussed in Section 3.3). If future tuning for higher beam current and energy involves large changes in the positions of the septum, deflector, and magnetic channel, a significant shift in beam position might occur. Since the targetry is designed to be per­ pendicular to the beam, this would require only a new transition piece.

An important area for future development is an extension of tan­ dem system to include uses other than production of and ^0. This 89

would result in a multipurpose system rather than a dual-purpose system, and would increase the capability to efficiently prepare for studies which are done on short notice. An example of an unscheduled study would be a patient suffering head trauma in an automobile acci­ dent, where cerebral angiography is necessary, and it is desirable to include a regional blood flow study using blood labeled with

This type of situation may arise on a day scheduled for research chemis­ try, when many bombardments of liquid or powder samples for a few seconds each may be scheduled. It would be desirable to continue such cyclotron utilization while on standby for a patient study, which could be accomplished if special purpose target rods were used in the section of the tandem system. The ^ 0 target could then be maintained in a leak-checked condition ready for a run on very short notice, since the target rod in use could be quickly removed and replaced by the dummy plug shown in Figure 3-2. Special purpose target rods could be designed for the bombardment of powders and liquids contained by

2.54 y Havar foil, and for solid targets used for reactions such as

^Na(d,p)^Na and ^^S(d,a)^®P. Provisions to protect the target flange from contamination may have to be considered.

Two gas targets other than the ^ 0 chamber used frequently are those for production of -*-®F using the ^®Ne(d,a)^®F reaction with neon 13 12 13 target gas, and the production of NH^ using the C(d,n) N reaction 18 with methane (CH^) target gas. A nickel F chamber is being fabri­ cated for use in the ^ 0 chamber position of the tandem hardware, which 90

will eliminate the inconvenience of removing the transition piece to 18 install the current single purpose F target. The chamber for CH^

target gas cannot be dealt with as simply because of the nature of the

present design. Figure 3-14 is a photograph of the glass chamber in

current use, which is clamped with an "C'-ring seal against a water-

cooled aluminum flange. The flange is attached to the transition piece

by two capscrews with heads sealing against "O^-rings in counterbored

holes in the flange. These capscrews pass through a Delrin twin window

frame and thread into tapped holes in the transition piece which were IT 15 used for the mounting studs in the C/ 0 tandem targetry. Since the

capscrew heads are inside the glass chamber, and the transition piece

is used with the targetry, the glass vessel must be removed

to detach the flange and make the transition piece available. Removal

of the intact assembly shown in Figure 3-14 by unbolting the transition

piece is not possible since the end of the glass vessel is very

close to the magnet frame and does not allow the transition piece

mounting studs to be cleared. These inconveniences are being circum­

vented by fabricating another transition piece with the flange slotted

to allow enough sideways movement for removal of the intact assembly.

■^NHg produced in the target gas is thought to react with metal sur­

faces struck by the beam, so this chamber made of glass for chemical

inertness and ease of cleaning was designed to be large enough to

provide for stopping the beam completely in the CH^. The resulting

size of 9.21 cm i.d. and 45.72 cm length precludes a tandem target

geometry because of the magnet frame Interference described above. Figure 3-14 Glass Target Chamber 92

An approach considered for achieving a shorter target without using an exit window was bombarding at target gas pressures up to three atmospheres, the limit for the twin window assemblies (see Section 2.2). 1*3 Yields of NH^ at pressures above atmospheric were evaluated and found to be drastically reduced (37), so this idea was not pursued. Another solution to the problem might be the design of a shorter gas chamber, possibly made of teflon, utilizing a gas-cooled 2.54 p Havar window on the exit end in order to bring the beam out at about 3 MeV without most of the undesirable surface effects occurring due to high local energy deposition in a thick metal beam stop. The cross section for 11 13 C(d,n) N drops rapidly below 3 MeV (24,38) so the chamber length could be reduced to about 25.4 cm with little reduction in yield. With an optimum cross sectional geometry, this type of design would result 3 3 in reducing the target volume from 3050 cm to about 500 cm and the 2 2 surface area from 1390 cm to about 460 cm . Undesirable quantities of carrier ammonia and other nitrogen compounds are formed by radiolysis

from nitrogen impurities in the target gas and on the vessel walls,

so the proposed reduction of volume and surface area are predicted to enhance the specific activity and yield of ^NHg (36). Dispersion of

the beam by the target gas prior to reaching the exit window of the proposed target will preclude using a standard twin window assembly.

Use of a larger window with its diminished ability to withstand diff­

erential pressure will probably mean that the proposed target will have

to follow any other target in a tandem system. This will not be a 93

disadvantage, since the most desirable applications would be in the 15 present 0 target location behind the target rod section, or to follow a specially designed ^ 0 target in a tandem system for rapid switching 15 1^ 15 between 0 and NH^ production. The latter would involve an 0 target chamber design with a standard twin window assembly on both ends to allow pumpdown and bombarding through both windows into the chamber containing CH^ target gas.

In concluding the discussion on future targetry for the production 13 of NHg, it should be pointed out that the proposed radiolytic mechanism 13 for the formation of NH- during bombardment of CH (39) is not well j 4 understood, and the proposed development of a teflon chamber with gas cooled exit window may be very difficult due to unanticipated effects.

Studies utilizing -^C-glucose are now in a transition period between extensive experiments with rhesus monkeys to validate methods for measuring various aspects of cerebral metabolism, and possible frequent future application of the developed techniques in regional metabolism studies on patients. If investigations involving patients begin to occur steadily at the rate of more than two per week, it may be desirable to consider purchasing duplicate equipment for the liquid chromatographic purification indicated in Figure 3-6. This part of the glucose synthesis is now done with equipment in a sixth floor chem­ istry lab, with the attendant inefficiencies of an elevator transfer from the cyclotron location in the basement using a shielded container.

By equipping the cyclotron hot lab, the entire ^C-glucose synthesis 94

could be carried out by the cyclotron operators while performing con­ current production of other labeled compounds. This would free the chemists and chemistry lab for other work, and increase the final -^C- glucose activity by eliminating an unnecessary transfer. Cost of the necessary equipment is about $6,000 (36). If the decision is made to pursue this installation, the entire synthesis process (including those portions now done in the hot lab) should be reviewed with regard to the feasibility of semi-automation and with emphasis on the shielding of equipment. Automation of the present hot lab portion of the synthesis was considered out of concern for radiation dose to the operators and chemists, but the shielded labeling cave and use of convenient fixtures and efficient procedures proved adequate. This adequacy may not contin­ ue if the entire "^C-glucose synthesis is centralized in the hot lab and used frequently for patient studies requiring the high activity pre­ dicted by Table 3-7.

The eluate from the sixth floor chromatograph column is presently passed through a tubing loop adjacent to a shielded sodium iodide crys­ tal, and the activity monitored by means of a scaler, digital print­ out, and strip chart recorder. If this equipment were available in the hot lab, it could also be used in conjunction with single or multiple small sodium iodide probes to monitor and record activity in the gas circulating loops during any type of production run. Such a capability 11 would have been useful in the Investigation of the uptake of CO2 by

chard leaves (Section 3.2) where shielded detector loops on the inlet 95

and outlet side of the leaf flask could have been used to determine the 11 extraction fraction of (X^ under different experimental conditions.

If the postulated future application of ^C-glucose to regional cerebral metabolism studies in patients does materialize, the stringent requirements for high activity outlined in the discussion of Table 3.7 will demand tnat every possible means be used to minimize radioactive decay losses. One way to approach this is to scrutinize the production process and the protocol used for the physiological study for procedures which contribute to unnecessary delays. In order to do this on a con­ tinuing basis, the cyclotron operators and investigators conducting the study should routinely record the activity and time at key points up to the time of physiological injection. This type of data has been collec­ ted in the past by the author for the purpose of investigating the ^ C - glucose production process (see Section 3.2) but it has not been neces­ sary for the cyclotron operators and the investigators conducting the studies to be involved in routinely recording such information on a basis that will allow detailed analysis. Activity has been adequate to accomplish useful studies even though some unnecessary time delays go unnoticed. One such example is the observation that during rhesus monkey studies the ^C-glucose injection frequently occurs about 20 minutes after the injectate is available, resulting in .a factor of two loss in activity and the possible decomposition of some labeled glucose due to radioly3is. Requirements for activity for future regional patient studies will press the capabilities of the cyclotron and label­ 96

ing chemistry to the limit, making unnecessary time delays intolerable.

Collaboration between the cyclotron operators and investigators on a consistent basis for routinely recording and reviewing activity and time data as suggested above should accomplish an awareness of pro­ cedural delays resulting in activity losses. Emphasis of this approach is probably most necessary in the case of ^C-glucose, but might also prove useful in relating production techniques to experiment protocols 15 in the case of compounds labeled with 0 when rapid injection sequences are desired to follow rapid drug-stimulated physiological changes.

This section has discussed the modification and extension of the current ^ C / ^ O target and labeling systems to include research chem- 15 1ft 2A istry applications and the production of NH3» F, Na, and P.

Improvements in the equipment and procedures for the production and utilization of -^C-glucose were also discussed in the context of future requirements dictated by the anticipated emphasis for regional cerebral metabolism studies on patients. CHAPTER 4

CONCLUSIONS

Section 4.1 Evaluation of Equipment and Procedures Implemented to Achieve Efficient Production of Compounds Labeled with -^C and

This dissertation has presented work culminating in the successful routine operation of new target and labeling systems producing com­ pounds labeled with ^ C and ^ 0 at the Washington University Medical

Cyclotron. These systems were developed to provide production modes dedicated to in vivo studies of regional cerebral metabolism and hemo­ dynamics in patients and animals. Criteria for efficient performance adequate to meet current requirements included improvements in activity, reliability, purity of labeled compounds, and rapid delivery of

sequences of labeled compounds. The methods implemented to satisfy

these criteria were presented in previous chapters, and are briefly

summarized in Table 4-1. A comparison of the performance of the

current and original systems is given in Table 4-2, along with an in­ dication of the goals felt to be attainable with the continuing program of improvements suggested in Section 4.3. The current performance

represents a twofold to threefold increase in the maximum activity of

all labeled compounds produced, with another factor of two increase

97 98

Table 4-1

SUMMARY OF METHODS IMPLEMENTED TO MEET REQUIREMENTS FOR

EFFICIENT PRODUCTION OF COMPOUNDS LABELED WITH 1;LC AND 150

High activity and rapid production (rapid serial measurements with good statistics) (1) High stable beam current delivered to target with minimum energy attenuation (a) gas-cooled twin window (b) cyclotron improvements (septum, deflector, ion source, dee collimator, magnetic channel, vacuum system) (2) -^C screen target (3) |^0 target (4) 11C-glucose photosynthesis improvements (5) Minimum volume independent labeling systems

High reliability (1) Six independent brazed nickel labeling systems improving chemical purity by reducing leaks and cross-contamination (2) Cyclotron improvements (see above)

Rapid target changes (serial measurements and concurrent experiments) 11 1S (1) Tandem C'/ 0 target system (2) Automatic pumpdown of all labeling systems not in use (independent gas lines)

Low personnel radiation dose (operators and chemists) (1) Remote tandem target changes (2) Quick-disconnect targetry (3) -^C-glucose on-line production process (4) ShieJ.ding 99

Table 4-2

ACTIVITY, TIME OF PRODUCTION, AND RELIABILITY FOR COMPOUNDS LABELED WITH U C AND 150

Past Current Future

Beam current, yA 18 30-35 50-80 11 a C-glucose, mCi 0.5 in 3 in 6 in 135 min 85 min 85 min

^30-blood, mCi/2 cm3 H20b 10 in 90 sec 10 in 45 sec 10 in 22 sec 0-Hgb 1 2 4 CO-Hgb 0.5 1 2

Reliability 75% 90% 90%

Target changes 20-30 min 1 min 1 min

Personnel dose acceptable no change lower

^ h e production time for the current and future categories can be shortened to a minimum of 75 minutes by eliminating a ten-minute bombardment of the target before the leaves are valved on line. This would decrease the Hc-glucose activities to 77% of the values shown. bThe current and future categories offer little advantage in the case of blood labeled with H2^ 0 . It is possible to achieve over 50 mCi/2 cm^at 18yA for 6 minutes. This is much more than the maximum application requirement of 10 mCi/2 cm3shown in the table, which is produced in an acceptably short time for all three categories. 100

predicted for the future. Realization of the future potential depends mainly on the success of efforts to increase the beam current and energy of the cyclotron, as the target and labeling systems are pre­ dicted to function without major modification in the beam current range indicated in Table 4-2.

Section 4.2 Applications to Other Accelerators

It is recognized that many features of the equipment developed in this work were dictated by the characteristics of a particular medical cyclotron, and by the requirements for the production and delivery of labeled compounds in support of particular research programs. For this reason the details of the designs implemented apply directly at only one installation, another Allis-Chalmers medical cyclotron located at

Massachusetts General Hospital in Boston. This machine is very similar to the Washington University unit, with present maximum external beam capabilities in the area of 100 yA of 6.6 MeV deuterons (40). It pre­ sently suffers from one of the problems initially encountered with the

Washington University machine. Beam current and energy available for bombardment of targetry is restricted by the use of a 25.4 y aluminum window which Is held in place by "C'-rings in a water-cooled frame.

The Massachusetts General Hospital facility is also utilized for the production of compounds labeled with short-lived positron-emitting isotopes, including ^C-glucose and gases labeled with ^0, so all the targetry and labeling systems presented in this dissertation are of 101

interest there. A fruitful visit to Boston for detailed discussions resulted in requests for sets of engineering drawings and other mater­ ial that will expedite construction of equipment.

Section 2.1 presented a method for constructing a contour map of deuteron density distribution in a cyclotron beam. This quantitative autoradiographic technique should prove useful in target design and accelerator beam analysis for other charged particle applications.

It may also be of general interest to those interested in the quanti­ tative analysis of autoradiographs produced by any method. Section 2.1 also evaluated several qualitative beam mapping methods and recommended the use of a Polaroid autoradiographic technique as an accurate and convenient way of performing routine beam mapping. This information should be of interest to many accelerator installations.

The gas-cooled twin window assembly presented in Section 2.2 can find general application where minimum energy attenuation of an intense charged particle beam is important. A basic virtue of this design is the high mechanical strength inherent in the use of the silicone rubber bonding technique. This permits the use of thinner foils and higher differential pressure than would be possible with other mounting methods. Experience with these window assemblies has also established their attribute of long life, probably due to the good ability of the silicone rubber to withstand high temperatures and exposure to large doses of ionizing radiation.

The B2O3 screen target described in Section 3.1 is applicable at 102

several medical cyclotrons where gases labeled with C are produced using the ^B(d,n) or ^B(p,n) reactions. Attributes of this target compared to those used previously are less activity trapped in the

target, long life, and small size. A unique application made possible by this type of target is a design proposed for the production of 11 11 multi-curie quantities of CO and CO2 at the Oak Ridge National

Laboratory 218 cm Cyclotron, for use by the Medical Division of Oak

Ridge Associated Universities which operates an USAEC-sponsored cancer hospital in Oak Ridge, Tennessee. A prototype target has been fabri­

cated by forming molybdenum screen into a triangular tube about 0.32 cm

on a side and 5.08 cm long, which is coated with natural B2O3 (80%

■^B, 20% -^B). This target is placed in a tube of 0.457 cm inside diameter, which is in turn placed inside a larger tube with a water-

cooling annulus in between. The resulting assembly is positioned with

axis perpendicular to and intercepting the last internal beam orbit of

22 MeV protons in the 218 cm cyclotron. Helium sweep gas is to pass

axially past the triangular target, supplementing the water cooling and 11 11 11 carrying away the CO and CO^ produced by the B(p,n) reaction in

the B2O3 , which is melted by the beam and retained on the screen. This

design will present thick target geometry to the incident protons which

are degraded to 15 MeV by the outer tubes and water annulus. Internal

beam up to 2500 yA is available, but heat transfer considerations

would obviously prohibit using this capability. Preliminary calcul­

ations indicate that 250 yA, producing about 15 curies of circulating 103

gaseous activity in a 40-minute bombardment, will be feasible. This work is in the proposal stage, and experimental testing of the proto­ type target is contingent on the availability of funding.

Investigations of the photosynthetic process for labeling glucose 11 with C as described in Section 3.2 produced information regarding the parameters affecting leaf uptake of (wavelength and intensity of illumination, leaf geometry and temperature, gas atmosphere effects).

This data is applicable to several medical cyclotrons where "^C-glucose is synthesized using this method. It may also be of interest to plant 11 physiologists in evaluating the possible general application of CC>2 to the in vivo study of the photosynthesis process in any organism of interest.

The particular application of the tandem target idea to the and ^ 0 labeling systems described in Section 3.3 is very parochial to cerebral metabolism and hemodynamics investigations supported by the

Washington University Medical Cyclotron. However, several visitors interested in medical cyclotrons have indicated that the general tandem target philosophy is not widely used, and has many potential applica­ tions. Depending on the beam characteristics and production require­ ments of the individual accelerator, there is no reason to preclude consideration of the use of any tandem combination. Solid-gas, gas- gas, gas-solid, or solid-solid arrangements are certainly all possible, with the obvious requirement that a solid target in the first position must be retractable. The geometry of a target station with a window on 104

each end that can be pumped down to allow beam passage to a second station may provide an economical solution to the problem of rapid remote target changes in many situations.

This dissertation may prove useful to those organizations which have committed themselves to the acquisition of a new medical cyclotron facility, particularly if extensive in vivo studies utilizing short­ lived labeled compounds are planned. Limitations on beam current and energy should not be problems with a new facility, but an awareness of target and labeling system requirements and pitfalls will be highly desirable in the effort to design economical and efficient equipment.

Experience at this facility has included considerations such as relia­ bility, purity of labeled compounds, and rapid remote changes between various target and labeling systems. The information on these factors, along with general theoretical and experimental approaches to target design, should be helpful in the planning of a new medical cyclotron facility.

Section 4.3 Future Work

A major purpose of this dissertation is to provide the background information necessary for a continuing program of improvements at the

Washington University Medical Cyclotron Facility. Such an effort should ultimately lead to conditions where routine production runs can be carried out reliably at beam currents in the range of 50-80 pA with a deuteron energy of 7 MeV, thus realizing the predictions of Table 4-2. 105

It is also hoped that future development will accomplish an extension of the tandem target hardware from the present capability for ^ C and 15 0 production to include provisions for all isotopes produced. The details of methods proposed to achieve these goals have been outlined in the relevant sections of Chapters 2 and 3, with a summary included in Sections 2.4 and 3.4. These recommendations have been generalized and grouped into the six categories summarized in Table 4-3 for reference purposes.

It is hoped that the development and evaluation of the target and labeling facilities described in this dissertation have created a basic framework which can be improved and extended to meet the future require­ ments of the research projects supported by the Washington University

Medical Cyclotron. 106

Table 4-3

SUMMARY OF PROPOSED FUTURE WORK

Reference Sections Recommendations

2.1, 2.2, Schedule time for extensive tuning to achieve 80 yA 2.3, 2.4 of 7 MeV deuterons. Modify cyclotron components if necessary. Perform routine weekly measurement of beam energy (transmission in Al) and daily logging of oscillator frequency.

2.2, 2.4 Periodically test twin window assembly under the most exhaustive conditions possible as machine improvements make higher beam currents available. Develop helium-cooling capability and other modi­ fications if necessary.

3.3, 3.4 Develop special-purpose tandem assembly target rods, similar to the present ^-C target rod, for research chemistry bombardments (liquids, powders, solids) and production of 24fla and 30p.

2.1, 2.4, Perform contour map analysis of attenuated beam for 3.3, 3.4 present ^ 0 gas chamber. Correlate with Polaroid autoradiographs. Redesign chamber if necessary to accommodate beam changes resulting from modification of magnetic channel.

3.3, 3.4 Develop improved gas chamber for the production of which is compatible with tandem target opera­ tions.

3.2, 3.3, Centralize ^C-glucose synthesis in cyclotron hot 3.4 lab to efficiently meet the anticipated high activity requirements of frequent regional metabolism studies on patients. APPENDIX A

PROCEDURES FOR IN VIVO STUDIES UTILIZING

COMPOUNDS LABELED WITH 150 AND n C

Measurement of Regional Cerebral Blood Flow, Oxygen Metabolism, and

Blood Volume with ^~*0

About 2 ml of the patient’s blood is labeled with ^ ^ 0 and rapidly injected into a selectively cannulated internal carotid artery for the measurement of regional cerebral blood flow (rCBF). Recordings of the time course of the radioactivity in the brain are obtained by means of six collimated scintillation probes (residue detection). The labeled water bolus passing through the capillary bed of the brain diffuses rapidly and equilibrates with tissue water, and then egresses by wash­ out after the bolus has passed. Correction of the data for decay, count rate loss, and recirculation effects is followed by calculation of the mean transit time (or average residence time) of the labeled water, from which cerebral blood flow in ml/min per gram of brain tissue can be derivea (11).

An identical procedure using blood labeled with ^O-hemoglobin is performed for the measurement of regional cerebral oxygen metabolism

107 108

(rCtK^). The detector recordings In this case reflect the delivery of

labeled oxygen to brain tissue, metabolic conversion of part of the

oxygen into water, and the washout of labeled water from the brain.

Oxygen delivered to the tissue is assumed proportional to the maximum

count rate as the bolus arrives, which is followed by an immediate

decrease in count rate as unused ^"*0 flows out with the blood. The

remaining count rate, before being slowly diminished by washout of 15 water, is assumed proportional to the 0 involved in brain metabolism.

The ratio of count rate after rapid egress of unused ^ 0 to maximum

count rate obtained immediately after injection is then the fraction

of oxygen utilized by the tissue and converted into the water of metabolism. This fraction of injected ^ 0 utilized by the tissues in

the field of view of a probe is multiplied by the blood flow as measured

by a subsequent injection of labeled water, and by the arterial oxygen

content as measured from a sample of the patient's blood. The product

of these three quantities gives the regional oxygen utilization in

ml/min per gram of brain tissue.

Measurement of regional cerebral blood volume (rCBV) is accomplished 15 using the injection of the vascular tracer C 0-hemoglobin under the

conditions described in Table 1-6 and along with the other measurements

indicated there. The indicated cerebral hematocrit determination was

investigated in rhesus monkeys and is not repeated for patient studies.

Injections of blood tagged with any of the three labels described

above are typically performed about six minutes apart in patient studies, 109

and labeled water Injections in animals have been made as frequently as three minutes apart in order to follow rapid drug-induced physio­ logical changes. Probe count rates peak at about 7000 cps for the bolus injection of 1 mCi of labeled blood using a 6-probe system with

7.62 cm o.d., 5.08 cm deep Nal crystals with 19-hole converging colli­ mators. A new 26-probe system with 5.08 cm o.d., 5.08 cm deep Nal crystals with 7-hole converging collimators is predicted to require twice the activity for the same count rate (34).

Measurement of Regional Cerebral Metabolism with ^C-Glucose

The labeled compound is injected intravenously and monitored with head probes and with a detector placed over an external extension of an indwelling peripheral arterial catheter. Time course of the activity is recorded for two to three minutes. The activity detected by a head probe is from tissue (primarily brain tissue) and from tracer in the blood in all parts of the head seen by the probe. Head probe activity is corrected for blood activity by simultaneously recording the glucose activity in blood withdrawn with the peripheral catheter, and by recor­ ding the probe and blood activity during a subsequent intravenous in- 15 jection of blood labeled with the vascular tracer C 0-hemoglobin.

This correction and the calculation of the regional cerebral glucose metabolism in mg/min per gram of brain tissue is accomplished through procedures set forth in a mathematical model (41) and executed by an

IBM 360 program. The determination of glucose metabolism with a single 110

large head probe was performed in nine rhesus monkeys under various conditions. It was also determined by measuring flow using labeled water as described previously and multiplying by the arterial-venous glucose difference across the brain determined from blood samples. A significant correlation was found, and the resulting validation offers the potential for a low-risk means of measuring regional glucose metabolism ir. man when employed with a properly designed multiprobe system. The safety of this method is a drastic contrast with the highly invasive nature of the carotid catheterization required for the 15 measurements using blood labeled with 0 described previously. Count rates two minutes after an intravenous injection of ^C-glucose are about 10,000 cps/mCi for the 7.62 cm o.d., 5.08 cm deep Nal crystal with straight-bore open collimator placed below the head of the rhesus monkeys studied (42). Corresponding count rate for a 6-probe system with 19-hole collimators is predicted to be 250 cps per probe. The prediction for a 26-probe system with 7-hole collimators is 125 cps per probe. APPENDIX B

BEAM ENERGY MEASUREMENTS

Mean range of the deuteron beam In aluminum was measured using the target hardware shown in various views by Figures 3-7, 3-8, and 3-9.

The flange on the extreme left bolts to the cyclotron vacuum chamber and is followed by a beam window assembly positioned by a Delrin frame, a solid target chamber with retractable target rod, another window assembly, and a gas target chamber. For the purpose of the range measurement the first window is replaced with a solid aluminum plate containing an 0.63 cm collimator hole with a 1.27 cm counterbore 0.32 cm deep on the incident beam side for retaining thin aluminum absorbers.

The second window assembly is replaced by an aluminum plate with full window opening but no foil (shown in Figure 2-3) so that all the hard­ ware can be pumped down and communicated with the cyclotron vacuum.

Deuterons passing through the absorbers and collimator hole pass through the solid target chamber (target rod retracted as in Figure 3-8) and are stopped by the back wall of the gas target chamber. The gas target is insulated and connected to a microammeter so that transmitted beam current can be read for absorber thicknesses from zero through that re­ quired to completely stop the beam. Absorbers were pure aluminum foils

111 112

In combinations of nominal 9 y, 51 y, and 152 y thicknesses. Exact thicknesses in mg/cm were obtained by weighing 10 cm by 10 cm squares.

The results of this measurement are shown in Figure 2-1, which indicates 2 a mean range of 45.2 mg/cm corresponding to a deuteron energy of 6.13

MeV (43).

An earlier determination was made of total range in air, including the effects of straggling and energy spread of the beam. This was done by mapping the beam with blue cellophane in room air after it had passed through a 6.35 y (0.00025 in.) single Havar window at an angle of 52° with the beam. The cellophane ceased to be affected at a distance of 2 31.8 cm, which corresponds to 37.5 mg/cm of air for the prevalent tem­ perature and pressure conditions. This represents a maximum energy of

6.25 MeV (43), which was corrected downward to a mean energy of 5.75

MeV based on the 500 keV mean-to-maximum range spread in the aluminum transmission data. From Table 2-3, the window geometry used results in an attenuation of 466 keV for 6 MeV deuterons. Adding this to the 5.75

MeV mean energy produces a result of 6.22 MeV initial deuteron energy from the air data analysis. 12 1^ Another correlation was obtained through the C(d,n) JN reaction utilizing data from graphite plates used to make Polaroid autoradio­ graphs for beam mapping in air. This was done using a window consisting of two 2.54 y Havar foils normal to the beam and separated by a 0.63 cm air cooling space. Energy attenuation for this geometry is 376 keV 13 (see Table 2-3). N activity in the graphite plates ceased at 25.4 cm 113

in air, at which point the energy of the deuterons should be around 2 1.25 MeV (24,38). 25.4 eta of air corresponds to 30.0 mg/cm at the prevalent temperature and pressure conditions, and the range of 1.25 2 MeV deuterons is 2.8 mg/cm , so the range of the deuterons. exiting 2 from the window would be predicted as 32.8 mg/cm , corresponding to an energy of 5.84 MeV (43). Adding the energy attenuation of the window produces a result of 6.23 MeV initial deuteron energy. 15 A final correlation was obtained from the N(d,n) 0 reaction utilizing data from NH^NOg cylinders irradiated at increasing distances

in air. This was done using two twin-window assemblies of the type

described in the preceding paragraph. The purpose and details of

these measurements are described in Appendix F. It was found that

150 activity ceased at 22.9 cm, at which point the deuterons should be 2 around 0.80 MeV (44,43). 22.9 cm of air corresponds to 27.0 mg/cm , 2 and the range of 0.80 MeV deuterons is 1.5 mg/cm (43), so the range

of the deuterons after passing through two twin windows would be pre- o dieted as 28.5 mg/cm , corresponding to an energy of 5.33 MeV (43).

Adding the energy attenuation of the two twin windows produces a result

of 6.08 MeV initial deuteron energy.

Table 2-1 is a comparison of the four beam energy determinations

described above. APPENDIX C

BEAM MAPPING MEASUREMENTS

Various techniques for determining the spatial distribution of the beam were evaluated for accuracy and convenience. Figure C-l is a photograph of the results of several approaches to beam mapping.

On the left is a series of 0.230 mm thick white paper cards bombarded at 5 pA for 10 seconds each. They were spaced at one inch intervals in air starting one inch from a single 6.35 p Havar window. The top card separated from those below it was bombarded at 20 pA for 17 seconds. It was located at the same position of 30.5 cm in air as the card directly below it.

Just to the right of the white paper cards is a series of 0.025 mm thick light blue cellophane sheets in cardboard frames. These were bombarded at 5 pA for 5 seconds in the same positions as the white cards discussed in the preceding paragraph. The three groups of cellophane patterns to the right of those just described were made at various deflector voltages and beam currents in an unsuccessful attempt to detect small shifts in beam location.

A tray for holding beam mapping materials in place during bom­ bardment is shown in the upper left corner. Two rods containing machined notches to locate the cards or plates are attached to a

114 Figure C-l 115

Beam Mapping Techniques 116

flange compatible with the window assembly. An aluminum tray supports

the lower edge of the beam mapping materials. The tray contents in

the photograph are (from left to right) cellophane, white paper,

1.59 mm thick graphite plate, and a small square aluminum block suppor­

ting a 0.6 cm o.d., 5.0 cm high vertical cylinder of NH^NO^ which was

referred to in Appendix B. Adjacent to the tray on the right are

three beam patterns made at the same location with (top to bottom) white paper, an unexposed developed Polaroid print, and blue cello­ phane .

The vertical row of white-on-black patterns in the center of

Figure C-l are autoradiographs made with graphite plates activated

by the ^C(d,n)^% reaction. 10.0-minute decays with the emission

of a 1.19 MeV 3"*" and the accompanying two 511 keV annihilation photons,

exposing Polaroid Type 55P/N 10.2 by 12/7 cm (4 by 5 in.) black and

white sheet film which was developed in a Polaroid #500 Holder for 20

seconds. The graphite plates were bombarded at 12 pA for times vary­

ing from one second to five seconds as the distance in air was increased.

Exposure was started when the radiation level from a plate reached

240 m /hr 3 + y on contact, and varied from two to six minutes as the

distance in air increased and the pattern became more dispersed.

Two patterns are shown extending to the right from the center of

the vertical Polaroid strip. These were made with an activated gra­

phite plate containing ten holes varying in diameter from 0.33 mm to

3.17 mm. The smallest hole that could be resolved on the autoradio­ 117

graph was 0.53 mm. Based on. this test, 1.00 mm holes have been rou­ tinely used as a means of providing convenient landmarks for referen­ cing and correlating autoradiograph patterns made with activated graphite and aluminum plates.

Three patterns extending to the left of the vertical Polaroid strip are white paper patterns made at the same location as the auto­ radiographs they are adjacent to. Note the drastic difference in size, which will be discussed later.

Directly below the graph on the right side of Figure C-l is a series of exposures made on Kodak Type M Industrial X-ray film with a single activated graphite plate. The graph is a curve of the maxi­ mum optical density of each pattern as a function of the exposure of each pattern. Exposure was calculated from the relative radioactivity at the start of each exposure period, the half-life of the activity, and the exposure time. Optical density was obtained from a measure­ ment of percent transmission using a Baird-Atomic Microphotometer

Model No. RC-2, with a 0.4 mm diameter light beam setting.

The center and upper group of autoradiographs along the right hand edge of Figure C-l were done in a manner similar to that de­ scribed in the preceding paragraph, except that Kodak Type AA Industrial

X-ray film was used. The group at the lower right hand corner of the photograph was done using Kodak Fine Grain Positive 20.3 by 25.4 cm

(8 by 10 in.) sheet film.

Work described by Figure C-l and the foregoing discussion provided 118

the experience necessary to develop and apply a method suitable to evaluate the various beam mapping techniques. These investigations indicated the. need for a quantitative autoradiographic method capable of mapping the deuteron beam density from its maximum value down to 1% of the maximum. Table C-l lists the films and developing procedures tested in arriving at a suitable choice. Figure C-2 gives the sen- sitometric curves for the conditions in Table C-l. The fast high- contrast x-ray film did not have the slope required to allow assess­ ment of density down to 1% of the maximum exposure. The slow low- contrast fine-grain positive would have required an initial exposure with a plate reading an estimated 100 R/hr on contact. Results from the Ilford Commercial test were unsuccessful because of low contrast which allowed the 511 keV annihilation radiation to contribute signif­ icantly to the film background surrounding the six highest exposures.

This could probably be remedied by going to a lower initial maximum exposure and increasing the developing time to increase the slope of the curve, but the Kodak Commercial film was tested instead due to the ease of obtaining it locally and the availability of fairly complete sensitometric data on various developers and developing time/tempera­ ture combinations.

Figure 2-2 is a contour map of the beam constructed using the film calibration curve for Kodak Commercial shown in Figure C-2 to evaluate the maximum exposure beam pattern. The percent transmission was mea­ sured for the points corresponding to 1%, 10%, 50%, 90%, and 100% 119

Table C-l

FILMS AND DEVELOPERS FOR QUANTITATIVE AUTORADIOGRAPHY

Film Developing conditions

Kodak Industrial X-ray 7 min at 20°C in Kodak X-ray Type AA-2, 25.4 by 30.5 cm developer

Ilford Commercial Ortho, 7 min at 22°C in Kodak DK-50 1:1 20.3 by 25.4 cm, ASA 80 daylight

Kodak Commercial #6127, 6 min at 22° C in Kodak DK-50 1:1 20.3 by 25.4 cm, ASA 50 daylight

Kodak Fine Grain 3 min at22°C in Kodak D-76 Positive #7302, 25.4 by 30.5 cm, ASA 1.2 daylight OPTICAL DENSITY 0 z 0 3 Film Calibration Curvesfor Quantitative Autoradiography 2 O RLTV EXPOSURE RELATIVE LOG 3+ y + /3

Figure C-2 OE AE N UFC, hr /h R m SURFACE, ON RATE DOSE OA INDUSTRIAL KODAK X-R A Y Y A X-R AA-2 COMMERCIAL KODAK

BACKGROUND POSITIVE OA FN GRAM FINE KODAK COMMERCIAL ORTHO ILFORD 2

» 120 121

exposure at nine vertical sections of the beam and at sufficient points on the ends to close the contour lines. Figure C-3 is a detailed ver­ tical profile of the beam at one of the sections used for the contour map, and indicates the relationship of the transmission, density, and exposure parameters.

The autoradiographs used to construct the contour map were made 12 13 using a graphite plate activated by the x C(d,n) N reaction as de­ scribed in the preceding discussion of Figure C-l. The film was placed in a 20.3 by 25.4 cm Ilford mammography holder with a 0.18 mm thick vinyl cover on the exposure side. This holder is convenient because

it is reusable and easily closable with a flap held in place with

Velcro fasteners. A measurement of the vinyl cover gave a result of 2 21 mg/cm , which is thin compared to endpoint and mean ranges of

512 mg/cm^ and 134 mg/cm^ (46) for the positrons with endpoint and mean energies of 1.190 MeV and .448 MeV (47). Aluminum plates 27 28 activated by the Al(d,p) A1 reaction were also used with this holder 28 to make commercial film autoradiographs. 2.31-min A1 decays with the

emission of a 2.85 MeV 8~ and a 1.78 MeV y (both 100%), and produces 1 3 an autoradiograph with poorer resolution than the JN type described

above (see discussion of Figure C-7 later in this section).

Figures C-4 through C-8 represent an evaluation of five different

qualitative techniques for beam mapping. Material characteristics,

irradiation conditions, and film exposure times in the case of the

autoradiographs are all given on the figures. An actual size celluloid DENSITOMETER SCAN INCREMENTS (I UNIT= 0.219 mm) ro bi LOWER LOWER EDGE ± • • WINDOW WINDOW UPPER EDGE 10 10 r AND AND EXPOSURE (E) CENTERLINE WINDOW OPTICAL DENSITY (D)«log FRACTIONAL TRANSMISSION (T) TRANSMISSION FRACTIONAL i o o> ro 0) 00 ro H CO > Z o *n 30 S 3 < m o o € m 30 n z m m 30 H O > o o m 3 *

Figure C-3 Densitometer and Beam Density Curves for a Vertical Profile ZZl ( ( (

0.230 mm WHITE PAPER Irradiation time Vertical % Horizontal %

1 sec 90-100 90

2 sec 90 50

3 sec 50-90 30

5 sec 50 20

Deuteron Beam Current *> 5 yA

Figure C-4 123 Evaluation of Beam Patterns on White Paper 0.178 Iran POLAROID PRINT Irradiation time Vertical % Horizontal %

1 sec 90-100 50-90

2 sec 50 50

3 sec 10-50 10

5 sec 10 10

Deuteron Beam Current = 10 yA

Figure C-5 Evaluation of Beam Patterns on Polaroid Prints ( f (

t; 0. 025 mm BLUE CELLOPHANE

Deuteron Beam Current = 3 yA

Figure C-6 125 Evaluation of Beam Patterns on Blue Cellophane (

ALUMINUM PLATE AUTORADIOGRAPH

Film exposure time Vertical % Horizontal %

10 sec 50 50

10-50 30 20 sec

40 sec 10-50 10

60 sec 10-50 10

Initial plate activity (2^Al) = 300 mr/hr at 7.6 cm Polaroid Type 57 black and white sheet film (10.2 bj 12.7 cm) developed for 15 sec

Figure C-7 no Evaluation of A1 Autoradiograph Beam Patterns ( ( (

GRAPHITE PLATE AUTORADIOGRAPH

Film exposure time Vertical % Horizontal L

13 Initial plate activity ( N) = 300 mr/hr y on contact = 65 mr/hr y at 7.6 cm Polaroid Type 57 black and white sheet film (10.2 by 12.7 cm) developed for 15 sec

Figure C-8 127 Evaluation of Autoradiograph Beam Patterns 128

overlay was made from the contour map shown below each group of beam

patterns, and used to evaluate the approximate horizontal and vertical

extent of the patterns. Results of this analysis are shown to the

right of each individual pattern. As an example, a 50 appearing under

the heading "Vertical %" means that the height of the pattern corre­

sponds to an envelope where the deuteron density at the top and bottom

of the pattern has dropped off to 50% of the maximum. A larger number

of closely spaced contour lines could have been used for this assess­ ment, but the variations due to Irradiation time or exposure time are

so drastic that more accuracy is unnecessary.

Paper, Polaroid print, and cellophane beam patterns (Figures C-4,

C-5, C-6) have limited usefulness for all but the crudest measurements

unless they have been evaluated with a contour map and are repeated

under exactly the same conditions at the same location. If these materi­

als are irradiated at a high enough dose to approach a 10% envelope,

they will burn through and sometimes catch fire, and the amount of

activity produced is sufficient to preclude handling for 10 to 15

minutes. Irradiations that produce an envelope in the 50% vicinity

are sometimes used to produce a cardboard or Polaroid print pattern

to see if the beam is centered vertically in the window, or too close

to the inside edge. The blue cellophane can be used to get a total

range in air (including straggling and energy spread) as described in

Appendix B, but this is not a particularly common or useful measurement.

These three materials are definitely not useful for beam mapping to 129

design a gas target, or for any other measurement where a knowledge of beam density distribution is important. The decision to test a twin window assembly with a 0.635 cm high opening (described in Appendix D) was based on cardboard and Polaroid beam patterns, which led to the observation that target activity was lower than with a 0.952 cm high window. If the information presented above had been available at that time, an autoradiograph beam pattern would have been made and the 0.635 cm high window assembly would have been eliminated without going to the trouble of building and testing it.

Polaroid autoradiographs made with an activated aluminum plate as shown in Figure C-7 are inferior to those made with an activated graphite plate as shown in Figure C-8 . Both plates were identically drilled with a pattern of 1 mm holes. The difference in energy of the

^®A1 radiation (2.85 MeV endpoint) and the radiation (1.19

MeV endpoint) causes a drastic difference in the resolution of the marker holes. Another advantage of the graphite plate is that the 3 28 10.0-min N activity is more convenient than the 2.31-min A1 activity for establishing the suitable exposure conditions described at the end of Section 3.1 (clearly resolved marker holes which occur with and indicate a 10% envelope). APPENDIX D

PROTOTYPE BEAM WINDOW DEVELOPMENT

The window in use from July 1968 through October 1971 is shown on

the left of Figure D-l in disassembled form. It consists of two 0.635

cm thick plates of aluminum which are 7.78 cm high by 17.46 cm wide.

The plates have a window opening 1.27 cm high by 5.40 cm wide surrounded by "0M-ring grooves on the inner mating surfaces. A foil of 6.35 p

Havar alloy (see Table D-l for properties) is sealed in place between

the elastomer "0M-rings by assembling the plates with screws around the perimeter. Foil cooling was improved where possible by directing re­

circulating target or sweep gas against the Havar surface. This window

assembly was sealed against an "0"-ring in the face of the beam valve

housing (shown in Figure 1-2), and was at an angle of approximately 52°

to the beam. The same assembly was used with a 25.4 p aluminum window

prior to the use of Havar foil (June 1965 through June 1968). Titanium,

tungsten, tantalum and molybdenum foils were also considered at the

time the change to Havar was made (July 1968), but they could not be

obtained in pinhole-free quality in thicknesses less than 12.7 p.

A prototype assembly which is interchangeable with the July 1968 design,

previously described, is shown in the center of Figure D-l. Cooling

gas is introduced and removed from the space between the 2.54 p foils

130 (

Figure D-l 131

Original, Prototype, and Current Beam Window Assemblies 132

TABLE D-l

DATA FOR HAVAR ALLOY

DESCRIPTION: HAVAR is a proprietary, non-magnetic, corrosion resis­ tant, cobalt-base alloy which exhibits high strength and a high fatigue endurance limit.

MANUFACTURER: Hamilton Watch Company, Precision Metals Division, Lancaster, Pennsylvania

SPECIFICATIONS FOR THIS APPLICATION: 7.6 m coil, as-rolled temper, 2.54 D thick ±5% by 8.89 ±0.32 cm wide, best effort basis for pinhole- free material.

COMPOSITION: Cobalt 42.5% Beryllium 0.04% Nickel 13.0% Manganese 1.60% Chromium 20.0% Tungsten 2.80% Molybdenum 2.0% Iron 17.86% Carbon 0.20%

A = 30.09 Z = 61.03

MECHANICAL PROPERTIES: As Rolled Ultimate Tensile Strength 18,300 - 20,400 kg/cm2

Yield Strength (at 0.2% offset) 16,500 - 18,600 kg/cm2

Hardness (Rockwell C) 44-50

Maximum Recommended Service 400°C Temperature

PHYSICAL CONSTANTS: Density 8.3 g/cm

Linear Coefficient of 12.5 x 10"6 cm/cm/°C (0-50°C) Thermal Expansion

Electrical Resistivity 91.4 microhn cm

Modulus of Elasticity 2,070,000 - 2,120,000 kg/cm (Tension) 4. 133

by 0.317 cm passages drilled In the aluminum 0.635 cm center plate.

Two outer plates of 0.317 cm aluminum are used with the foils bonded in place between the plates with an adhesive sealant material. The window size shown is 0.952 cm high by 6.032 cm wide. Work on the twin window design was initiated in July 1971, and by October 1971 the assembly shown in Figure D-l was in routine use, allowing use of the maximum available beam current of 40-50 yA.

Selection of a suitable adhesive/sealant and developing acceptable methods for assembling and disassembling the window hardware were the major problems encountered. Silicone rubber compounds were chosen for initial testing due to their desirable elastomeric properties, high heat and radiation resistance, low weight loss in vacuum, good resis­ tance to ozone and chemicals, and availability in a wide variety of formulations. Table D-2 lists those compounds evaluated.

Seven unsuccessful attempts were made to complete an assembly with RTV-112. This material proved too viscous and fast-curing, resulting in severe wrinkling of the fragile 2.54 y Havar foils.

RTV-602 requires the use of a catalyst, and all surfaces to be bonded must be primed. The uncured rubber viscosity is low enough and the pot life long enough to allow satisfactory assembly. Primer is very thin and can be easily applied to the foil surface surrounding the window area without wrinkling the foil. Drying time for the primer is about 10 minutes for handling and 30 minutes for proper bonding.

Six assemblies with window openings 0.635 cm high by 7.935 cm wide were 134

Table D-2

COMPARISON OF SILICONE RUBBER BONDING MATERIALS

Advertised Advertised Material Catalyst Viscosity pot life cure

G.E. RTV-112 Moisture 300 poises <5 min 24 hr in air (self-leveling)

G.E. RTV-6022 1% SRC-04 12 poises 8 hr 5-8 hrs at (easily pour- 80°C able)

G.E. RTV-602£ *5% SRC-05 12 poises 20-30 min 24 hrs at (easily pour- room temp. able)

Dow-Coming 734 Moisture 250 poises <5 min 24 hr in air (pourable)

aRequires use of G.E. SS-4004 Primer on all surfaces to be bonded. NOTE: General Electric products from Silicone Products Department, Waterford, New York; Dow Corning products from Dow Corning Corporation, Midland, Michigan. 135

completed and tested (one at a time ) using the temperature-activated

SRC-04 catalyst (1%). Cure rates were somewhat hard to control and varied from about 12 to 36 hours at an oven temperature maintained at

90 ±5°C. There was a problem with wrinkling of the foils when the assembly was cooled to room temperature, due to the coefficient of expansion of the aluminum being higher than that of the foil. There was also a problem with the rubber slowly running onto the window area due to the long pot life and the low viscosity during the oven curing.

These oven-cured assemblies were tested by installing them on the cyclo­ tron in the standard way between the beam valve housing and a target chamber. Target and cyclotron sides were pumped down separately, with the cooling space at atmospheric pressure. Three assemblies failed both foils within a second or two after target-side pumpdown was initi­ ated. The other three held the vacuum on both sides, but when removed for inspection it was noted that the initial wrinkling had become more pronounced. Overnight pumping on the three assemblies that survived resulted in two failures. The one successful assembly was beam tested at 30° pA for 6 minutes with an gas target. "^0 activity was about equal to that produced with the 6.35 p Havar single window using an

18 pA beam current. It was noted that the silicone rubber at the lower edge of the window opening was blackened. A beam pattern deter­ mination on the target-side surface of the window assembly indicated the beam was below the vertical center of the window, causing a problem with the 0.635 cm high twin window, but not with the 1.270 cm high 136

single window. The twin window assembly was leak-tested with soap solution by pressurizing the cooling space, which indicated a pinhole

leak in the area of the blackened rubber on the cyclotron side.

Due to the problems encountered with the over-cured window assem­ blies, the use of the SRC-05 catalyst (room temperature cure) was

investigated. The short pot life of 20 minutes necessary to achieve

a one day curing time turned out to be an advantage in that the unwan­

ted flow of rubber onto the foil surface in the window opening was

eliminated. The wrinkling problem was also eliminated since the

silicone rubber did not have to cure with the plates and foils held at

an elevated temperature. An assembly with a window opening 0.635 cm

high by 7.935 cm wide was cured for 48 hours. The window opening was

raised vertically so as to be centered on the beam. Foils were smooth

and tight before and after pumpdown and beam testing at 40 yA, but ^ 0

and activity generated at 18 yA were about three-fourths of that

produced using the 6.35 y Havar single window. A twin window with an

opening of the same size as the single window (1.270 cm high by 5.397

cm wide) was assembled, but failed during pumpdown. A twin window with

an opening 0.952 cm high by 6.032 cm wide was then tried, proving 15 11 successful in all respects. 0 and C activity was equal to that

produced with the single 6:35 y Havar window at 18 yA, and the foils

remained smooth and tight after runs at beam currents up to 45 yA.

This window assembly was used for all regularly scheduled cyclotron

runs for two weeks .at beam currents between 20 and 30 yA. It then 137

developed a pinhole leak In the target-side foil. The performance of this twin window design was felt to be adequate to justify further testing during regular runs to gain operating experience and make further improvements.

Non-destructive disassembly of the twin window hardware in a con­ venient manner was a requirement not easy to meet due to the properties of the silicone rubber adhesive/sealant. Several commercial products for achieving bond release and digestion of cured silicone rubber com­ pounds were obtained and tested. These are listed in Table D-3 along with results of a 24-hour soak on test plates. Ten prototype twin windows were disassembled by soaking overnight in WS-1 and prying apart with a knife edge. They were then soaked for a few hours in

MX-38 to allow removal of the primer and rubber residue by light scrubbing with a wire brush. This was a time-consuming procedure and there was some evidence that prying the plates apart was deforming them and causing vacuum leaks due to discontinuous bonding. To eliminate these problems, six 10-32 holes were tapped in both outer plates as shown in Figure D-l. It was then possible to break the bond without soaking by evenly tightening jacking screws against the center plate.

An overnight soak of the separated plates in MX-38 then allowed easy removal of the rubber and primer, followed by a hot water wash.

Assembly of the twin window assembly was made more convenient by scribing lines on the outer plate surfaces to coincide with the peri­ meter of the foil, allowing easy alignment of the unprimed window area 138

Table d -3

COMPARISON OF SILICONE RUBBER DIGESTORS

Material Results of 24-hour soak on test plates2

Diverstrip RTV-112 plates fell apart and rubber easily WS-1 Stripper rubbed off. RTV-602 plates forced apart easily but primer and rubber still tena­ ciously adhere. 25% evaporation.

Diverstrip Same results as WS-1. 100% evaporation in D-90A Stripper 8 hours (evaporated material not replaced).

Silicone Stripper RTV-112 and RTV-602 plates forced apart SXS/1 easily and primer and rubber removed with light scrubbing. Deep pitting and etching of aluminum plates. 15% evaporation.

CEE-BEE MX-38 RTV-112 plates forced apart easily. RTV-602 Silicone Digestor with slightly more effort. Primer and rubber removed with light scrubbing. 5% evaporation.

0.317 cm thick by 5.080 cm square aluminum plates bonded together .and immersed in 200 ml beaker of test liquids. NOTE: WS-1 and D-90A from Diversey Corporation, Chicago, Illinois; SXS/1 from Southern California Chemical Company, Santa Fe Springs, California; MX-38 from Cee-Bee Chemical Company, Downey, California. 139

when placing a rubber-coated foil on a rubber-coated plate. An assembly fixture shown to the right of the prototype twin window in Figure D-l was also fabricated so that each successive plate could be aligned on dowel pins as it was coated and placed in contact. All plate and foil interfaces are primed and rubber-coated before assembly.

Twin wirdows of the type shown in Figure D-l were used routinely for all runs from October 1971 to October 1972. During this time it was found that lapping the mating surfaces of the plates to achieve maximum flatness would reduce the likelihood of initial or later leaking and thus increase life of the assembly. It was also found that excess rubber samples left over from the assembly of a window seemed to increase in strength up to about three days, so a three-day curing time was tried. This resulted in an increase in average window assembly life of about one third. Another factor contributing to the success of window assembly procedures was the requirement for obtaining 2.54 y

Havar as free as possible of pinhole defects and rigorously inspecting it before use. Ordering a mill run (8 m or more) of 8.89 cm wide material and requesting production on a best effort basis for pinhole- free material was found to be far superior to ordering in 1 m lengths from stock. Inspecting the purchased foil in a darkroom with a flash­ light held against the side opposite the viewer completely eliminated

inadvertent assembly using foils containing defects.

Five air-cooled window assemblies processed in accordance with the above procedures each operated for three to eight weeks of routine 140

cyclotron operation at beam currents up to 40 yA before developing leaks in the silicone rubber bonding around the window opening on the target side of the foil. These prototypes provided the design bases for the window assemblies used with the new tandem target system de­ scribed in Section 3.3. The best prototype was used from April 11 to

June 13, 1972 at beam currents in the 30 to 35 yA range.

The twin window assembly used with the tandem target system is pictured in Figure D-l (right side) which emphasizes its small size compared to the prototype. This has the advantages of less area to lap and remain flat, easier disassembly and cleanup, shorter cooling gas passages, and foil size same as plate size for easy alignment. External dimensions are 5.080 by 8.890 cm with plate thickness identical to the prototype and a window opening 0.952 cm high by 5.715 cm wide. Figure

2-3 shows the window assembly with its Delrin frame which provides electrical insulation, mechanical location, and the air supply connec­ tion for window cooling. The center plate has an inlet and outlet nipple (0.475 cm o.d. nickel tubing threaded on one end) on which an

H0"-ring is placed. These plug into the air passage holes in the frame

(see Figure 2-3). The two thick plates shown on the right are blank and open-window dummy assemblies used for test purposes. APPENDIX E

PARAMETER STUDY OF UPTAKE OF 1;LC02 BY SWISS CHARD LEAVES

Figure E-l is a photograph of one particular array of chemistry hood equipment in the cyclotron hot lab which was used to study the

I I photosynthetic incorporation of CO2 into the leaves under the con­ ditions of the on-line process described in Section 3.2. Sweep gas 11 3 containing C02 is conducted to and from the 125 cm Pyrex flask by the two plastic lines in the lower right hand corner. A single 5 cm o.d. piece of Swiss chard leaf weighing approximately 0.5 gram is shown on the bottom of the flask with the sun side down and the shade side just under a stainless steel gas inlet tube. An intact chard leaf is shown in the lower right hand corner on the floor of the hood.

The lamp under the flask is a W-45 daylight-white fluorescent using

Sylvania phosphor #5656. It is wound in a 12.7 cm o.d. tight spiral

of 8 mm O.D. by 4 mm I.D. soft glass tubing (special order from Aristo

Grid Lamp Products). The lamp at the side of the flask is identical

to the lower lamp except for the phosphor which is Sylvania #7452.

Two Aristo power supplies for the lamps are shown to the left of the

intact leaf. The lower lamp supply is 45 mA at 650 volts and the side

lamp is 30 mA at 650 volts. Supply outputs are fixed but can be used

141 Figure E-l 142 Hot Lab Area Used for Prototype Development of 11C-Glucose Production 143

with either lamp. A Color-Tran Model LQF-6 focusing-reflector unit

(Color Tran Industries of California) using a tungsten iodide 650-watt

120-volt 3200°K G.E. Quartzline FAD lamp is shown in the upper left with a bulb-to-flask distance of 79 cm. Heating of the fluorescent lamps due to the infra-red energy radiated by the incandescent lamp is re­ duced by air cooling with the Dayton 3020 rpm fan shown to the left of the flask (Aristo recommends maximum surface temperature of 52°C for lamp tubing to achieve optimum efficiency in operation). The alcohol extraction step is performed by removing the gas line stopper from the leaf flask at the end of the illumination, adding 40 ml of 90% ethanol, and transferring the flask to a 150°C oil bath shown on the left. Nitrogen cover gas is supplied by the tube just above the mouth of the flask. After boiling for 5 minutes, 2 ml of 3N HC1 is added and boiling continued for 5 minutes to hydrolyze the sucrose present to glucose and fructose. At this point 15 ml of sterile ^ 0 is added to the flask, the flask is removed from the oil bath, and the liquid is decanted into a rotary evaporator flask for a concentration step.

The details of the remaining purification steps (indicated in Figure

3-6) have been published elsewhere (48) and will not be discussed since they were not studied or altered. Other items shown in Figure

E-l are the ethanol and sterile water containers with a 100 ml flask used for measuring and adding them, a 3 cm syringe for adding the HC1, a scalpel for cutting the circular piece of leaf, and long tweezers for inserting the leaf in the flask. A Victoreen Model 440 Survey Meter 144

with ranges of 3, 10, 30, 100, and 300 mr/hr (Victoreen Instrument Co.,

Cleveland, Ohio) shown in the center foreground was used to measure the activity present in the leaf flask. An identical meter (not shown in the photograph) was placed about six feet away from the leaf flask at an 8 cm long unshielded section of the gas circulating lines to monitor the activity of the gas due to this short section of exposed inlet and outlet line. The position of the meter shown in the photo­ graph was such that its reading during ^CO^ circulation with no leaf in the flask was the same as the meter placed adjacent to the exposed gas lines. The plastic gas lines connected to the leaf flask run about

60 cm to the right of the edge of the chemistry hood and terminate in a manual valving arrangement that allows the leaf flask to be isolated so that gas will recirculate through the target system only. The valves and lines from this point on were shielded with lead except for the short exposed section described previously.

Prototype studies utilizing the type of arrangement illustrated by the equipment just described were performed to investigate ^C-glucose production parameters. Those parameters assessed were listed in Table

3-3 , which provides an outline of the topics discussed in subsequent paragraphs. The parameter studies to be discussed suffered from the fact that they were performed concurrently with development of the prototype versions of the ^ C screen target, gas-cooled twin window 11 assembly, and on-line C-glucose system. Because of the changing nature of the components and conditions available at any given time, 145

it was impossible to use consistent techniques and results should be considered qualitative when compared to the on-line system finally developed.

Assessment of illumination variables affecting leaf uptake of

was initiated with the W31 (warm white) fluorescent lamp used with the trapping process. The lamp (described in the second para­ graph of this section) used a different phosphor than either of the lamps pictured in Figure E-l and had an inferior geometry of four rings of concentric tubes (array o.d. of 15.24 cm) with a 6.35 cm aperture in the center. This ring source lamp geometry was originally purchased as a standard available part (the spiral disc sources were special orders). Half a leaf (light-starved for 48 hours) was cut into two pieces and placed sun side down in the bottom of a 600 ml flask. Gas

* was admitted by a tube reaching just short of the bottom of the flask so that the flow was initially directed over the sun side. The flask was placed about an inch above the lamp for a 20 pA on-line bombardment for 20 minutes with concurrent illumination for the last 17 minutes.

The leaf pieces were then removed, rinsed in tap water, transferred to a small beaker, and placed in a well-type ionization chamber (Baird

Atomic Model 11-140, 17.8 cm o.d. by 17.8 cm high with 7.6 cm i.d. by

12.7 cm deep well, calibrated in mCi for positron emitters). The same test was repeated with the fluorescent lamp replaced by the incandescent lamp (described previously) at a distance of 76 cm, resulting in an activity ratio of 1.8 for fluorescent to incandescent illumination under 146

these conditions. The fluorescent test was repeated with the leaf pieces shade side down and an activity ratio of 2.4 was obtained for sun side to shade side. Another test was performed with square leaf pieces 2.54 cm on a side in the bottom of a 51 ml flask in order to get the entire leaf surface evenly and intensely illuminated. The bottom of the small flask could be placed directly against and be com­ pletely covered by the width of the four rings of tubes in the annular disc geometry of the fluorescent light source. A 10-minute bombardment at 14 yA with concurrent illumination was performed with the flask bottom directly on and also 10 cm above the lamp, resulting in a ratio of 9.3 for leaf activity. These simple tests indicated that wavelength and intensity of the illumination were important, and that it was de­ sirable to give priority to illuminating the sun side of the leaf sur­ face if both sides could not be illuminated equally. These observations were substantiated by textbook information on the action spectrum for photosynthesis (49) and by a discussion of the structure of plant leaves

(50). Relative effectiveness in photosynthesis for wavelengths in the visible spectrum is shown in Figure E-2, along with the emission spectra

for the two spiral lamps pictured in Figure E-l. Complete spectral

data for the original four-ring fluorescent lamp used in the tests de­

scribed above was not available, but it was specified as having a peak

at 480 nanometers with a span of 400-720. This peak is just above the photosynthesis peak at 460 nanometers, which it probably overlaps fairly well, but the spectrum probably drops off drastically in the region of 147

i

u. ac 100

a 80 LJ m t» u i o z < 6 0 . I d _ ac >• O m o w - ac ii, to I d 4 0 - I d Z X I d I d r- l i ( 2 I d > 2 0 I- O < H J O I d X I d ac o . ac 400 500 600 700 WAVELENGTH, nanomiftrs

P PHOTOSYNTHESIS ACTION SPECTRUM BF SYLVANIA 5 6 5 6 FLUORESCENT LAMP ( OAYLIGHT WHITE) RF SYLVANIA 7452 FLUORESCENT LAMP (COOL WHITE) T TUNGSTEN FILAMENT LAMP

Figure E-2 Emission Spectra for Various Light Sources Compared to Relative Effectiveness in Photosynthesis 148

the 650 photosynthesis peak. The four-ring fluorescent was thus judged to be clearly better for stimulating photosynthesis than the incandes- 9 cent lamp (curve labeled T in the figure) if the illuminance (lumens/m ) of the leaf surface was about the same for both types of illumination.

This will be verified in a comparison to be described later where an incident light meter was used. The sun side versus shade side behavior can be explained by noting that the typical structure of a leaf has the chloroplasts (large cytoplasmic organelles containing the green pigment chlorophyll and the yellow-orange pigments called carotenoids) located in palisade mesophyll cells just under the sun side surface.

The shade side of the leaf contains spongy mesophyll cells which com­ municate through intercellular spaces with holes (stomata) in the lower surface of the leaf which give access to the atmosphere. Thus the illumination test results and structural evaluation suggested 11 giving illumination priority to the sun side and (X>2 delivery pri­ ority to the shade side for the on-line process being developed.

The foregoing considerations prompted the purchase of the two spiral-wound lamps and the 45 yA power supply described previously in the discussion of Figure E-l. One of these lamps has a phosphor described by the vendor's spectral emission data received with the delivered lamp and shown in Figure E-2 marked BF (for blue fluorescent because the peak at 470 nanometers is in the 430-490 range of the color blue). The other has a phosphor described by the curve marked RF (for red fluorescent because the peak at 610 nanometers was an attempt to 149

approach the 650 nanometer peak in the photosynthesis action spectrum in the 640-700 range of the color red). These lamps and the power supply were obtained for the purpose of achieving a source of illumi­ nation closely matching the spectral distribution of the photosynthesis action spectrum along with an increase in intensity due to the improved spiral geometry and the higher operating current. Unfortunately, the phosphor marked RF in Figure E-2 fell short of what was expected, since it does not do a very good job of overlapping the 650 peak in the photosynthesis spectrum. The phosphor marked BF is an almost perfect match for the 460 peak in the photosynthesis spectrum and has the added advantage of good breadth in the direction of longer wavelengths.

A comparison test of the blue spiral and red spiral lamps was conducted in June 1972 with the equipment shown in Figure E-l using the lower spiral position with the 45 yA power supply and no incandes­ cent lamp. 30 yA bombardments for 20 minutes with concurrent illumina­ tion for the last 17 minutes produced a leaf uptake for the red lamp about two-thirds of that obtained for the blue lamp. Leaf uptake for the blue lamp was crudely estimated to be about double that which could have been obtained with the original four-ring lamp and 30 yA power supply. The original lamp burned out shortly after the new lamps arrived, and the prototype target had been improved since the earlier test, so it was not possible to predict this improvement accurately or verify it experimentally. Since the blue spiral lamp produced superior results, the geometry shown in Figure E-l with this lamp directly under 150

the sun side of the leaf and the red spiral as close as possible to the side of the flask was chosen as optimum. The 45 pA power supply was used with the blue spiral and the 30 pA supply used with the red spiral.

With the fluorescent lighting optimized, the remaining illumina­ tion geometry to be determined was an effective placement of the 650 watt incandescent lamp. About 90% of the power supplied to this lamp is converted into radiant flux, but only about 20% of the power is delivered as luminous flux. The very large infrared component of the radiant flux causes heating of leaf to be the major consideration.

This is not desirable since leaf temperature is a variable which has a significant effect on photosynthesis. A field study (51) indicated

that a plant native to Death Valley, California, achieved its optimum photosynthetic carbon dioxide uptake during noon sun illumination when

its leaf temperature was held at 47°C (117°F), and that uptake dropped

off about 20% at 37°C (99°F). The distance between the incandescent

lamp bulb and the leaf flask for the equipment arrangement shown in

Figure E-l was varied over a range of 46 to 91 cm, during a series of

eight bombardments at 30 pA for 30 minutes with the leaf chamber valved

on line and illuminated for the last 20 minutes. The leaf was badly

wilted at the 46 cm distance, and leaf activity was negligible. Uptake

increased with increasing distance, leveling off at around 79 cm and

falling slightly at 91 cm. Leaf temperature for the 79 cm distance was

estimated by inserting a mercury-in-glass thermometer into the leaf flask

through a rubber stopper, wrapping a piece of leaf tightly around the 151

bulb with thread and repeating the run. The inlet sweep.gas tube was positioned sc that its shadow shielded the thermometer shaft above the bulb from the incandescent lamp. Twenty minutes of illumination pro­ duced a rapid rise in temperature from the initial 25°C (77°F) to 35°C

(95°F) in four minutes, followed by a slower rise and leveling off at

39°C (102°F) during the last sixteen minutes. This run was also repea­ ted with the incandescent lamp off, resulting in a temperature rise from

25°C (77°F) to 28°C (82°F). The combined effects of the thermal and luminous radiation from the incandescent lamp at 79 cm increased the leaf activity about 50% over that obtainable with only the fluorescent lamps.

The relative illuminance (luminous flux incident per unit area) of the leaf surface by the blue spiral and incandescent lamps for the geometry shown in Figure E-l was measured with a photographic light meter. A Kalimar Model CdS-31 meter (Kalimar, Inc., St. Louis, Mo.) was used in the incident light mode with a diffuser bubble covering the CdS window. A reading of 12.1 on the high range exposure value

(EV) scale was obtained for only the blue spiral lamp with 45 yA power supply by holding the meter diffuser bubble over the lamp in the posi­ tion normally occupied by the leaf. A reading of 11.8 was obtained for only the incandescent lamp by orienting the diffuser bubble from the leaf position toward the lamp bulb 79 cm away. The color response of this meter peaked at 620 nanometers (according to the manufacturer's specifications) which is about where the spectral emission curves for 152

these two lamps are equal in Figure E-2. Since the illuminance con­

tribution of the lamps tested is about equal, the spectral emission

curves marked BF and T as drawn can be roughly compared to the photo­

synthesis curve on the same basis for the geometry discussed above.

The results of this relative illumination measurement substantiate

comments made earlier in this section indicating that the spectral

emission curves would explain the experimental data if the illuminance by fluorescent and incandescent sources were about equal.

The optimum geometry described above was used with the on-line

equipment shown in Figure E-l for all routine production runs from

June 1972 when it replaced the trapping method until November 1972 when the hardware was moved into the shielded labeling cave

described in Section 3.3. During this time several additional problem

areas affecting the reliability of the leaf uptake of were en”

countered and procedures to cope with them implemented. Another effort

made during this time period was the recording of activity at key

points during routine runs from the targetry through the synthesis to 11 the physiological injection of C-glucose in an animal or patient study.

These considerations will be discussed in the remainder of this appendix.

In the original trapping method described at the beginning of this

section, the illumination vessel containing the leaves was evacuated

for times varying from several seconds to around thirty seconds during

a normal procedure. Evacuation of the leaf flask used in the on-line

process was also done initially in order to fill the circulating system 153

with the helium sweep gas. This practice was soon abandoned when it was noticed that ieaf uptake of the ^ C O ^ was drastically reduced if the evacuation was performed several times. In one case the flask containing the leaf was inadvertently left pumped down for about two minutes, and leaf uptake of in the subsequent run was negligible for a twenty-minute illumination period. It is logical to assume that the cells in the leaf would be damaged by the sudden application of such a large differential pressure, so the evacuation step was elim­ inated by valving the leaf flask to permit purging with helium. The flask is now bypassed after a helium purge for one minute so that the remainder of the circulating system can be evacuated, filled, and operated to build up ^ C O activity prior to valving the leaves on-line for the illumination period.

Leaf uptake of in the original trapping system was enhanced by the addition of 0.2 ml of water to the illumination chamber. Main­ taining humid conditions on the leaf flask was also found to produce higher and more consistent uptake in the on-line system. This could be due to the response of the guard cells on the lower surface of the leaf which regulate the size of the stomatal openings and thus control the gas atmosphere admitted to the interior cells. Also, sufficient water must be available to the cells to provide the continuous supply of hydrogen ions and electrons needed for the photosynthetic incorpor­ ation of the atom of into the glucose molecule.. It was found that the leaf could be kept sufficiently moist by rinsing the flask 154

and leaf with water just prior to Inserting the leaf In the flask.

The water vapor present in the recirculating system did not cause any

problems for the "^C-glucose production process, but it was a disad­ vantage because the same peristaltic pump loop and stainless steel

circulating lines were used for ^ 0 runs. Runs involving ^"*0 produc- 11 tion are frequently made immediately after C runs, and any water not

removed during the pumpdown provides a large pool of non-radioactive

oxygen atoms with which the ^ 0 atoms in 0^ 0 , C ^ 0 , or can

exchange. The resulting sporadic decreases in activity of the ^ 0

labeled gases were very undesirable. Another problem caused by the

water vapor was deterioration of the 1000°C charcoal furnace used in 11 the production of CO. In an attempt to reduce these problems, a

water bubbler was inserted in the leaf flask inlet line and a Drierite

filter was placed in the outlet line in order to remove the water vapor 15 from the rest of the circulating system. This helped the 0 problems

but decreased leaf uptake of ^y 38 muc^ 88 a factor of two. A

detector placed adjacent to the filter during the runs indicated a

radiation level nearly equal to that obtained for the leaf flask,

revealing the ability of the filter to trap as well as water

vapor. Use of the Drierite filter was discontinued and the "^0 prob­

lems were tolerated until October 1972, when the shared circulating

lines were replaced by independent circulating systems for each labeled

compound (described in Section 3.3). 155

Contamination of the sweep gas by Improper purging and small leaks due to fittings, valves, and pinholes in the beam window were found to cause a drastic decrease in leaf uptake. This is not surprising when one considers that a typical activity of 100 mCi in the gas cir- 12 11 culating system corresponds to 6 x 10 molecules of CO^ as compared 15 3 to the normal concentration of 9 x 10 molecules of CO^ in a cm of

3 atmospheric air. Thus 1 cm of air in the system would provide 1500 11 times as many non-radioactive molecules as CO2 molecules com-

3 peting for photosynthesis. System volume was around 670 cm for the prototype target with leaf flask and shared circulating lines and 3 11 is around 430 cm for the current C target with leaf flask and separate circulating lines, and the pressure at the start of the run is set at approximately 1.1 atmospheres. A leak into the system due to a vacuum window pinhole can occur with a twin window since the cooling air between the foils is either at 1.3 atm (two window assemblies) or 3 ? 1.6 atm (one window assembly). A one cm leak of this type in a 500 cmJ system would raise the pressure only about 0.002 atm, which is not very noticeable even on the large-face absolute pressure gauge with smallest divisions of C.007 atm that is used to monitor the pressure. Pinhole leaks in the window adhesive/sealant and the foil occurred frequently until experience using the prototype assemblies resulted in better assembly and inspection techniques. Window leaks are not a significant problem with the current assemblies, but leaks out of the circulating system continue to require routine attention and maintenance. Careful 156

purging and a check to see that the system will maintain a pressure with

and without the window cooling air supply activated are necessary pre­

cautions before starting a ^C-glucose run.

A class of variables affecting the leaf uptake of "^COg which are

very difficult to separate from more frequently occurring hardware

problems are occasional seasonal variations in the growing conditions

of the leaves. The chard leaves are grown, cut, light-starved for 48

hours, and delivered to the cyclotron at least twice a week by a

cyclotron operator. A potted chard plant is shown in Figure E-3. The

plants are grown indoors in the winter and transplanted outside in

the summer. Around the end of June when the days are long and the

temperatures high, their sun exposure must be reduced by a properly

positioned shading cover. An example of the effect of the shading can

be given because of a problem which occurred in June 1972. Figure E-4

shows the detector readings made during ^C-glucose production runs for

the detector locations described in the discussion of Figure E-l. The

data for 6/13/72 is a typical run for the state to which the on-line

system had progressed at that time, and produced 60 mCi in the leaves

at the end of 20 minutes of illumination and a desirable glucose/

fructose ratio in the final product ready for injection estimated (48)

at around 5 to 1. This was just before about two weeks of very hot and

sunny weather during which the shading cover was not erected, and the

data for 6/27/72 is a run under identical conditions which reflects

the difference in uptake of thought to be due to this change. Figure E-3 Swiss Chard Plant DETECTOR READING AT GAS LINES OR LEAF FLASK, m R /h r 200 150 100 50 0» BA O TARGET ON BEAM 30/»A BPS LA FAK HOG ILMNTD EF FLASK LEAF ILLUMINATED THROUGH FLASK) LEAF (BYPASS -77 RN (ABNORMAL) RUN 6-27-72 • B1'2 U (TYPICAL) RUN B-13'72 e EF FLASK LEAF ■ ■ ■ - gas

nes e in l I ATR TR O BMADET mln. BOMBARDMENT, OF START AFTER E TIM Uptake of 0*A EM N AGT N CRUAE "CO, CIRCULATE AND TARGET ON BEAM A 30/* 1 ;LC0 Figure E-4 2 by Swiss ChardLeaves

20 530 25 35 159

Uptake was much less rapid at the start of the illumination, circulating gas activity was higher, and it took five more minutes of illumination to produce 56 mCi in the leaves. Glucose and fructose were estimated to be present in about equal amounts. The problems of abnormal uptake be­ havior and high fructose content faded away a few days after the sun shade was erected over the plants, and the detector data for production runs returned to the familiar pattern indicated by the 6/13/72 curve of

Figure E-4. This type of curve was frequently plotted during the develop­ ment of the on-line process, and could be useful in the future to diagnose a subtle problem, particularly if there is some reason to suspect the growing conditions of the chard plants.

The 6/13/72 curve of Figure E-4 also justifies the continued use of the 20-minute illumination period established in the original work by Lifton and Welch (33). With the batch process they found no improvement in ^C-glucose yield with illumination beyond 20 minutes, presumably because of the symthesis of some of the labeled glucose into starch, competing with the synthesis of new glucose. This result does not apply directly to the on-line process where a continuous supply of ^ C 02 activity is supplied to the leaves, but does lend support to terminating the illumination based on the empirical obser­ vation of the leaf flask activity curve. The._curve is leveling off at

20 minutes and coiltinuing further would produce little increase in leaf uptake, while promoting the starch competition problem. 1 60

11 Development of the basic on-line process for uptake of CO^ by

Swiss chard leaves, and problems encountered and solved while gaining experience with its use have been presented in this appendix. Current

■^C-glucose production data, obtained after this work was completed, is shown in Table 3-4 and 3-5, and discussed in Section 3.2. APPENDIX F

DESIGN OF GAS CHAMBER FOR THE PRODUCTION OF 150

15 Design of the 0 target body described in Section 3.3 was deferred until fabrication of the tandem components preceding it (including the gas target flange ) was complete. This allowed direct experimental determination of the beam path and ^ 0 yield parameters relevant to target design. The cross sectional dimensions of the gas target were determined by using Polaroid autoradiographs made with graphite plates as described in Section 2.1, using a beam mapping box attached to the gas target flange as shown in the upper left corner of Figure C-l.

Length of the chamber was determined by bombarding a solid material con­ taining nitrogen at increasing distances in room air, and measuring the

150 activity produced by ^N(d,n)^0. Attenuation of the beam by air is sufficiently close to attenuation by the 97% N£/3% Og target gas mixture to negate the need for correction factors. Finding a suitable solid was difficult, and the compound used was far from ideal from a standpoint of stability and ease of fabricating it into a desirable geometry. It was found that ammonium nitrate (NH4NO3) could be cast into cylinders 0.6 cm diameter and about 5 cm long in a split aluminum mold.

White granular crystals of NH^NO^ were put in the mold and heated to

170°C in air in an oven. As melting occurred more was added until the

161 162

mold was full. Slow cooling for about ten minutes was necessary to

avoid a large central void in the cast cylinders. The mold had to be opened at about 100°C to prevent the cylinders from sticking to the mold and breaking. Cast cylinders were trimmed to a uniform length of 3.8 cm and kept in a sealed jar partially filled with Drierite since they were hygroscopic and would start to swell noticeably in about thirty minutes.

These cylinders were held vertically in a hole drilled in a small alu­ minum block so that they could be positioned in the path of the beam.

The block with a NH^NOg cylinder in place is shown on the floor of the beam mapping tray in the upper left corner of Figure C-l. Height of

the cylinders was sufficient to include all of the beam strike, and the

cylinders were kept in the most dense and uniform portion of the beam

in order to minimize effects of horizontal dispersion. The NH^NOg

cylinders were bombarded at 5 yA for 5 seconds, and allowed to decay

for 5 minutes in order to allow the 66-second ^ F activity from the 16 17 0(d,n) F reaction to become negligible. They were then placed on

the surface of the metal end cap of a Victoreen Model 440 Survey Meter

for two minutes to verify that the remaining activity was 2.067-minute

150. The cylinders were rolled until the bombarded side was down and

the detector reading was maximum before recording the value at 7 minutes

after bombardment. Figure F-l gives the results of these measurements

as a function of distance in air from the surface of the second window.

The drop in activity with distance is accentuated by horizontal disper­

sion of the beam, so this data cannot be used to construct a true thick- mR/hr ON CONTACT FROM aO ACTIVITY IN NH4 NO CYLINDERS 7 min AFTER BOMBARDMENT 120 100 40 60 20 80 . 0 IDW N ,0 ADM AGT SEBY cm ASSEMBLY, TARGET TANDEM /,80 C " IN WINDOW ITNE N I FO SRAE F EOD TWIN SECOND OF SURFACE FROM AIR IN DISTANCE

150 from Yields Irradiation ofIn Cylinders NH^NO^ Air 4

• • 8

12 Figure Figure F-l

16

• • 20

24

28 163

164

target yield curve. It would have been desirable to have thick flat plates of NH^NO^ sufficiently large to contain the entire beam strike, but an attempt to achieve that geometry would have been very difficult and time consuming. A target gas path length of 19 cm was chosen based on the results of Figure F-l. The present target chamber width of 5.715 cm and height of 1.270 cm based on graphite autoradiographs made before the magnetic channel width was increased and before beam energy mea­ surements indicated a possible increase of 300 keV. These changes in beam characteristics (both discussed in Section 2.3) dictate a repeat of the gas target design measurements in order to evaluate the potential for increasing ^ O activity by modifying the target geometry. The present target body was designed with sufficient wall thickness to allow some increase in internal cavity dimensions, so it may not be necessary to machine a completely new part if a change in dimensions proves warranted.

Three 6-minute bombardments at 30 yA were made on 97% ^ 7 3 % °2 target gas to determine the total production using the new target. 3 Activity of a 50 cm aliquot of the gas in the circulating system was measured in a well counter 2 minutes after shutdown and corrections applied for radioactive decay and total system volume. The average value of 361 mCi was used in the equations presented in Appendix H, 14 15 producing a result of 76 mb for the average cross section for N(d,n) 0 between the initial energy of 5.38 meV entering the target gas at the window surface and the final energy of 2.14 meV striking the end of 165

the target chamber.

Literature on the ^N(d,n)^0 excitation function and yield (44,45,

27,28,31,52) was reviewed in order to calculate correlations with the measured average cross section. Variations were extremely wide, with cross sections in the 2 to 5 MeV region ranging from a nearly flat

22 mb (44) to a sharp increase from 100 mb at 2 MeV to over 300 mb at

3 MeV (27). It was not possible to correlate any yield data because of much different or unspecified bombardment conditions. APPENDIX G

RANGE AND ENERGY LOSS FOR DEUTERONS IN VARIOUS MATERIALS

The results of range and energy loss calculations were presented in Section 2.1 for beam energy determinations (range In aluminum, range

In air, activation thresholds after attenuation by air), In Section 2.2 for various window geometries (attenuation by Havar alloy, air, helium), in Section 3.1 (range in boron oxide), In Section 3.3 for the *^0 gas target (attenuation by nitrogen), and in Section 3.4 for the ^% H 3 gas target (attenuation by methane). This appendix presents the methods used in these calculations.

Range and stopping power data were taken from the 1966 compilation of Williamson, Boujot, and Picard (43). Mean ranges are the actual physical range of the deuterons in the absorber, and are not strictly equal to the physical thickness of the absorber because small angle multiple scattering will make the path traversed by the deuteron slightly

longer than the physical thickness of the absorber. This effect was

evaluated with an approximate correction for multiple scattering given by Williams (53 ): RmZ /M 167

where 8 mean range without multiple scattering correction

Robs ** mean ran8e with multiple scattering correction

Z 8 atomic number of stopping material

M 8 electron rest mass o 8 mass of incident particle

r * 1.7 from the approximate expression R(E) Er

The correction was negligible for most applications in this disserta­ tion since high-Z materials were not encountered. For example, it was about 0.1% for the range of 6 MeV deuterons in aluminum.

Range in compounds was evaluated using the Bragg-Kleeman rule (54) for relative range of charged particles in a compound with respect to range in an element. The element in the range tables were atomic number closest to the average atomic number of the compound was used. The

Bragg-Kleeman rule is fR0 i ■ fapi y o a,, r where RQ 8 range in reference element

Rl 8 range in compound

pQ 8 density of element

8 density of compound

A0 8 atomic weight of element

8 effective atomic weight of compound

from £ 7 - nlAi.* n2-A^ -:-- J 1 n 1J S ^ + n2Jk£ + ... 168

where n^, n^, ... are the atomic fractions of the elements whose atomic weights are A^, Aj, ...

Applying the relations given above to Ng, enriched I^O^, and air

(three materials commonly encountered in this dissertation) produced a 2 multiplier of 1.0167 to adjust nitrogen range in mg/cm to enriched boron oxide, and a multiplier of 1.0145 for nitrogen to air. For many applications it was sufficiently accurate to use nitrogen data for the other two materials without correction.

Ranges for energies intermediate between tabulated energies were determined by plotting data at several higher and lower energies. Tabu­ lated energy increments were 0.05 MeV for 0.00 to 1.00 MeV, 0.1 MeV for

1.0 to 3.0 MeV, 0.2 MeV for 3.0 to 5.0 MeV, and 0.5 MeV for 5.0 to 10.0

MeV.

Energy loss in gas targets and other applications where the loss was not small compared to the energy of the deuterons was simply calcu­ lated using the tabulated ranges:

R(Ef) » R(E±) - A(px) where R(E^) ■ range to be calculated for deuterons with unknown final energy Ef, g/cm2 2 RCEji) » range for deuterons of initial energy E^, g/cm 2 A(px) » thickness of target, g/cm

E^ corresponding to R(Ej;) is then used to determine the energy loss AE ■

E± - Ef MeV. 169

Energy losses for window materials were calculated utilizing stopping power which Is more appropriate when the energy loss Is small compared to the Initial energy of the charged particle beam. Using a method outlined by Williamson and Boujot (55) which expresses the energy of a particle traversing a thickness of absorber by a Taylor’s series expansion, the following relation results:

AE » [S(E)] [A(px)]{-1 + Jg[D(E>] [A(px)]} + ... where AE = energy lost in the absorber, MeV 2 S(E) = stopping power at incident energy E, MeV cm /g

D(E) = derivative with respect to energy of the stopping power, cm2/g 2 A(px) «• thickness of absorber, g/cm APPENDIX H

YIELD CALCULATIONS FOR 10B(d,n)1;lC AND l4N(d,n)150

The results of yield and average cross section calculations for

^®B(d,n)^C and ^^N(d,n)^0 were presented in Section 3.1 and 3.3. This appendix describes the methods used in these calculations.

An average cross section, 7, introduced by Ricci and Hahn (29) in considering the activation of thick targets by charged particle bombard­ ments, has been evaluated by Tilbury (31). Ricci and Hahn define

E / ae(dx/dE)dE T w ' - - —

E ^ d x / d E M E ( H ’ 1 J where oe is the energy dependent cross section, and insert the change in energy E with distance of penetration x (the stopping power) in matter for non-relatlvistic charged particles. This is given by the expression (56)

- g - A r e V NZln^ (H-2) dx m0V2 1

- (k/E)ln(E/I) (H-3) where ez is the charge of the bombarding particle, V is its velocity, mQ is the rest mass of the electron, N and Z are the number and charge of the target atoms, and I is the mean ionization potential of the tar­ get atoms.

170 Insertion of equation (H-3) in (H-l) gives

(° -1 _ E J aeE{Kln(E/I)} AdE a “ ------(H-4) g ^°E{Kln(E/l)}_1dE

For a given target, I is constant, so In (E/I) varies slowly with energy, or K In (E/I) is approximately constant, resulting in

(H-5)

To a good approximation, o’ is independent of any properties of the tar­ get materialr and is constant for a given nuclear reaction and a fixed bombarding energy. Ricci and Hahn tested these approximations and found good results for target Z *» 4 to Z » 95. They claim that these findings can be applied to reduce charged particle activation analysis to the simplicity of neutron activation analysis.

In numerical integration form, equation (H-5) becomes

(H-6)

where o^ is the cross section for the nuclear reaction at E^ and AE is a small increment of energy. This relation was used for numerical inte- gration under the excitation function for ^B(d,n)^C shown in Figure

3 - 5 using an energy increment of 0.1 MeV. The same figure shows the resulting ~a plotted for En up to 10 MeV. 172

For the case of a constant cross section, activity resulting from a charged particle bombardment can be calculated from a simple expres­ sion which assumes the particle beam strength is constant and only its energy is attenuated as it penetrates the material (57). This is given by:

A(t) - ION (REi"REf)

I ** intensity of beam in particles/sec

o » constant cross section for reaction of interest from final energy Ef to initial energy Ef, cm2

N » target atom density, target atoms/g material. 2 Rg a range of particles at Ef, cm /g material

9 Rjj a range of particles at E^, cm /g material

(l-e”^t) a fraction of saturation activity present at end of bombardment time t for nuclide produced which has decay constant A

For the case of the thick-target saturation yield curve for ^B(d,n)^C shown in Figure 3-5, equation (H-7) becomes:

A - IoNRg (H-8) where cr is the cross section defined by Ricci and Hahn and R^ is the to­ tal range for particles of energy E incident on the target material.

The average cross section for ^N(d,n) ^ 0 given in Section 3.3 was cal­ culated using equation (H-7). 173

Experimental yields measured for the ^B(d,n)^C target described

In Section 3.1 correlated well with the yield calculated by the Rlccl

and Hahn o' method (see Figure 3-5 discussed In Section 3.1). The

simplicity and apparent accuracy of this method deserve consideration when frequent and rapid calculations of production yields are required. 174

LIST OF REFERENCES

1. A.P. Wolf, D.R. Christman, J.S. Fowler and R.M. Lambrecht, Synthesis of Radlopharmaceutlcals and Labeled Compounds Utilizing Short-Lived Isotopes, Symposium on New Developments in Labelled Compounds including Radiopharmaceuticals, March 26-30, 1973, Copenhagen, Denmark, 1AEA/SM-171.

2. D.K. Bewley, R.J. Post, and D.J. Silvester, Eds., Symposium on the Use of Cyclotrons in Medicine, MRCUK 1969. MRC Cyclotron Unit, Hammersmith Hospital, London.

3. C.B. Amphlett, Ed., The Uses of Cyclotrons in Chemistry, Metallurgy and Biology, Butterworths, London, 1970.

4. J.S. Laughlin, R.S. Tilbury, and J.R. Dahl, Recent Advances in Nuclear Medicine, Chapter 3, Vol. 3, Progress in Atomic Medicine, (J.H. Lawrence, Ed.), Grune & Stratton, New York, 1971.

5. H.I. Glass, D.J. Silvester, Cyclotrons in Nuclear Medicine, Brit. J. Radiol. 43:589, 1970.

6 . B. Hoop, Jr., J.S. Laughlin, and R.S. Tilbury, Cyclotrons in Nuclear Medicine, Vol. II, Chapter 13, Instrumentation in Nuclear Medicine, (G. Hine and J. Sorenson, Eds.), Academic Press, New York, 1973.

7. W.J. Lorenz, K.E. Scheer, H.W. Lindner and M.A. Chaudhri, Cyclotron for Medical and Biological Research. I. The Importance of a Cyclotron for Nuclear Medicine and Nuclear Biology, Atompraxis 13: 251, 1967.

8 . E.E. Stlckley, Availability and Production of Short-Lived Positron Emitters by Compact Cyclotrons, Radiol. Clin. No. Amer. 7_:207, 1969.

9. M.M. Ter-Pogossian and H.N. Wagner, Jr., A New Look at the Cyclotron for Making Short-Lived Isotopes, Nucleonics 24:50, 1966.

10. D. Comar, D.J. Silvester, S. Rothchild and A.P. Wolf, manuscript in preparation for the IAEA.

11. M.M. Ter-Pogossian, J.O. Eichllng, D.O. Davis, M.J. Welch, and J.M. Metzger, The Determination of Regional Cerebral Blood Flow by Means of Water Labeled with Radioactive Oxygen 15, Radiol. 93:31-40, 1969. 175

12. M.M. Ter-Pogossian, J.O. Eichling, D.O. Davis, and M.J. Welch, The Measure In Vivo of Regional Cerebral Oxygen Utilization by Means of Oxyhemoglobin Labeled with Radioactive Oxygen-15, J. Clin. Invest. 49:381-391, 1970.

13. M.M. Ter-Pogossian, D.O. Davis, J.O. Eichling and C.C. Carter, Regional Cerebral Oxygen Metabolism and Regional Cerebral Blood Flow Compartments Measured by Means of Oxygen-15 in Patients with Cerebral Abnormalities, Cerebral Vascular Diseases, Grune & Stratton, Inc., 1971, pp. 103-108.

14. C.C. Carter, J.O. Eichling, D.O. Davis, and M.M. Ter-Pogossian, Correlation of Regional Cerebral Blood Flow with Regional Oxygen Uptake Using the ^ 0 Method, Neurol. 22:755-762, 1972.

15. M.M. Ter-Pogossian, J.M. Taveras, D.O. Davis and J.O. Eichling, A Study of Regional Cerebral Oxygen Supply and Utilization by Means of Radioactive Oxygen-15, Recent Advances in the Study of Circulation, (J. Taveras, H. Fischgold and D. Dilenge, Eds.), Charles C Thomas, Publishers, Springfield, Illinois, 1969, pp. 156-174.

16. M.E. Raichle, private communication, Division of Radiation Sciences.

17. M.M. Ter-Pogossaan, R.L. Grubb, Jr., M.E. Phelps, M.E. Raichle, and M.J. Welch, The Importance of Cyclotron Produced Carbon-11, Nitrogen-13, and Oxygen-15 in In Vivo Metabolic Studies, presented at the Third International Conference on Medical Physics, July 30- August 4, 1972, Goteborg, Sweden.

18. M.S. Livingston and J.P. Blewett, Particle Accelerators, McGraw- Hill Book Company, New York, 1962, pp. 140, 142.

19. A.I. Brown and S.M. Marco, Introduction to Heat Transfer, Third edition, McGraw-Hill Book Company, New York, 1958, p. 155.

20. J.C. Clark and P.D. Buckingham, The Preparation and Storage of Carbon-11 Labelled Gases for Clinical Use, Int. J. Appl. Radlat. Isotop. 22:639-646, 1971.

21. M.B. Windstead, H.S. Winchell and R.A. Fawwaz, The Synthesis of Sodium H-C-Benzoate and its Use in Renal Visualization, Semiannual Report, 31ology and Medicine, Donner Laboratory UCRL-18793, Fall, 1968, University of California, Berkeley. 176

22. M. Kamen, Radioactive Tracers In Biology. First edition. Academic Press, New York, 1947.

23. B.W. Wieland, Redistribution of Lithium in Polythermal Two-phase Systems, presented at the Fall Meeting of Southwestern Ohio Section of American Nuclear Society, Battelle Memorial Institute, November 1967.

24. O.D. Brill and L.V. Sumin, Excitation Curves for the Reactions B^l(d,2n)cH, B9 (ct,2n)C1;L, BlO(d,n)Cn and C12 (d,n)N13, Sov. J. Atom, Energy Jj 856, 1961.

23. Ref. 3, p. 341.

26. G.U. Din, M.A. Nagarajan and R. Pollard, The Study of ^®B(d,n)^^C and ■*-^B(d,n) C Reactions, Nucl. Physics A93:190-200, 1967.

27. V.K. Wohlleben and E. Schuster, Aktivierungsanalyse mit Deuteron- en, Radiochim. Acta 12:75, 1969.

28. H.W. Newson, Transmutation Functions at High Bombarding Energies, Phys. Rev., April 15, 1937, p. 620.

29. E. Ricci and R.L. Hahn, Theory and Experiment in Rapid, Sensitive Helium-3 Activation Analysis, Analyt. Chem. 37:742, 1965.

30. Ref. 3, p. 258.

31. R.S. Tilbury, Activation Analysis with Charged Particles, Subcom­ mittee on Radiochemistry, National Academy of Sciences, National Research Council, Clearinghouse for Federal Scientific and Technical Information, National Bureau of Standards, U.S. Depart­ ment of Commerce, Springfield, Virginia.

32. Ref. 3, p. 383.

33. J.F. Lifton and M.J. Welch, Preparation of Glucose Labeled with 20-minute Half-lived Carbon-11, Radiat. Res. 45:35, 1971.

34. M.E. Phelps, unpublished data, Division of Radiation Sciences.

35. J.T. Hood, Washington University Physics Department, Cyclotron Unit, private communication.

36. M.J. Welch, private communication, Division of Radiation Sciences. 177

37. M.J. Welch, Division of Radiation Sciences, unpublished data.

38. R.J. Jaszczak, R.L. Macklin and J.H. Gibbons, ^C(d,n)^% Total Cross Section from 1.2 to 4.5 MeV, Phys. Rev. 181:1428, 1969. 13 39. M.J. Welch and M.G. Straatmann, The Reactions of Recoil N Atoms with Some Organic Compounds in the Solid and Liquid Phases, in press, Radiochlm. Acta.

40. B. Hoop, Physics Research Laboratory, Massachusetts General Hospital, Boston, Mass., private communication.

41. M.E. Raichle, M.E. Phelps, K.B. Larson, R.L. Grubb, Jr., M.J. Welch and M.M. Ter-Pogossian, In Vivo Measurement of Cerebral Glucose Metabolism Employing C-Labeled Glucose, presented at the Sixth International CBF Symposium, June 6-9, 1973, Philadelphia, Pa., to be published in the Proceedings of the Meeting.

42. M.E. Raichle, unpublished data, Division of Radiation Sciences.

43. C.F. Williamson, J.P. Boujot and J. Picard, Tables of Range and Stopping Power of Chemical Elements for Charged Particles of Energy 0.05 to 500 MeV, Report CEA-R 3042, July, 1966, Departe- ment de Physique Nucleaire, Service de Physique Nucleaire a Basse Energie.

44. T. Retz-Schmidt and J.L. Weil, Excitation Curves and Angular Dis­ tributions for N^(d,n)ol5f Phys. Rev. 119:1079, 1960.

45. S. Morita, N. Kawai, J. Goto, T. Maki and M. Mukae, The N^Cdjn) 015 Reaction in the Energy Range from 1.5- to 2.9-MeV, J. Phys. Soc. Japan 15:2170, 1960.

46. R.E. Lapp and H.L. Andrews, Nuclear Radiation Physics. Third edition, Prentice-Hall, Inc., Englewood Cliffs, N.J., 1963, pp. 156, 157.

47. Journal of Nuclear Medicine Supplement Number 4, Pamphlet 6 , March 1970, Volume 11; Journal of Nuclear Medicine Supplement Number 2, Pamphlet 4, March 1969, Volume 10.

48. M. Straatmann and M.J. Welch, The Liquid Chromatographic Puri- ficiaton of Carbon-11 Labeled Glucose, Int. J. Appl. Radiat. Isotop. 24:234-236, 1973.

49. W.T. Keeton, Biological Science, W.W. Norton & Company, Inc., New York,, 1967, pp. 116-119. 178

50. Ref. 43, pp. 129-134.

51. 0. Bjorkman, R.W. Pearcy, A.T. Harrison and H. Mooney, Photo­ synthetic Adaptation to High Temperatures: A Field Study in Death Valley, California, Science 175:786, 1972.

52. N.A. Dyson, P. Hugh-Jones, G.R. Newbery and J.B. West, The Prepa­ ration and Use of Oxygen-15 with Particular Reference to its Value in the Study of Pulmonary Malfunction, Proceedings of the Second International Conference, United Nations, Peaceful Uses of Atomic Energy, Geneva, Switzerland, 1958, Pergamon Press, London, 1959.

53. E.J. Williams, Rev. Mod. Phys. 17:217, 1945.

54. R.D. Evans, The Atomic Nucleus, McGraw-Hill Book Company, New York, 1955, p. 652.

55. C. Williamson and J.P. Boujot, Tables of Range and Rate of Energy Loss of Charged Particles of Energy 0.5 to 150 MeV, Report CEA No. 2189, 1962, Service de Physique Nucleaire a Basse Energie.

56. Ref. 53, p. 638.

57. W. Liben, Computation of Nuclide Yield in a 60-inch Cyclotron, Technical Memorandum, The Johns Hopkins University, Applied Physics Laboratory, TG-837, July 1966.