United States
Environmental Protection
Agency
Office of
Radiation Programs
Washington, D.C. 20460
Technical Note
ORP/TAD-79-.i
August 1979
332 EPA
Radiation
A Study of
Radioactive Airborne
Effluents From
Particle Accelerators
-
-------
ERRATUM
This technical note, A Study of Radioactive Airborne Effluents
From Particle Accelerators, was printed with the wcong sequential number.
The number on the cover and title page should read OKP/TAD-79-12. Please
make the appropriate changes with pen and ink.
-------
Technical Note
URP/TAD-79-12
A STUDY OF RADIOACTIVE AIRBORNE EFFLUENTS
FROM PARTICLE ACCELERATORS
Final Contract Report
Principal Investigator: Joel I
Teknekron Research, Inc.
1486 Chain Bridge Road
McLean, Virginia 22101
Cehn
August 1979
Prepared for
U.S. Environmental Protection Agency
Under Contract No. 68-01-4997
Project Officer
Frederick C. Sturz
Office of Radiation Programs (ANR-459J
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
DISCLAIMER
This report has been reviewed by The Office of
Radiation Programs, U.S. Environmental Protection
Agency (EPA) and approved for publication. Approval
does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or
recommendation for use.
11
-------
PREFACE
Secion 122 of the Clean Air Act Amendments of 1977, Public Law 95-95,
directed the Administrator of the Environmental Protection Agency, to review
all relevant information and determine whether emissions of radioactive
pollutants into the ambient air will cause or contribute to air pollution which
may reasonably be anticipated to endanger public health. As part of this
review, the Agency has been assessing the public health impact resulting from
emissions of radioactive material into the air from a broad spectrum of source
categories. This study was performed to assess the extent of radioactive
airborne emissions from non-government owned particle accelerators.
Readers of this report are encouraged to inform the Office of Radiation
Programs of any omissions or errors. Comments or requests for further
information are also invited.
David S. Smith
Director
Technology Assessment Division (ANR-459)
Office of Radiation Programs
iii
-------
ABSTRACT
This report discusses radioactive gas production around non-Federally
owned particle accelerators, and the release of such gases to the atmos-
phere. The estimated 1200 non-Federally owned accelerators in the United
States are categorized by type and energy, and the potential for radiogas
production is discussed for various types. The results of field monitoring
around two machines (a cyclotron and a Van de Graaff) are also presented.
Estimates of annual radiogas releases from generic accelerator facilities
are made. The isotopes of interest are: tritium (target gas) from a
generic Van de Graaff; nitrogen-13, oxygen-15 and carbon-11 from a generic
cyclotron; and argon-41 from a generic linac. Finally, control technologies
to reduce airborne releases of these isotopes are discussed.
IV
-------
TABLE OF CONTENTS
Section
1.0 INTRODUCTION 1-1
2.0 PRINCIPLES OF ACCELERATOR OPERATION 2-1
2.1 General Design Considerations 2-1
2.2 Principal Accelerator Design Features 2-3
2.2.1 Constant Field Machines 2-4
2.2.2 Incremental Acceleration Machines 2-8
2.2.3 Magnetic Field Accelerator (Betatron) . . . 2-17
2.3 Isotope Production 2-21
2.4 Isotope Applications 2-30
3.0 EXTENT OF ACCELERATOR USE IN U.S 3-1
3.1 Growth in Accelerator Use 3-1
3.2 Use by Type of Machine 3-6
3.3 Machine Use by Location 3-10
4.0 POTENTIAL FOR RESIDUAL AIRBORNE ACTIVITY 4-1
4.1 Accelerator Radiation Hazards 4-1
4.2 Mechanisms for Air Activation 4-2
4.3 Calculation of Production Rates 4-2
5.0 EFFLUENT MONITORING DATA 5-1
5.1 Existing Data 5-1
5.2 Data from Study Facilities 5-3
5.2.1 Sampling Techniques 5-5*
5.2.2 Detection Techniques 5-5
5.2.3 Results 5-9
5.3 Release Estimates 5-14
6.0 EFFLUENT TREATMENT SYSTEMS 6-1
6.1 Existing Standards for Air Treatment 6-1
6.2 Treatment Systems 6-2
6.2.1 Ventilation 6-2
6.2.2 Air Cleaning Devices 6-6
6.3 Costs and Effectiveness 6-9
-------
TABLE OF CONTENTS (Continued)
Section Page
7.0 GENERIC FACILITIES 7-1
7-1 Characteristics 7-1
7.2 Release Estimates 7-3
7.2.1 Constant Field Accelerator 7-3
7.2.2 Electron Linac 7-4
7.2.3 Cyclotron 7-4
7.3 Dispersion Estimates 7-5
7.4 Demography 7-6
8.0 CONCLUSIONS 8-1
9.0 REFERENCES 9-1
APPENDIX A
APPENDIX B
vi
-------
LIST OF FIGURES
Figure Page
2-1 Operating Principle of a Voltage Doubler 2-5
2-2 Design of a Van De Graaff Generator 2-7
2-3 Basic Design of a Linear Accelerator 2-9
2-4 Principles of a Cyclotron 2-11
2-5 Cutaway View of a Betatron 2-18
3-1 Growth Trend of Particle Accelerators in the
United States 3-3
3-2 Accelerators Reported by Type 3-8
3-3 Megavoltage Radiation Therapy Equipment Projections
for the United States Through 1980 3-9
3-4 Accelerator Use by Type 3-12
3-5 Accelerator Use by Location 3-13
5-1 Schematic of Cyclotron Facility 5-4
5-2 Schematic of Van De Graaff Facility 5-6
5-3 Sample Train Schematic 5-7
5-4 Decay of Activity on Charcoal-A Sample, Feb. 6. ... 5-10
5-5 Decay of Activity on NaX-A Sample, Feb. 6 5-11
5-6 Decay of Activity of 2-Liter Gas Sample #1, Feb. 6. . 5-13
7-1 Particle Beam Intensity vs. Particle Energy 7-2
7-2 Population Distribution Around a Generic Accelerator
Facility 7-10
vii
-------
LIST OF TABLES
Table
2-1 Radionuclides of Medical Interest 2-26
3-1 Particle Accelerators in the United States 3-2
3-2 Adjusted Estimates of Particle Accelerators by
State - 1977 3-4
3-3 Accelerator Use by Type and Energy 3-11
4-1 Nuclear Reactions Responsible for Most Airborne
Radioactivity around Accelerators 4-3
4-2 List of Air Activation Radionuclides 4-4
4-3 Production Rates of 150, 13N, and 1]C in Air . . . . 4-6
5-1 Duct Sampling Equipment 5-8
6-1 Cost of Air Treatment Controls (1978 Dollars). . . . 6-11
o
7-1 Normalized Concentrations, C/Q (Sec/m ) Downwind
of an Accelerator Facility 7-7
7-2 Normalized Concentration as a Function of Distance
for Short-term Releases 7-8
viii
-------
1-1
1.0 INTRODUCTION
On August 7, 1977, the President signed into law the Clean Air Act Amendments
of 1977. These amendments extended th£ definition of "air pollutant" to Include
radioactive material that is emitted into the ambient air. Also, Congress
directed the Administrator of the Environmental Protection Agency (EPA) to
investigate the emissions of radioactive pollutants with respect to public
health. This study is an investigation of the effluents from selected
particle accelerator facilities.
The sources of these potential radioactive effluents are discussed in Sections
4.0 and 5.0. Section 5.0 also contains the results of effluent monitoring
performed at two study facilities. A discussion of air treatment systems capable
of reducing these effluents is contained in Section 6.0.
Also presented in this report is an overview of the use and distribution of
accelerators in the United States (Section 3.0) and a discussion of the various
types of accelerators and how they work (Section 2.0). Finally, a summary of
our findings and our conclusions is presented in Sections 7.0 and 8.0.
It should be noted at this point that this study is restricted to non-Federally
owned accelerators. Airborne effluents from the large and very large machines
at the national laboratories (e.g., Brookhaven) and elsewhere (e.g., National
Bureau of Standards) are the subject of a concurrent study. Thus, we have
focused on machines at universities and in the private sector, which generally
range in energy up to about 100 MeV.
Finally, this study focuses on machines not used exclusively for the
production of radiopharmaceuticals. Discussions of airborne effluents from
these sources are included in a separate report, "A Study of Airborne Radio-
active Effluents from the Radiopharmaceutical Industry" recently completed
by Teknekron for EPA/EERF.
-------
2-1
2.0 PRINCIPLES OF ACCELERATOR OPERATION
2.1 General Design Considerations
Generically, accelerators are devices for imparting high kinetic energies
to electrons or to positively charged particles (such as alpha particles,
protons, and deuterons) by electric or magnetic fields. In typical operation,
the stream of accelerated particles, which travel in an evacuated tube or other
enclosure within the machine, impinges on a metallic or gaseous target, pro-
ducing various types of secondary radiation. Those types with the highest
associated energies, neutrons and X-rays, are of principal importance in terms
of practical application, such as industrial radiography, X-ray therapy, and
radioisotope production. In some designs the target is enclosed within the
accelerating tube, while in others it is external to the vacuum. In the
latter case, the particle beam leaves the tube via a foil window so that it
travels through air for an appreciable distance before reaching the target.
In certain applications, such as electron beam tumor therapy, the primary
particle beam is used directly, rather than as a means for generating second-
ary radiation.
lii terns of basic design, accelerators are often categorized according to the
means used to achieve the desired particle velocity. For the most part, they
fall into the following three main classes:
t Constant D.C. Field Machines
Accelerators in this class, sometimes called "potential-drop"
machines, share the common feature of a constant D.C. electric
field through which the charged particles "fall." Depending
on the polarity of the accelerating field with respect to the
particle source, machines of this type can be used for accelerating
either negatively or positively charged particles. Within7 this
class, the most important accelerator designs are the Cockcroft-
Walton and Van de Graaff machines which are named after the
developers of the original prototypes. The Dynamitrcn and the
insulated core transformer (ICT) accolerntor arp constant D.C.
field machines based on the Cockroft-Walton design.
-------
2-2
• Incremental Acceleration Machines
In these machines, the particles are incrementally accelerated
by time varying electric fields, so that their velocity increases
in a stepwise rather than continuous manner. The most common
types in general use are the linear accelerator (linac) and
the cyclotron and its derivative designs.
• Magnetic Field Accelerator
The only current example of this category is the betatron in which
electrons are accelerated by a time varying electric field which,
in turn, is derived from a time varying magnetic field. Accelera-
tion is not incremental.
Some machines do not fit exactly into any of the above three categories. For
example, neutron generators are so classified in terms of function, rather
than by the accelerating principle employed. In these devices, neutrons are
typically produced by the impingement of accelerated deuterons on a tritium
target. Some designs employ Cockcroft-Walton type high voltage power supplies
for generating the accelerating field, so that these are constant D.C. field
machines. Others, however, are self-rectifying, so that the electric field
actually consists of a series of half-wave pulses having a repetition rate
equal to the frequency of the source A.C. voltage. Such a field is equivalent
to a D.C. field with a superimposed "ripple" component. It is not a constant
D.C. field. Another accelerator type that is not categorizable under the above
classification is the resonant transformer machine. In this design, the high
voltage is produced by resonating the transformer inductance with the distributed
capacity of the system. The frequency of the current supplied to the primary
winding is usually higher than ordinary line frequency (60 Hz) because the
system capacity is relatively low. Resonant transformer machines typically
operate in the self-rectifying mode, so that the accelerating field generated is
not one of constant D.C. potential.
Before a discussion of the machines themselves, it is important to consider
the topic of relativistic mass increase because this factor strongly influences
the design of several accelerator types. As predicted by Einstein, the mass
of a moving body has been found not to be independent of its velocity when
measured with respect to a given reference framework. The quantitative
relationship between the mass of a body moving at a velocity (v) and the mass
-------
2-3
of the body at rest is shown by the following expression in which b = v/c
(c is the velocity of light = 300,000 kilometers/second), m is the moving
mass, and mQ the rest mass:
m/m =•
1 - b'
As is evident, when the velocity is very small compared with that of light, the
term b is essentially zero so that m is virtually equal to m . However, at
velocities that are comparable with that of light, b is no longer negligible,
with the result that the m/m ratio can become quite large. For example, for
b = 0.9, m/mQ is about 2.3; for b = 0.999999999, m/mQ is approximately 22,360.
In the limiting case, m approaches infinity as v approaches c.*
At relatively low energy levels, the relativistic increase in mass is small
for accelerated positive particles but may be quite appreciable for electrons.
The kinetic energy of a particle (K), in terms of its mass and velocity, is
2
given by K = mv /2. Inasmuch as the rest mass of an electron is only about
1/1830 that of a proton, for any value of K an electron must be accelerated
to a considerably higher velocity than a proton. For example, at K = 11 MeV**
an electron must have been accelerated to a velocity of about 0.999 that of
light. At this velocity, its m/m ratio would approximate 22. On the other
hand, a proton accelerated to the same energy level would be moving at a
velocity of only a little more than 0.1 light speed and its m/m ratio would
be less than 1.005.
2.2 Principal Accelerator Design Feature^
This section describes the major design features of the different accelerator
classes identified in Section 2.1.
*Note that if a body were to be accelerated to a velocity approaching c, the
required force (force = mass x acceleration) would approach infinity as the
mass approached infinity. As the expression shows, if the force failed to
approach infinity, the acceleration produced by it would progressively fall
as the mass increased, reaching zero in the limiting case. The speed of light
is, therefore, an unattainable velocity limit for anything possessing mass.
**This expression means 11 million electron volts. The incremental kinetic
energy gained by an electron as a result of its acceleration by a one-volt
electric field is defined as one electron volt. The electron volt is a
common unit of measure in the context of particle acceleration.
-------
2-4
2.2.1 Constant Field Machines
Cockcroft-Walton Accelerator
This design was employed in the first accelerator constructed (1932) and is
still employed for particle acceleration to energy levels up to about 1 MeV.
The high voltage D.C. accelerating field is produced by rectification of the
output of a high potential power transformer in a voltage doubling circuit.
The net result is the development of a D.C. potential that is theoretically*
twice the peak of the A.C. voltage appearing across the transformer secondary
winding. The principle of operation is shown in Figure 2-1. When the high
voltage appearing at terminal A of the secondary winding passes through the
positive half of the A.C. cycle, rectifier R, conducts, thus charging capacitor
C, to a positive D.C. potential equal to the peak voltage of the half sine
wave, as shown in the figure. During the second half of the cycle, the sign
of the potential at A is negative and R, does not conduct. However, R? now
conducts so that the capacitor C~ charges to the peak negative value of the
applied voltage. Since C, and C2 are connected in series, the total voltage
appearing across the two is twice the peak voltage.
The effectiveness of this system in maintaining a constant D.C. accelerating
voltage largely depends on the relationship between the magnitudes of the
charging and load (due to the particle beam) currents and on the capacitor
sizes. If the load current is small in comparison with the charging current,
the potential across the capacitors will decline only slightly between
successive peaks in the A.C. cycle. As the load current increases, the D.C.
potential will drop progressively between these peaks so that its average
value could fall significantly below twice the peak value. This effect is
called "regulation." It is significant because a change in the field potential
will directly affect the kinetic energies of the accelerated particles
*The D.C. potential would be exactly twice the peak voltage if the electrical
components were ideal and if there were no current drain from the circuit,
either from extraneous losses or the particle beam.
-------
2-5
A
- R2
HIGH VOLTAGE
D.C. OUTPUT
S~^f
/ T PEAK VOLTAGE
FIGURE 2-1
OPERATING PRINCIPLE OF A VOLTAGE DOUBLER
-------
2-6
which must usually be accurately maintained at known values in experimental
studies. Typically, these accelerators are equipped with voltage control
devices that permit precise setting of the D.C. field potential under the
load imposed by the beam current which may range from microamperes to
several milliamperes.
Van de Graaff Generator^
This machine employs a moving belt of insulating material which operates as
an electric charge transfer device. High potentials are generated by charge
accumulation. As Figure 2-2 shows, the charge is produced by a high voltage
power supply and transferred to the belt via electrode A. As the belt moves
in the direction shown by the arrows, the charge is transferred by electrostatic
induction to the metal shell C via electrode B. This device is based on
the principle that the potential (V) at a point, produced by a charge (q),
at a given distance from the point (r), in a medium of dielectric constant (n)
is given by the relationship:
V= Tip-
Inasmuch as n is a constant (unity in a vacuum, and very close to unity in
most gases) and r is determined by the machine design, the voltage developed
depends solely on how much charge can be transferred to C and maintained. A
major limiting factor is the breakdown potential of the air surrounding the
charged sphere. This potential can be substantially increased by operating
the machine in either a vacuum or a high-pressure ambient atmosphere. The
latter procedure is generally employed. It is usual practice to house the
machine within an enclosure that is pressurized (with either air or some other
gas) to about 10 atmospheres. Under these conditions, potentials of the order
of 10-12 megavolts can be maintained. Small machines designed to generate
relatively low potential accelerating fields (less than one megavolt) do
not require pressurization.
-------
2-7
ION OR ELECTRON SOURCE
HIGH VOLTAGE TERMINAL
CORONA POINT
VOLTAGE CONTROL
INSULATING COLUMN &
EQUIP OTENTIAL RINGS
PRESSURE TANK
ACCELERATING COLUMN
HIGH VOLTAGE
POH€R SUPPLY
z
PRESSURIZING
GAS SUPPLY
BELT DRIVE
VACUUM PUMP
t
BEAM
FIGURE 2-2
DESIGN OF A VAN DE GRAAFF GENERATOR
-------
2-8
Higher particle energies can be attained through "tandem" arrangements of
these machines. For example, in the two-stage tandem Van de Graaff generator,
negative ions are accelerated in the first stage toward the positive ter-
minal located at the center of the machine. This terminal houses a gas-filled
channel through which the ions travel. During this exposure, they lose their
negative charges and are converted to a positive ion beam. In the second
stage, this beam is further accelerated toward the ground potential terminal
at its end. Three-stage Van de Graaff accelerators have also been constructed.
All constant field machines have the disadvantage of requiring high voltage
D.C. sources with the attendant problems of insulator leakage and corona
loss due to ionization of the surrounding air. These problems become more
severe with increasing voltage and limit the maximum potentials that are
practicably attainable and, hence, limit the maximum kinetic energies that
can be imparted to the accelerated particles. In the next class of machines
to be discussed, these difficulties are eliminated by successively applying
relatively weak accelerating fields whose cumulative effects permit the
attainment of far higher energy levels than can be achieved with constant
field machines.
2.2.2 Incremental Acceleration Machines
Linear Accelerator (Linac)
The basic design of this machine is illustrated in Figure 2-3. In essence,
this accelerator consists of a series of coaxially arranged metal tubes, with
gaps between them, alternately connected to the same terminal of a high fre-
quency sine wave generator. The tube assembly is enclosed within a vacuum.
Electrons are injected at point A and travel in the direction shown by the
arrow. In the drawing, an electron is shown as about to enter the gap between
tube 1 and tube 2. Assume that the time of entry coincides with the time at
which the phase of the applied voltage is such that tube 2 is positive with
respect to tube 1. The electron will thus accelerate during its passage
-------
2-9
ELECTRON
GUN
A
TARGET
MICROWAVE OSCILLATORS
FIGURE 2-3
BASIC DESIGN OF A LINEAR ACCELERATOR
-------
2-10
across the gap. By the time it has reached the beginning of the next gap
(between tubes 2 and 3), the driving voltage has passed through a complete
cycle so that tube 3 is electrically positive with respect to tube 2,
thus further accelerating the electron. (Since there are no potential
gradients within the tubes, the electrons experience acceleration only when
moving through the gaps between them.) It will be noted from Figure 2-3
that the tube lengths increase progressively in the direction of electron
beam travel. This increase in length compensates for the decrease in the
time required for an electron to travel a given distance along its total
path as its velocity increases. If the tube lengths are appropriately
selected with respect to the frequency of the driving potential and the
increasing particle velocities, with consequent relativistic mass effects
taken into account, the electrons will arrive at each gap at a time when
the tube it is approaching is positive with respect to the tube it is leaving.
The above discussion presents the linac as an electron accelerator because all
the early machines were designed for this purpose. At present, however, proton
linear accelerators are also made. Their basic principle of operation is the
same, except that proton entry into an inter-tube gap occurs when the tube
ahead is negative with respect to the one behind. A major application of
electron linacs is radio- and electron-therapy. These machines are often
designed so that either X-rays (resulting from the impingement of the electron
beam on a target) or the primary electron beam itself can be employed. Proton
linacs are used primarily for research purposes.
Cyclotron
Like the linac, the cyclotron uses an alternating electric field for incre-
mentally accelerating the particle beam. However, as the name of the machine
implies, the beam path is circular, rather than straight. Figure 2-4, an
illustration of the features of the cyclotron, shows that the key components
of the cyclotron include two semicircular hollow shells (called "dees" because
of their shape). The dees, which are made of metal, are enclosed within a
-------
2-11
Deflector
Plate
Outlet
Target
Chamber
Metal Tank
FIGURE 2-4
PRINCIPLES OF A CYCLOTRON
-------
2-12
metal tank which is evacuated and connected to an alternating current
generator that supplies the accelerating voltage. Not shown in the
figure is a magnet structure whose pole pieces are parallel with the
upper and lower dee surfaces and which exerts a centrally directed uniform
radial magnetic field. The gap between the two dees is functionally
analogous to those between the tubes in the linac. The particles are
accelerated only when moving across the gap because there is no potential
gradient within the dees. The basic operating principle is as follows:
a positive ion introduced from the source S at a time when dee B is
negative with respect to dee A would be accelerated toward dee B and,
in the absence of the imposed magnetic field, would move within it in a
straight line at constant velocity. Because of the magnetic field, however,
the ion will be constrained to a path of constant radius while it remains
within dee B. The radius is constant because while the particle is within a
dee, its velocity is not exposed to tangential accelerating forces and remains
unchanged. During this time, the centripetal and centrifugal forces on it
2
are equal. The centrifugal force is expressed by Mv /r, where M is the
particle mass and r the radius of its path. The centripetal force is given
by Hev, where H is the magnetic field intensity and e is the charge associated
with the particle. Thus,
M 2
Hev = v ^ S
Therefore, for a constant velocity, r is constant because M, V, H, and e are
also constant. (It is assumed for the moment that relativistic effects can ^
be ignored, permitting M to be treated as constant. These effects will be
considered later.) Hence, the path described by a particle traveling within
a dee is a semicircle (ignoring the half-width of the gap). When the particle
leaves dee B and enters the gap in its approach to dee A, it is exposed to the
accelerating field between the dees, and thus its velocity increases. Since
r = Mv/He, r must increase. Thus, on entering dee A, the ion describes a
semicircular path of slightly greater radius than that followed previously in
dee B. From the foregoing, it is clear that most of the total ion path within
-------
2-13
the cyclotron consists of a series of semicircles of progressively larger
raclii. When the ion reaches the periphery of dee A, it is deflected out-
ward by a negatively charged electrode into the target chamber.
In order to accelerate the particles with a time varying electric field of
constant frequency, it is necessary that they always arrive at a gap when
the potential gradient across the dees is of the correct polarity regardless
of their path radii at the arrival time. Expressed somewhat differently,
their angular velocity (w) must be constant and, therefore, independent of
their linear velocity (v). That this condition is met (assuming particle
mass to be constant) is easily demonstrated. The angular velocity (w) is
equal to v/r. Thus, r = v/w. If v/w is now substituted for r in the equation
r = Mv/He, we find that
w = He/M
For any given positive particle (proton, alpha particle, deuteron, etc.), the
e/M ratio is characteristic of the particle type. Therefore, the angular
velocity and thus the traverse time per semicircle is determined only by the
magnetic field intensity, once the category of particle to be accelerated has
been chosen. Note that v does not appear in the equation, so that w is
independent of v and, hence, of r. In practice, it is usual to maintain the
field at a fixed intensity (regardless of particle type) and to adjust the
frequency of the accelerating field so that it will be in synchronism with w,
as determined by both the e/M ratio of the particle type to be accelerated
and -the magnetic field intensity setting. At typical magnetic field inten-
sities, the accelerating electric field frequencies are of the order of several
megacycles per second (MHz). For example, if a machine's magnetic field
intensity has been set to 15,000 gauss, the accelerating frequency for protons
2+
would be approximately 23 MHz. In the case of deuterons (He ions), the
e/M ratio is one-half that of protons, so that the accelerating frequency for
these ions (assuming the same value of H) would be 1/2 x 23 MHz.
An important characteristic of a cyclotron is the maximum final energy attain-
able for a given particle accelerated within it which, as will be shown, is
-------
2-14
directly related to the machine dimensions. This energy can be readily
derived by squaring the expression for the path radius shown above, giving
,2.2
Kinetic energy = K = 1/2 Mv
As the last expression indicates, the kinetic energy of a particle of a given
type constrained to a circular path by a given value of H is directly propor-
tional to the square of the radius of that path. If r is taken to be the
radius of the final semicircle traversed by the particle before ejection from
the dee, the final attainable energy is thus essentially determined by the
radii of the dees. This is why high energy cyclotrons must be larger than
lower energy machines.
The final energy attained by a particle is independent of the incremental
kinetic energy acquired during each of its passages across the gaps between
the dees. Nevertheless, it is obvious that the final energy must be the
aggregate of these increments. It follows that the total number of revolutions
made by a particle within the cyclotron to attain a given energy level will be
fewer for stronger than for weaker electric accelerating fields. Ordinarily,
the number of complete orbits made by a particle is on the order of hundreds
of thousands.
In the discussions of the cyclotron to this point, any effects due to rela-
tivistic increases of particle mass have been deliberately ignored (i.e.,
the mass was assumed to remain constant) so that the angular velocity (w)
could then be considered as constant for a given magnetic field intensity
and particle type. Because of the constancy of w, the applied accelerating
A.C. voltage could be maintained at an invariant frequency and the arrival
of a particle at a gap would always occur when the potential across the gap
was appropriately phased. Cyclotrons of this type are sometimes called "fixed
frequency machines." The assumption of constant particle mass is essentially
-------
2-15
valid for positive particles, provided that they are accelerated in relatively
low energy cyclotrons. For example, if a proton were accelerated to an energy
level of about 4.7 MeV, the ratio of moving mass to rest mass (m/m ) would be
only 1.005 (the corresponding velocity would be approximately one-tenth that
of light). The constant mass assumption does not, however, apply to electrons.
As previously explained, because of their relatively low rest masses (approximately
1/1830 of a proton mass) they must be accelerated to quite high velocities to
attain even moderate energy levels. At 4.7 MeV, an electron would have an m/m
ratio of over 10 and a velocity of more than 0.995 that of light. Because of
the large increase of electron mass at even comparatively low energy levels,
it is evident that their angular velocities would not remain constant if
accelerated in a cyclotron, but would decline substantially so that phase
synchronization with a fixed frequency accelerating field would soon be lost.
For this reason, cyclotrons are not suitable as electron accelerators.
Relativistic increases in proton mass cannot be ignored when these particles
are accelerated to higher energy levels. However, these increases are far
less extreme than the increases in electron mass. For example, the proton
m/m ratio at 625 MeV is somewhat less than 1.7; at 1200 MeV it is about 2.29.
The magnitudes of these ratios are low enough so that they can be accommodated
by appropriate modifications of the basic cyclotron design without loss of phase
synchronization.
In some machines, the relativistic mass effect is compensated for by a nonum'form
magnetic field which is stronger at the periphery of the dee than at its center.
It will be recalled that the angular velocity (w) of the accelerated particles
is equal to He/M. From this relationship, it can be seen that if the magnetic
field is shaped so that its intensity increases radially at the rate of the
relativistic increase in proton mass, the ratio H/M will be constant and, con-
sequently, w will also remain constant. Machines designed in this manner can
accelerate protons to higher energy levels with fixed frequency A.C. generators
than can those machines with radially uniform magnetic fields. They suffer,
however, from the fact that radial field shaping tends to defocus the particle
-------
2-16
beam. This defocusing effect can be largely overcome by special contouring
of the faces of the magnet pole pieces. Accelerators incorporating this
feature are sometimes called "sector focused" or "azimuthally varying field"
(AVF) cyclotrons. These machines permit the attainment of energy levels of
the order of hundreds of MeV.
Frequency modulation is a quite different approach to compensating for rela-
tivistic mass increase. As pointed out, the effect of this mass increase,
given a uniformly radial magnetic field, is to reduce the angular velocity
of the particles. This limits the final energy that can be attained in a
fixed frequency machine because of progressive loss of synchronization between
the particle arrivals at the gaps between the dees and the accelerating A.C.
voltage. It is evident that higher energy levels could be attained if the
accelerating voltage frequency were gradually reduced as the particles spiraled
outward so that it always stayed "in step" with the decreasing angular velocity
of the particles. This technique has been implemented, and machines based on
this principle are called "synchrocyclotrons."
It was originally believed that achieving synchronization between the
accelerating field frequency (f) and the decreasing angular velocity (w) of the
particles might prove difficult. However, the "principle of phase stability,"
discovered in 1944-45, indicates that sychronism will occur more or less auto-
matically. Suppose that the machine is designed so that the particles enter
the gap during the 90°-180° portion of the sinusoidal accelerating waveform.
Now assume that there is a slight mismatch between f and w such that a particle,
in crossing a gap, receives slightly more energy than it should. In this case,
its relativistic mass increase will be slightly higher than it would normally
be, and its semicircular path through the dee will require correspondingly more
time. Therefore, it will be late in its arrival at the next gap with respect
to the accelerating voltage wave which will have passed its 90° peak. In con-
sequence, the particle now will receive somewhat less energy than it would have
if synchronization had been precise. Conversely, a particle that originally
received too little energy during a gap crossing will experience a smaller
relativistic mass increase and it's angular velocity will be correspondingly
-------
2-17
higher. Therefore, it will arrive at the next gap somewhat earlier during
the 90°-180° portion of the accelerating voltage wave and will thus
receive more energy during its transit through the gap. This self-correcting
mechanism does not attain phase stability in the absolute sense of a rigid
relationship in time between particle arrival at a gap and the appearance
of a particular point of the accelerating voltage waveform. Actually, the
orbit periods tend to oscillate about a mean value from which they vary by
quite small amounts. These oscillations are known as "synchrotron oscillations,"
and their frequencies are usually considerably less than the revolution
frequency (1/w).
As is evident from the foregoing, particle acceleration occurs during only
a fraction of the total sinewave cycle. In a synchrocyclotron, stable orbits
are formed by particles reaching the first transit during approximately 1%
of the cycle, so that the beam is actually discontinuous. It consists of a
series of bursts of pulses. Because of this low duty cycle (i.e., ratio of
beam pulse duration to the total period of one sinusoidal cycle), the average
beam currents generated in these machines tend to be lower than those produced
in constant frequency machines in which the effective duty cycles are longer.
2.2.3 Magnetic Field Accelerator (Betatron)
First developed in 1940, the betatron was initially conceptualized as an
approach to imparting higher energies to electrons than had been attained by
other means. Betatrons can produce electron beams with energies on the
order of several tens of MeV (one machine in operation at the University of
Illinois is rated at about 300 MeV). In the past, the therapeutic use of
betatrons in hospitals was not uncommon, but they have been displaced by
linacs for this application to a considerable extent.
The principal betatron design features are shown in Figure 2-5. Basically,
this machine consists of an alternating current magnet within whose field is
located a circular evacuated tube ("doughnut") within which the electrons
-------
2-18
INDUCED ACCELERATING
FIELD
4-
MAGNETIC FIELD
\\N\\\\\\\\\\\\\\\ |
•ft
DOUGHNUT
VACUUM TUBE
STABLE ELECTRON
ORBIT
FIGURE 2-5
CUTAWAY VIEW OF A BETATRON
-------
2-19
are accelerated. The electrons are usually injected into the doughnut
in pulses from an electron gun (i.e., an electron source and associated
D.C. accelerating electrodes) with initial energies of a few tens of keV.
Between the pole pieces of the magnet, identified as P in the illustrations,
is a "flux bar" (F). By inserting flux bars with different magnetic satura-
tion characteristics, the maximum energy attained by the electrons can be
controlled.
The betatron operates according to the following principles. The sinusoidally
alternating magnetic flux, whose frequency is typically a few hundred Hz
(cycles per second), induces an A.C. emf (electromotive force) of the same
frequency that is exerted tangentially with respect to the electron orbit
within the tube. The magnitude of the accelerating emf is maximum when the
time rate of change of the magnetic flux is maximum. The latter condition
occurs when the sinusoidal flux is passing through its zero value. The
pulse repetition rate of the injected electron burst is synchronized with
the frequency of the magnetic field and is so phased that the electrons enter
the doughnut about when the field is changing sign (i.e., passing through zero).
Thus, the rate of particle acceleration is highest at the time of injection.
The magnetic field that generates the accelerating emf also exerts a centripetal
force that constrains the electrons to a circular path. As the electron velocity
(v) increases, the path radius might also be expected to increase, as in a
cyclotron. However, the magnetic field intensity and hence the centripetal force
is increasing at the same time. By shaping the pole pieces for appropriate adjust-
ment of the field intensity as a function of radial distance, the electron orbits
can be well stabilized.
In theory, the phase interval of the magnetic field waveform during which
electron acceleration can be maintained could continue from the zero crossing
to the point at which the sine wave reaches its peak value. This interval
is 90°. At the peak, although the magnetic field intensity is maximum,
its derivative (i.e., time rate of change) is zero, and, consequently, the
-------
2-20
accelerating emf is also zero. If the electrons were permitted to remain
within the doughnut after the peak, an emf would appear again but it would
be opposite in sign to the accelerating emf and would hence decelerate
the particles. Therefore, the phase interval during which acceleration
takes place never exceeds 90° and may, in fact, be substantially less,
depending on the desired energy level.
The deflection of the electrons from their orbital path to the target may be
controlled in either of two ways. As the magnetic field intensity increases,
the flux bar saturates, with the result that the radius of the electron path
diminishes. The particles then impinge on a target positioned inside of the
orbit they traversed while accelerating. Through the use of flux bars with
different saturation characteristics, the energy attained by the electrons
prior to leaving the acceleration orbit can be predetermined. Alternatively,
the electrons can be magnetically deflected from this orbit at any time during
the acceleration period by a suddenly applied field. This field is typically
generated by the discharge of a condenser into an appropriately positioned
coil. This second method obviously provides more options in the placement of
the target with respect to the orbit.
From the above description, it is clear that the betatron is a "one shot"
machine in the sense that the electrons receive only one accelerating pulse
from their initial injection to deflection from the orbit. However, the dura-
tion of this pulse is considerable when compared with the time required for an
electron to complete a single orbit, so that the particles are continuously
accelerated during many orbital revolutions, although at a decreasing rate.
For example, a 300 Hz magnetic field has a period (i.e., time required for one
complete sinusoidal alternation) of about 3.3 milliseconds. Assuming that the
electrons are accelerated during a 90° interval, the actual acceleration
time would be 25% (90°/360°) of this, or approximately 0.83 milliseconds.
During this brief period, the electrons traverse several tens of thousands
of orbits, attaining terminal velocities approaching the speed of light.
(An electron accelerated to a moderate kinetic energy level of about
-------
2-21
11 MeV would have a velocity greater than 0.999 that of light.) It should
be recalled, however, that the electrons as initially injected into the
doughnut have already attained considerable velocity. For example, an
electron injected with a kinetic energy of 100 keVwill be moving
at nearly 0.6 light speed before it is further accelerated by the betatron.
2.3 Isotope Production
Radionuclide Production - General Principles
Artificial (i.e., man-made) radionuclides are produced in four different ways:
a. Fission
The fissioning of uranium and plutonium in nuclear reactors results
in a number of fission-product isotopes, both stable and radioactive.
Among the latter are 144ce, 90$r and- 137cs.
b. Neutron Irradiation
An appropriate target is bombarded by neutrons, typically in a reactor.
Typical products include 32p (from 32$ ), 14C (from I^N) and 47Ca (from
c. Accelerators
An appropriate target is bombarded by positively charged particles.
Cyclotrons are generally employed. Typical products include 129cs (from
"2/1), 62zn (from 60Ni) and 52Fe (from 50Cr).
d. Radioisotope Generators
An artificial radioisotope generator (cow) spontaneously decays to a desired
daughter radionuclide. If the half-life of the cow is sufficiently
long, the daughter nuclide can be "milked" as required. In terms of
medical application, a generator of particular interest provides
99mjc from 99Mo. (The "m" means metastable and indicates in this
case that 99mjc has a relatively brief half-life, actually about 6
hours, decaying to 99jc).
In general, nuclides that are transmuted to radionuclides by bombardment are
converted to an unstable state of either (a) proton excess with respect to the
number of neutrons in the bombarded nucleus, or (b) neutron excess with respect
-------
2-22
to the number of protons. As the bombarding particle is incorporated within
the nucleus, other particles or gamma radiation is emitted. In the case of
neutron bombardment, a common nuclear reaction is one in which the neutron
is "captured" by the nucleus with the concurrent emission of a gamma ray.
As a consequence, the atomic mass number of the target nucleus is increased
by one, but its chemical identity remains the same.
A simple reaction illustrating this process is:
27A1 + neutron becomes Al and a gamma ray is emitted
Note that the Al has not been transmuted to some other element; its atomic
number remains the same (13) because it has neither gained nor lost nuclear
charge. However, its atomic mass has increased from 27 to 28 because of the
captured neutron. 28A1 is not a stable isotope. The nucleus emits an electron,
becoming 28Si. 28Si is stable and has an atomic number of 14 in consequence of
the unit increase in nuclear charge. Not all neutron-induced nuclear reactions
necessarily result in the formation of radioactive isotopes. For example, if
B is bombarded by neutrons, Li is formed and alpha particles are emitted.
This is the "neutron-alpha reaction." Li is stable.
Radionuclides produced by accelerators are always formed by positively charged
particle bombardment. The "projectiles" used may be protons, deuterons or
alpha particles. Depending on the conditions (i.e., identity of target, nature
of particle, etc.), any of a number of nuclear reactions may occur. The follow-
ing is a list of some of the more important of these reactions in terms of the
incident particle and the emitted particle or radiation.
Nuclear Reactions Caused by Positive Particle Bombardment
Incident Particle Emissions
Proton Deuteron
Alpha particle
Neutron
Gamma ray
-------
2-23
Deuteron Proton
Alpha particle
Neutron
Alpha particle Proton
Neutron
A standardized form of representation has been developed for symbolizing
nuclear reactions such as those categorized above. As an illustration, if
12 1?
C is bombarded with deuterons it is transmuted to N and a neutron is
emitted for each deuteron absorbed. This would be written as
12C(d,n)13N
Note that the bombarded nuclide is shown first with its mass number. Within the
parentheses are shown the bombarding and emitted entities in that order, separated
by a comma. The nuclide resulting from the reaction is shown last, also with its
mass number identified. It is usual to omit the atomic numbers (not shown)
from these nuclear reaction equations because these numbers are inherent in the
identity of the nuclides (i.e., all carbon atoms, for example, regardless of
their mass numbers, have the same atomic number - 6).
In the reaction shown above, it can be seen that both the atomic mass and the
atomic number have increased by one (that is, the atomic mass change is from
12 to 13 and the change in atomic number is from 6 to 7). This is to be expected
because the deuteron consists of one proton and one neutron. Since a neutron
is emitted, the net effect is that of the addition of a proton to the bombarded
carbon nucleus with a consequent gain of one unit of atomic mass as well as of
13
charge. The resulting N is unstable. It emits positrons and has a half-life
of only 10 minutes. As another example,
55Mn(p,4n)52Fe
Since, in this reaction, for each proton absorbed there are four neutrons emitted,
there is a net loss of three mass units per nucleus, but a gain of one unit of charge.
An illustration of alpha particle bombardment is seen in the production of
28Mg from Al. For each interaction of a Al nucleus with an alpha particle,
three protons are emitted. This is represented as follows:
27Al(a,3p)28Mg
-------
2-24
Remembering that an alpha particle consists of two protons and two neutrons,
it follows that the atomic mass has increased by one. On the other hand,
the nuclear charge has decreased by one (two protons gained, three lost), so
that the atomic number has dropped from 13 (Al) to 12 (Mg).
A final example is presented here to demonstrate the possibility of a change
in atomic mass without a concurrent change in atomic number. If 84Kr is
bombarded with deuterons, one proton is lost for each deuteron absorbed as
follows:
84Kr(d,p)85mKr
This is the equivalent of adding one neutron to the nucleus, so that there
is a gain of one unit of atomic mass with no change in atomic number. The
letter "m" following the mass number of 85 means that the product nucleus
is a metastable form of Kr, with a short half-life. mKr is a beta and
gamma ray emitter with a half-life of 4.4 hours. For comparison, non-
metastable 85Kr has a half-life of over 10 years.
Radionuclide Production - Methods
The cyclotron is by far the most commonly used accelerator in radionuclide
production, although Van de Graaff generators and Cockcroft-Walton machines
are occasionally used. In the discussion of linacs presented earlier in this
report, it was mentioned that while these devices were originally developed
for electron acceleration, they have recently also been used for positive
particle acceleration.
The kinetic energies to which the bombarding particles are raised are usually
moderate in terms of the capabilities of very large machines. They are
typically in the range of 10-30 MeV, with the exact values determined by the
identities of the accelerated particles, the target material, and the desired
nuclear reaction.
Although the basic principles of radionuclide production by accelerators
appear simple, their actual implementation may pose many practical difficulties.
For example, the bombardment of the target produces considerable heat which can
-------
2-25
rapidly deform, fuse, or even vaporize the target. For this reason, water
cooling is generally employed. Furthermore, the target material may be a metal
foil, powder, liquid, or even a gas; each of these forms imposes its own
particular requirements. To a considerable extent, success in the production
of many radionuclides has been due to the ingenuity exercised in the mechanical
design of the target structure.
The radionuclide yields practicably attainable with accelerators are lower than
those that can be achieved by neutron bombardment, and, therefore, accelerator-
produced isotopes tend to be more expensive than those produced in reactors.
Accelerator production does, however, have certain general advantages. For
example, if the desired radionuclide has a short half-life, such as a few
hours, it must be produced at or near the site of use. Further, in some cases
it may not be practical to obtain a specific desired radionuclide except by
positive particle bombardment.
After the target has been irradiated, the extraction of the produced radio-
nuclide may be simple or complex. In some cases, a salt may be dissolved
directly in water. Other procedures, such as precipitation, elution, ion
exchange, or chemical separation, may be necessary. The methods used are
essentially the same as those used in the processing of radionuclides produced
through neutron bombardment.
Radionuclide Production - Specific Examples
As Table 2-1 suggests, the number of medically useful radionuclides is con-
siderable. Actually, the examples shown are by no means comprehensive.
11 1 "3 1*5 4^ 1 ?ft
Other isotopes of medical interest include C, N, 0, K, Ba, and
77Br, to name a few. The variety of these radionuclides precludes discussion
of each, but the technical considerations associated with specific isotopes
of particular importance are summarized below.
-------
USE
NUCLJDE
TABLE 2-1. Radionuclides of Medical Interest
FORM TREATMENT/STUDY USUAL DOSE STATE
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vitro
In vitro
In vitro
In vitro
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
In vivo
Internal therapy
Internal therapy
Internal therapy
Teohnotlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technotlum 99m
Technetlum 99m
Technetlum 99m
Technetlum 99m
Technotlum 99m
Technetlum 99m
Technetiura 99m
Iodine 125, 131
Iodine 125, 131
Iodine 125, 131
Iodine 125, 131
Iodine 123
Iodine 123
Iodine 125,131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Iodine 125,131
Iodine 131
Iodine 125,131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Iodine 131
Medronate
Pertechnetate
Pertechnetate
Portechnetate
Pertechnetate
Perteohnotate
Stannous polyphosphate
Sulfur colloid
DTPA (Iron-ascorbate)
DTPA (tin)
DTPA (tin)
DTPA (tin)
HSA mlcrospheres
Dl sodium etidronate (tin)
Labelled macroalbumin (tin)
Iodide
Labelled HSA
Labelled renal function compounds
Labelled fats, fatty acids
Iodide
Iodide
Iodide
Iodide
Labelled USA
Labelled HSA
Labelled HSA
Labelled HSA
Rose Bengal
Rose Bengal
Renal (unction compounds
Renal function compounds
Sodium lodlpamlde
Labelled macroalbumin
Labelled mlcroalbumln
Iodide
Iodide
Iodide
Bone Imaging
Brain imaging
Thyroid Imaging
Blood pool imaging
Salivary gland Imagine
Placenta localization
Bone Imaging
Liver/spleen Imaging
Kidney Imaging
Renal function
Kidney Imaging
Brain imaging
Lung imaging
Bone Imaging
Lung Imaging
Thyroid function
Blood volume
Renal function
Fat absorption
Thyroid function
Thyroid imaging
Thyroid function
Thyroid Imaging
Brain tumor localization
Placenta localization
Cardiac Imaging
Clsternography
Liver function
Liver Imaging
Renal function
Kidney Imaging
Cardiac Imaging
Lung Imaging
Liver Imaging
Hyperthyroldism
Cardiac dysfunction
Thyroid cancer
to 15 mCi
to 15 mCl
to 2 mCl
to 15 mCl
to 2 mC 1
to 1 mC 1
to 15 mCl
to 3 mCl
to 5mCl
to 10 mCi
to 10 mCl
to 15 mCl
to 4 mCl
to 15 mCl
to 4 mCl
5-25 iid
to 20 uCl
20-50 |iCl
25-100 (jCl
100 uCi
400 yCi
5-25 (jCl
to lOOfiCl
to 500 jiCl
5 (jCt
to 100 uCl
70-100 fiCl
20-50 (iCl
to 300 (jCl
20-50 jjCl
100 fiCI
300-500 ^Cl
to 300 (jCt
to 300 MCl
liquid
liquid
liquid
liquid
liquid
liquid
liquid
suspension
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
capsule
capsule
llq/capsule
liq/capsule
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
5-15 mCl/dose llq/capsule
30-50 mCl/course "
5-15 mCi/dose llq/capsule
30-40 mCl/cours« "
to 100 mC I/dose "
300-400 mCl/course '
ro
i
ro
-------
Table 2-1 (Continued)
USE
NUC LIDE
FORM
Internal therapy
In vivo
In vivo
In vivo
In vivo
In vitro
In vitro
In vivo
In vivo
In vivo
In vivo
In vitro
In vivo
In vivo
In vivo
Internal therapy
Internal therapy
In vitro
In vivo
In vivo
In vivo
Internal therapy
Internal therapy
Internal therapy
Internal therapy
Internal therapy
In vivo
In vivo
In vivo
In vivo
Iodine 131
Xenon 133
Xenon 133
Xenon 133
Xenon 133
Chromium 51
Chromium 51
Chromium 51
Chromium 51
Chromium 51
Chromium 51
Cobalt 57, 58, 60
Fluorine 18
Gallium 67
Gold 198
Gold 198
Gold 198
Iron 59
Iron 59
Indium 111
Krypton 81m
Phosphorus 32
Phosphorus 32
Phosphorus 32
Phosphorus 32
Phosphorus 32
Selenium 75
Strontium 85
Thallium 201
Yttarbrum 169
Iodide
Gas In saline
Gas In saline
Gas in saline
Gas
Chromate
Labelled HSA
Chromate
Chromate
Chromate
Labelled HSA
Labelled vitamin B12
Fluoride in saline
Citrate
Colloid
Colloidal
Colloidal
Citrate, chloride, sulfate
Citrate, chloride, sultate
DTPA
Gas
Soluble phosphate
Soluble phosphate
Soluble phosphate
Colloidal chromic phosphate
Colloidal chromic phosphate
Labelled melhionlne
Nitrate, chloride
Chloride
OT PA
TREATMENT/STUDY
Thyroid ablation
Cardiac abnormalities
Muscle blood Dow
Cerebral blood flow
Pulmonary function
RBC mass/survival
GI protein loss
Spleen imaging
Placenta localization
Red cell sequestration
Placenta localization
B 12 absorption
Bone cancer imaging
Tumor localization
Liver Imaging
Plcural effusions
Peritoneal effusions
Iron turnover
Iron distribution
CI stenography
Pulmonary function
Polycythemia vera
Leukemia
Osseous metastases
Pleural effusions
Peritoneal effusions
Pancreas Imaging
Bone cancer Imaging
Cardiac Imaging
CisUrDocraphy
USUAL DOSE STATE
50-100 mCl
to 50 mCl
to 200 MCl
to 1 mCl
to 10 mCl
to 200nCi
to 50 MCl
to 300 MCI
10 MCI
to 200 MCl
5-35 MCi
0.5 MCI
to 4 mCl
to 4 mCl
to 200 MCI
to 125 mCl
to 200 mCl
10-35 ;jCl
10-35 MCI
500 fiCl
3 mCl
3-7 mCi/dose
liq/capsule
gas in liq
gas in liq
gas in llq
gas
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
suspension
suspension
suspension
liquid
liquid
liquid
gas
liquid
8-10 mCI/course
1 mCl/dose liquid
8-10 mCl/course "
1.25 roCI/tiose liquid
20 rod/course "
to 20 mCl
to 25 mCl
to 250 MCI
SO-100 MCl
to l.SaCI
MOpCI
suspension
suspension
liquid
liquid
liquid
liquid
ro
I
ro
-------
2-28
123j
131
Although I has a long history of use in thyroid function
studies and in thyroid imaging, 123i offers several
distinct advantages which account for the growing interest in this
isotope.
These are:
a) Half-life (T]/2)
12*3
The half-life of I is about 13 hours as opposed to that of
131i which is approximately 8 days. The shorter T]/2
reduces cumulative patient exposure.
b) Gamma ray energy
The principal photonic emissions of ^"1 have energies of 159
KeV. This level is consistent with the spectral response of
current gamma cameras. The equivalent energies of 1311 photons
are about 364 keV.
c) Beta radiation
I emits abundant beta radiation which serves no medically
useful purpose in vivo but contributes to patient exposure. 1
is free of significant beta emission.
TOO
I is generated by either of two methods:
a) Bombardment of antimony or tellurium
121
In one procedure, Sb is bombarded with alpha particles, with a
consequent emission of two neutrons for each alpha particle
absorbed, as follows:
121CK, . J23T
Sb(a,2n) I
Both the mass and atomic numbers increase by two units in this
reaction. In another procedure, l23Te is bombarded with protons,
with a consequent emission of one neutron per absorbad proton.
The reaction is
123Te(p,n)123I
The atomic mass remains the same, but the atomic number increases
by one.
-------
2-29
b) As a decay product of 123Xe
123
Xe can be produced by
according to the reaction
123 122
Xe can be produced by bombarding Te with alpha particles
122Te(a,3n)123Xe
123
Xe is unstable, having a half-life of slightly over 2 hours.
It decays by positron emission (one positron per nucleus) with
a consequent drop in atomic number from 54 to 53.
99mTc
This radionuclide has a half-life of slightly over 6 hours. Its principal
radiation - gamma rays - has an energy level of about 140 keV, which is
excellent for imaging purposes. Reference to Table 2-1 will show that 99mTc
is used for visualizing a wide variety of organs and tissues. At present,
99myc -js usec| more extensively than any other isotope in imaging procedures.
QQrn
As previously mentioned, Tc is now produced mainly in radioisotope
generators in which the cow is 99^|0- 99fv|o itself is generally made either
by fission or neutron bombardment of 98M0. In recent years, however, there
have been studies addressing the feasibility of generating 99mjc using
accelerators. It appears that this can be achieved through protron bom-
bardment of lOOMo. in this reaction two neutrons are lost per nucleus,
so that the mass number decreases by one, while the atomic number increases
by one (i.e., 100 to 99 and 42 to 43), as is shown in the following:
100Mo(p,2n)99mTc
99
The possibility of producing Mo cow in accelerators has also been
explored. This can be accomplished by bombarding '00Mo with protons.
The resulting reaction is
100Mo(p,pn)99Mo
As is clear from the above, the net effect of the bombardment is a drop
in atomic mass of one unit with no change in atomic number.
-------
2-30
2.4 Applications
This section focuses on radionuclide labeled compounds of medical interest.
The applications of these materials fall into two broad categories:
• in vivo (within the living organism)
Labeled materials designed for clinical application are generally
administered orally or by injection. Their principal uses are:
(a) imaging; (b) function studies; and (c) therapy. Such materials
are "radiopharmaceuticals." Labeled compounds are also to be used
in vivo for medical or biological research purposes as distinguished
from clinical diagnosis or treatment.
• in vitro (in an artificial environment - literally "in glass")
This class spans a wide gamut of clinical laboratory tests and
analyses in which labeled compounds are employed as reagents.*
Detailed treatment of the medical applications of radionuclide labeled com-
pounds is outside of the scope of this report. However, the following
summary provides a general orientation.
a) Imaging
As the name suggests, imaging is a procedure for enabling the visual
examination of internal structures or functions. In a sense, it is the
reverse of conventional X-ray techniques in which an external source of
radiation is differentially transmitted through anatomic structures
on the basis of their different densities. In imaging, gamma-emitting
radionuclides incorporated in appropriate compounds may be selectively
absorbed by or become bound by the organ or tissue being
examined. For example, in the case of bone, the labeled compound would
be a phosphate. In other imaging procedures, the radionuclide may be
introduced into a body fluid (e.g., blood, cerebrospinal fluid) for
such purposes as delineating some portion of the cardiovascular system
or the ventricles of the brain. A considerable amount of imaging is
''The distinction between radiopharmaceuticals and labeled laboratory reagents
is evident. For convenience, however, the term "radiopharmaceutical industry"
is sometimes used as including the manufacture and distribution of all types
of labeled materials designed for medical and related uses.
-------
2-31
performed by scanning techniques in which a detector (typically, a
crystal that scintillates when excited by gamma radiation of appro-
priate energy level) is progressively moved in steps across the
area to be imaged. The generated presentation is thus a composite
or mosaic composed of many individual elements. With other techniques,
using a large stationary crystal, it is possible to photograph the
radiation from a given area simultaneously (gamma ray camera). In
this case, however, the size of the field that can be observed at one
time is limited by the dimensions of the crystal. Scanning methods
are useful when the structure to be visualized is large and static.
The imaging of dynamic processes, such as blood flow, obviously requires
simultaneous area photography. The principal purpose of imaging is
to detect and identify abnormalities, such as malignancy, vascular
defects (e.g., aneurism or stenosis), and other pathologic changes.
b) Function Studies
These studies are performed for the purpose of assessing the phy-
siological or biochemical activity of the organ or tissue of interest.
Such studies are performed essentially in vivo and include respiratory
exchange, blood volume measurements, gastro-intestinal protein loss,
vitamin B,9 absorption, iron metabolism and thyroid activity as measured
by 131I uptake.
c) Therapy
Compounds containing radionuclides are used in vivo for various therapeutic
purposes, including the treatment of metastatic bone tumors, thyroid
hyperfunction and malignancy, and polycythemia vera (excessive red blood
cell count). Where applicable, in vivo radiotherapy is considered to
have an advantage over irradiation from an external source in that the
former is more selectively directed with consequently less damage to
normal tissues adjacent to the treated site.
The number of isotopes that have been and are being used for medical purposes
and the variety of these uses are considerable. Table 2-1 lists several examples
of commonly used radionuclides and their applications, both in vivo and in vitro.
Note that in some instances the nuclides are used as tags or labels, as in the
case of iodine, while in others they are used in elemental form (e,.g., Kr).
-------
3-1
3.0 EXTENT OF ACCLERATOR USE IN UNITED STATES
3.1 Growth in Accelerator Use
Estimates of the number of particle accelerators by states in the United States
are published in the report of State and Local Radiological Health Programs^
by the Food and Drug Administration, Bureau of Radiological Health (BRH).
These data have been reported annually and were obtained for the years 1968
and 1970-1977. For the most part, the states exclude Federally owned machines.*
In many cases, data are missing due to failure of some of the states to
respond to the BRH questionnaire, upon which the report is based. These
missing data have been developed to the extent possible by using data from
previous and subsequent years to interpolate to the year of interest. In
some cases, we asked the state for the missing data.
The estimated number of particle accelerators in the United States reported to
BRH for the years 1968-1977 and the adjusted totals are shown in Table 3-1; a
breakdown by state is in Table 3-2. The adjusted totals for these years are
plotted in Figure 3-1. A least squares linear regression analysis was
performed on the data to obtain the growth curve shown in the figure. For
such a curve it is desirable to know how well the linear curve actually fits
the data; this measure is defined as the correlation coefficient. For this
curve, it was determined to be 0.99. A linear growth rate of 65 accelerators
per year was obtained from which the number of particle accelerators can be
projected:
1980 - 1355 machines
1981 - 1420 machines
1982 - 1485 machines
1983 - 1550 machines
*BRH defines accelerators to include machines producing electrons greater than
300 keV or ions greater than 150 keV.
-------
3-2
TABLE 3-1
PARTICLE ACCELERATORS IN THE U.S.
YEAR
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
TOTAL NUMBER REPORTED ^
576
N/A
639
790
714
875
963
833
845
1113
ADJUSTED TOTAL*
600
625
650
790
845
877
965
1029
1055
1191
*Total reported plus number of additional machines obtained by
extrapolation and by direct contact with states.
-------
3-3
1500
T 1 1 1
N=65.34Y-128,019
1968 69 70 71 72 73 74 75 76 77 79 79
FIGURE 3-1
GROWTH TREND OF PARTICLE ACCELERATORS IN THE UNITED STATES
-------
3-4
TABLE 3-2
ADJUSTED ESTIMATES OF PARTICLE
ACCELERATORS BY STATE - 1977
Number of Number Number of
State Accelerators Registered Users
Total 1191 982 825
Alabama 24 18 59
Alaska NA MA NA
Arizona 36 13 8-
Arkansas 755
California 177 177 85
Colorado 12 7 7
Connecticut 13 10 10
Delaware 737
District of Columbia 10 10 10
Florida 59 59 38
Georgia 21 21 15
Hawaii 222
Idaho 2 NA 2
Illinois 57 57 41
Indiana 26 33 33
Iowa NA NA NA
Kansas 10 10 6
Kentucky 10 10 8
Louisiana 15 18 14
Maine 000
Maryland 26 26 NA
Massachusetts 33 33 33
Michigan 57 57 NA
Minnesota 7 7 11
Mississippi 11 11 7
Missouri 11 11 6
Montana 222
Nebraska 856
Nevada 6 NA NA
New Hampshire 333
New Jersey 31 31 19
New Mexico 444
NA - Data not available
-------
3-5
TABLE 3-2 (continued)
ADJUSTED ESTIMATES OF PARTICLE
ACCELERATORS BY STATE - 1977
Number of Number Number of
State Accelerators Registered Users
New York 121 0 66
North Carolina 7 13 5
North Dakota 1 1 1
Ohio 30 23 23
Oklahoma 13 13 13
Oregon 19 19 19
Pennsylvania 83 69 67
Rhode Island 552
South Carolina 51 36 13
South Dakota NA NA NA
Tennessee 25 25 18
Texas 75 70 95
Utah 5 NA 5
Vermont NA NA NA
Virginia 11 8 8
Washington 26 26 26
West Virginia 666
Wisconsin 22 22 15
Wyoming 1 NA 1
Puerto Rico 3 2 1
NA - Data not available
-------
3-6
3.2 Use by Type of Machine
The annual BRH reports, discussed above, do not break down accelerator use
by type of machine. However, the questionnaire BRH distributes to obtain the
use data asks for such a breakdown. The responses to the 1977 questionnaire
in the area of accelerator use were made available to Teknekron. These re-
sponses were compiled into the seven categories of accelerators found in
the questionnaire. The results are presented in Figure 3-2, which clearly
shows that linacs are the most widely used machine.
This trend is largely due to the increasingly wide use of linacs in the medical
(2)
therapy field. A BRH report on electron linacs in medical radiation therapyv '
indicates that these machines are replacing teletherapy sources (e.g., cobalt-60)
that require periodic replacement due to radioactive decay, and older medical
therapy accelerators (e.g., Van de Graaffs, resonant transformers). This
report estimates a 15% per year growth in medical linacs, and projects 400 to
450 such machines in use in 1977 (see Figure 3-3). The present study estimates
that 590 linacs (all types) were in use in 1977, and about 70% of them (or 420)
were used in medical applications.
The growth of medical linacs at the rate of 15% per year is expected
(2)
to continue through 1980. ' There is a basis for projections beyond 1980
since the market for these machines is far from saturated and includes hospitals
and clinics currently operating cobalt-60 units and those considering the
purchase of a megavoltage therapy unit. The total number of machines reported
by type is 870, or 73% of the estimated total in 1977. We can see no bias in
the results and assume that the remaining 27% of the machines are distributed
in the same way. However, data obtained from accelerator industry sources^3'
indicate that about 400 Cockcroft-Walton type machines are in use as ion
implanters.* Most of these machines should appear in the "other" category in
Figure 3-2, which shows only 25 machines. Ion implanters are either unregistered
*Ion implanters are used in the semi-conductor industry to implant ions in
silicon and germanium.
-------
3-7
or greatly underreported by many states. Because of their design and low energy
(^0.2MeV), ion implanters do not create airborne radioactivity, so this anomaly
was not investigated further.
While implantation is the most popular application of industrial accelerators,
other applications include industrial radiography (X-ray and neutron generators),
analysis of materials food processing and sterilization, and curing of coatings
and paint. Most X-ray generators used for industrial radiography are not included
in our count of accelerators because most fall below the energy cutoff of 300 keV
(electrons). Since their average potential is 150 kVp, only a small fraction
of the estimated 3300 radiographic X-ray generators are electron accelerators
of greater than 300 keV.^3a'
-------
500-
FIGURE 3-2
ACCELERATORS REPORTED BY TYPE
(From BRH Questionnaire)
400
O1
I
Q
LU
a:
o
a.
LU
a:
a:
LU
CO
300 _
5 200.
100-
15%
17%
6%
50%
3%
3%
Van de
Graaff
Neutron Resonant Linacs
Genera- and Insu-
tors lated Core
Transformers
Cyclotrons Betatrons Other
PERCENT OF TOTAL BY ACCELERATOR TYPE
co
oo
-------
I I I I
ilOOO
900
800
700
600
500
400
300
200
100
Operational
Megavoltage
Radiation
Therapy Units
Cobalt-60
Units
1 I I I F71 I I f
60/yr
Sales-^//*"-Operational Units
// (Assuming 15 yr.life)
/ !
Uncertainty-^/
i /-l
Projected
Increase
15%/yr
// -
Medical
Linacs
(Van de Graaffs
< and Resonant
(Transformers
Betatrons
i I I I
1965 'l970
1000
900
800
700
600
500
400
300
200
100
975
1980
FIGURE 3-3
MEGAVOLTAGE RADIATION THERAPY EQUIPMENT PROJECTIONS FOR THE
UNITED STATES THROUGH 1980^
co
i
UD
-------
3-10
Additional sources of data on accelerator use were found in the American
(4 5)
Institute of Physics (AIP) Handbooks. The first and second editions v ' of
the handbooks contain listings of all U.S. accelerators for the years 1955
and I960, including location, operator/owner, and machine parameters (e.g.,
energy, particle). The data on non-Federally controlled machines were com-
piled under the same general machine classifications used earlier. However,
the class of "neutron generator" was removed since it is an application
(usually of a Cockcroft-Walton machine) that was not identified in the AIP
listings. In order to obtain current figures for comparison with the data
available from the above sources for 1955 and 1960 and the 1977 BRH study,
Teknekron contacted a number of directors of State Radiological Health Programs.
The information obtained represented a total of 376 machines and provided
their classification by kind, particle type, and energy. The results are
presented in Table 3-3, which also gives similar information from the AIP for
1955 and 1960.
The percentages of the total number of machines by type from this table, in
addition to the same breakdown from the BRH questionnaire (Figure 3-2), are
depicted for comparison in Figure 3-4. The shift away from Van de Graaffs
and cyclotrons and towards linacs is evident from this figure.
3.3 Machine Use by Location
An attempt was made to characterize accelerator use by location. References 4
and 5 were relied on for location data (i.e., city) for the years 1955 and 1960,
while current information was obtained from the Teknekron survey. Of the 129
machines operating in 1955,' ' 44 were classfied by location. The Teknekron
survey yielded location data on 376 machines.
A simple classification scheme was used: urban, suburban, and rural. An
urban area was defined as a city with a pooulation of at least 50,000; a sub-
urban area was defined as a city with a population of at least 30,000 and bor-
dering an urban area. All other locations were considered rural. Having located
the accelerator by city, the location was then classified accordingly.
-------
(a) TEKNEKRON - 1979
Energy
MeV
<1
1-19
20-100
>100
Total
5! Of
Total
Van de Graaff
e~ * 1on
10 3
26 23 6
1
1
-70-
18.1
Resonant and
ICT**
e" * 1on
7
8
-15-
4.0
L1nac
e" * 1on
9
134 41 1
4 3
-192-
51.1
Cyclotron
P * d
2
2 2
6 2
1 3
-18-
4.8
Betatron
e"
4
4
22
30
8.0
Neutron Generator
8
14
-22-
5.9
C-W, Other
e" * 1on
5 4 18
1 1
-29-
7.7
(b) AMERICAN INSTITUTE OF PHYSICS - 1960
(c) AMERICAN INSTITUTE OF PHYSICS - 1955
TABLE 3-3
ACCELERATOR USE BY TYPE AND ENERGY
Energy
MeV
<1
1-19
20-100
>100
Total
% of
Total
Van de Graaff
e" photon 1on
8
30 33 50
-121-
48.5
Resonant and
ICT
e" 1on
1
2
-3-
1.1
Unac
e" 1on
14 1
9 2
1 2
-29-
11.3
Cyclotron
p d a
2
13 5
1 1
6
-28-
10.8
Betatron
e"
1
27
2
-SO-
IL 7
Synchrotron
e" P
4
4 2
-10-
3.9
C-W, Other
1on e"
31
2 1
-34-
13.2
Energy
MeV
<1
1-19
20-100
>100
Total
« of
Total
Van de Graaff
e" photon ion
1 5
7 9 37
-59-
45.7
Resonant and
ICT
e" Ion
-0-
0
Unac
e" 1on
1 1
5
1 2
1
-11-
8.5
Cyclotron
p d a
5 11
1 1 3
5
-25-
19.4
Betatron
e"
1
7
3
-11-
8.5
Synchrotron
e" P
2
7 1
-10-
7.7
C-W, Other
1on e"
13
-13-
10.1
OJ
I
*Particle not specified
**Insulated core transformer
-------
3-12
1955 1960
1970
1977 1979
YEAR
FIGURE 3-4
ACCELERATOR USED BY TYPE
-------
70
60
O 50
u.
O
40
Z
UJ
o
£ 30
a.
20
10
I
URBAN SUBURBAN
1955 (AIP)
1960 (AIP) -
1979 (TEK)
RURAL
FIGURE 3-5
ACCELERATOR USED BY LOCATION
-------
3-14
The results are presented in Figure 3-5 and snow a definite trend toward use
in urban areas. This is most likely due to the widespread use of linacs as
cancer therapy units in urban hospitals, and, to a lesser extent, accelerators
being used in commercial applications in urban industrial areas.
-------
4-1
4.0 POTENTIAL FOR RESIDUAL AIRBORNE ACTIVITY
4.1 Accelerator Radiation Hazards
Accelerator radiation fields are of two distinct types. The prompt radiation
field,, produced during the existence of a beam or operation of the accelera-
tor, has the greater energy content, and thus constitutes the greater poten-
tial hazard. However, this prompt radiation, from primary and secondary beams
and scattered radiation, does not generally present a biological hazard
because shielding and exclusion are employed to limit the exposure of radiation
workers to allowable levels.
The residual field, remaining after the beam or accelerator is off, is created
when the energy of the prompt radiation exceeds the threshold for nuclear
reactions. Interactions in materials of the accelerator complex and its con-
tents produce lingering radioactivity with different half-lives and decay
radiations. In higher energy accelerators, the exposure of personnel to
the activation products in certain machine components, such as target and
extraction equipment, constitutes the principal biological hazard from
external exposure.
If the primary beam emerges into air before reaching its target, ozone will
be produced and, if its energy is sufficient, nuclear reactions with the
gas molecules will occur. Since air is composed chiefly of nitrogen and
oxygen, the majority of the radioactive products have short half-lives.
Radionuclides with lives of less than one minute (e.g., N - 0.01 seconds),
are not of concern since they are quickly removed by decay. Long-lived
activities (e.g., 3H - 12 years) do not constitute a hazard because of their
low activity. The principal hazard is exposure to personnel entering the
target area immediately after machine shutdown, if area ventilation is
inadequate.
-------
4-2
4.2 Mechanisms for Air Activation
The most important mechanisms by which radioactivation of air atoms may occur
as a consequence of accelerator operation are:
• Direct exposure of air to the primary beam, if the beam passes
from the accelerator tube vacuum into the atmosphere, and
• Exposure of air to secondary radiation from the target.
In addition, radioactive gases may be produced within the target and subsequently
escape to the local environment (if the target is external to the vacuum); in some
instances, the target may contain inherently radioactive material (e.g., tritium).
In general, the identities of the radionuclides generated by air activation and
their rates of generation depend on several factors, including the type, energy,
and intensity of the primary beam and the nature of the target.
Protons and heavy ions accelerated to high energies produce nuclear reactions
directly. All energetic accelerated particles give rise to protons and neutrons
as secondary radiations from interactions in various targets. Table 4-1 lists
the important nuclear reactions initiated in air by protons, neutrons, and
photons. Reaction threshold energies and cross sections are also shown.
Table 4-2 lists the products of the above reactions and other nuclides capable
of being produced in the particle fluxes near accelerators.
4.3 Calculation of Production Rates
A representative equation useful for the calculation of production rates of
airborne radionuclides comes from Patterson and Thomas, p. 520, Equation 12:
-------
4-3
TABLE 4-1
NUCLEAR REACTIONS RESPONSIBLE FOR MOST
AIRBORNE RADIOACTIVITY AROUND ACCELERATORS
Average
Threshold Cross Section
Reaction Parent Energy (MeV) (mi Hi barns)
(Y.n)
(n,2n)
(P.pn)
(n,Y)
14N
160
12C
14N
160
12C
160
14N
40Ar
10.5
15.7
18.7
11.. J
18
20
10
10
3*
11*
10*
6
40
20
33
10
610
*Resonance cross sections at about 22 MeV range from 50-150 mb.
-------
4-4
TABLE 4-2
LIST OF AIR ACTIVATION RADIONUCLIDES
Isotope
150
13N
16N
140
"c
41Ar
7Be
3H
38S
18p
Half-life Principal Means of Production
2
min
10 min
7
1
20
1.
53
12
37
2
sec
min
min
9 hr
days
160
14N
160
14N
12C
40Ar
12C
(n,
(n,
(n,
(P,
(n,
2n)
2n)
P)
n)
2n)
» (Y»n)
, (Y,n)
and 15N
and (Y
and (p,
and (p,
(n.Y)
.n)
pn)
pn)
(n.Y)
(3He,
2 a) and
others
yr various
min
hr
40Ar
16o
(Y
, 2p) and
37r. ,
Ci (a,
3p)
( o,pn) and others
-------
4-5
where S is the total specific activity of radioactive air (per liter),
C is a constant,
*r *Uv and *HE are lhe average photon, thermal neutron,
and high-energy particle flux densities,
°iJ7' °ijth and °ijHE are lhe corresponding average cross sections.
N- is the number of target nuclei of type j in a liter 01 aimosprteiK
air.
X . is the decay constant of the radionuclide i,
T is the irradiation time,
and t is the decay time.
A simpler equation is foun^l in Sarbier, p. 15, Equation 3.8:^ '
n: (specific activity of isotope i produced per unit time) =
m
where N is the number of gram-atoms of target material,
A. is the atomic weight of air (15g),
U
tm is the mean life of isotope i, and
other terms are defined as above.
Essentially, the production rates of air activation products are functions
of the following factors:
1. the incident flux () or secondary flux produced by the accelerator,
2. the density of target gas atoms in the beam path,
3. the appropriate cross section for the reaction and the interaction
energy,
4. a buildup factor, based on exposure time, having a maximum in the
equilibrium condition, and
-------
4-6
5. the decay that occurs between termination of the irradiation
and the time of exposure to the decay radiations.
As an example, calculations were performed to estimate the production rates
of 150, 13N, and ^C in the presence of a high-energy neutron flux. The
conditions assumed were a 1 uA proton beam striking a thick beryllium target
12
at 100 MeV. According to reference 6, this arrangement produces 2 x 10
neutrons per second with energies distributed around 100 MeV. Assuming a
2
1 cm beam traversing 1 meter of air, and using the abundances of C, N, and 0
isotopes in air, production rates were calculated. They are presented in
(19)
Table 4-3. Assuming 10 air changes per hour,v ' the total activity produced
in 4 hours was also calculated. This is equivalent to the activity released
to the atmosphere if it is assumed that activation products are exhausted
immediately after production.
TABLE 4-3
PRODUCTION RATES OF 150, 13N, and ]1C IN AIR
Production Rate Amount Produced
Reaction atoms/sec in 4 hours (Ci)
12C(n,2n)]1C 6xl04 7xlO"5
14N(n,2n)13N 5xl07 5.3xlO"2
160(n,2n)150 9xl07 9.4xlO"2
An estimate of the thermal neutron flux is not available for this situation;
thus production rates of Ar and other (n,x) products cannot be calculated.
However, it is possible to use the measured thermal flux around an 18 MeV
electron linac, a machine widely used in the United States (see Section 3.3).
This flux was reported to be about 440 n/cm2/sec.^17^ Assuming a 2000-hour
work-year, the activity produced annually of 41Ar and 14C in a 27 m3 room with
10 air changes per hour is 10 and 10 curies, respectively.
-------
5-1
5.0 EFFLUENT MONITORING DATA
5.1 Existing Data
Of the existing monitoring data reviewed, references Sand 9 are the most
closely related to this study, in that the production of radioactive gases
was investigated at three 40 MeV electron linacs operating at up to 0.5
rnA. Two linacs^ were operated with either aluminum targets which produced
bremsstrahlung (maximum energy ^20 MeV), or with tungsten targets which
produced photoneutrons.
Full power (25 kW) operation with a tungsten target produced, at one acceler-
ator, a maximum equilibrium gas concentration of 5 x 10 yCi/cc. ' This
15 13
was composed of 45% 0 and 55% N. However, similar operating conditions
at the second linac produced a concentration of 4 x 10 uCi/cc. This dif-
ference was attributed to the lack of ventilation at the second facility
which limited the amount of air available for interactions with the post-
target beam. The authors also observed a strong dependence of gas concen-
tration on the target arrangement, especially local shielding. Lack of such
shielding allows more radiation to interact with air, producing higher concen-
trations of radiogases. Use of an aluminum target in place of tungsten did
not significantly affect gas production.
Measurements at a 40 MeV linac, discussed in reference 9, detected trace amounts
of Cl and Ar, as well as the more common air activation products C, N,
and 0. Production rates of the latter three nuclides ranged from <.l MCi
per pulse of 20 MeV electrons, without a bremsstrahlung target, to 2 yCi per
pulse of 45 MeV electrons, with a bremsstrahlung converter. Relative proportions
of the nuclides were found to be extremely dependent on the operating energy,
since many of the operating energies are near or below one or more of the thresh-
olds of the fy, n)production reactions. Total facility release quantities were
given as being "in the curie range."
The calculation of radioactive gas production rates at a 100 MeV electron linac
is the subject of reference 19. Using a theoretical bremsstrahlung energy spectrum
-------
5-2
and gamma-neutron cross sections for oxygen and nitrogen, production rates
were calculated for 13N and 150. Equilibrium concentrations of these nuclides
were calculated to range from 10 to 1000 yCi/m3, depending upon room size and
ventilation rate. It was pointed out that electron linacs must be capable of
energies of 50 MeV or more to activate nitrogen and oxygen via the (y,n)
fractions.
The author also calculated an occupational (MPC) of 4 x 10" jaCi/cc based on an
n
external dose due to N. It was shown that a wait corresponding to 10 air
changes (M hr) is necessary after shutdown before uncontrolled access to
the target room can be permitted. Finally, the author mentions that exposure
hazards outside the target rooms will be generally insignificant if the exhaust
is discharged from a stack extending above the roof of the building. '9'
Reference 10 discusses in great detail the release of gaseous tritium from a
Cockcroft-Walton neutron generator. This machine produces neutrons by
deuterium bombardment of a tritium/titanium target [T(d,n) reaction]. During
irradiation of the targets, which typically contain about 5 curies of tritium,
some tritium is sputtered off of the target surface. Although the target is
contained in the evacuated beam tube, tritium can be released to the room
through the vacuum pump exhaust. The amount released in this manner is esti-
mated to be 100 to 300 mCi per target expended. Finally, the authors recommend
a simple tritium trapping system for the vacuum pump exhaust.
The remaining articles deal with two main topics: 1) gas and dust activation
at high energy (BeV) accelerators, and 2) stray radiation characteristics of
smaller (MeV) machines. The high energy accelerators are essentially all excluded
from this study due to being Federally controlled or non-U.S. machines. Never-
theless, the published information on radiogas formation at these machines will
be briefly reviewed.
References 11 through 15 discuss radiogas production at two proton accelerators
having maximum energies of 600 and 800 MeV, and two high energy electron linacs.
-------
5-3
The predominant gaseous nuclides formed at all machines were 150, 1]C, 13N,
41 c
and Ar. Gas concentrations at the proton accelerators ranged from 2 x 10
to 7 x 10" yCi/cc. Those at the 150 MeV electron linac averaged 1.0 x 10"6
yCi/cc (150 + 13N) and 2.4 x 10~7 yCT/cc (41Ar), during bean runs.(15) Investi-
gators at the 550 MeV electron linac were able to detect 7Be, 24Na, and 56Mn
produced by dust activation/14'
From the dosimetry discussions in these references, it was concluded that the
skin is the critical organ due to the predominant beta dosej12^ and that this
dose contributes only a small fraction of the total annual offsite
Neutron flux around smaller accelerators is the subject of references 16, 17,
and 18. Knowing the thermal neutron flux allows calculation of the
production rates of those isotopes produced (by n.Tl) reactions. A useful
t
n/cm'
(17)
2
value obtained was 440 n/cm /sec measured in the treatment room of a Clinac
18 medical accelerator.
5.2 Data from Study Facilities
Two facilities were selected for gaseous effluent monitoring in this study:
the 100 MeV isochronous cyclotron at the University of Maryland and the
6 MeV positive ion Van de Graaff located at the University of Kentucky.
Three potential sources of airborne radioactivity were investigated at each
facility before beginning the actual sampling. They were the machine cavity,
the beam tube vacuum pump(s) exhaust, and target preparation hoods. At the
cyclotron, a large, centrally located exhaust duct (see Figure 5-1) carried the
cavity exhaust air and the vacuum pumps' exhaust to a roof vent at 481,000 liters/
nin (17,000 cfm). The air turnover rate in the cavity was estimated to be 6 to 7
air changes per hour. Target preparation hoods were not utilized during the study
Deriod.
"he Van de Graaff cavity employed a closed air circulation system operating at about
13,000 liters/min (4000 cfm) (return side). The return air was sampled,
-------
5-4
I_ j tztc.ec fL A!
^j>'L'iJ-L -,^_
> tJ VI -
iW—U^!—-<
J-'vv
1 <. • o •
Note: At the center of this top-view schematic is the "magnet shaft."
Here the beam is carried vertically, as well as horizontally (as
shown), to another level of experiment areas. The single ventilation
air exhaust duct is located at the top of the magnet shaft. Thus ,
the air flows from the cyclotron room and experiment areas into
the magnet shaft (through breaks in the shield walls), into the
exhaust vent and up to the discharge point on the roof.
FIGURE 5-1
SCHEMATIC OF CYCLOTRON FACILITY
-------
5-5
although this air was not released to the atmosphere. The vacuum pumps'
exhaust was piped to the top of the cavity "silo" where it was vented near
a small roof ventilation fan (see Figure 5-2). Finally, a fume hood where
gaseous tritium targets were handled was sampled where it was vented atop
the building. The flow in this vent was measured to be approximately 17,000
liters/min (600 cfm).
No effluent treatment systems were in place or in use on any of the ventila-
tion systems described above.
5.2.1 Sampling Techniques
Sampling was performed by drawing a gaseous sample from the release plenum or
stack and collecting the sample in a chamber or on a trap. The flow rate of
the sample line was about 28 liters/min (1 cfm) through the filter cartridge
and about 7 liters/min (.25 cfm) through the silica gel. The schematic of the
sample train is shown in Figure 5-3. A summary of the sampling equipment used
is presented in Table 5-1.
A specially fabricated 2-liter gas collection chamber was also used. The
chamber was designed to fit over the gamma detector in a Marrinelli flask
configuration. The chamber was evacuated prior to sample collection. Details
of the sampling techniques are given in Appendix B.
5.2,2 Detection Techniques
Three instrumentation systems were used in this study for the detection of radio-
activity in gas and filter samples. For the detection of gamma (and positron)
emitting nuclides, a Ge(Li) spectrometry system was used. A 60 cc lithium
drifted germanium detector was coupled to a 4096 channel pulse height analyzer
(Canberra 8180). Spectra were stored on magnetic tape for later analysis. This
allowed sequential spectra to be taken on the same sample.
Siltca gel was used in the sampling train to trap HTO. After sampling, the
-------
o * ~l
Silo >•
j Exhaust Fan
Sample
#11
Taken
Here
Cavity
Pump
Sampl
O
0
O)
o
(J
«=c
-
Exhausty
e_Poijrt
Beam
*^Tube
Building Separat
30' at Closest P
Vacuum Pump
Vent Line
io
oi
P\ Cavity
\ ) Ground /Ventilation
U Love! / Register
Vacuum ^
Pumps 1 ]_-*-
u
\\
^
_.
i^
_J/
Cavity
n
nt
Isolation
/Dampers
V ,
^ f
r\ Hood Vent
/
Lab and
Classroom
Building
^
Cavity
"* Exhaust
Sample
Air Point
Is Conditioned
and
Returned
(
r^ Sample Point
v Blower
IL/
^
/ \
/ \
Fume
Hood
i
0^
FIGURE 5-2. SCHEMATIC OF VAN DE GRAAFF FACILITY
-------
coupling
glass probe
duct wall
r
— f3
cr
f i 1 ter
(Ccartridge
k
_1
L
-
-
^
r
side stream
V'flow meter
flow
i/^control
L
\s
J
^
*
(l
T
silica gel
/* cartridge
^^ breakthrough
indicator
1
__ /
exhaust
sample ptnfip
vac vac
run
time
flow
meter
flow control fuse
FIGURE 5-3. SAMPLE TRAIN SCHEMATIC
en
i
-------
5-8
TABLE 5-1
DUCT SAMPLING EQUIPMENT
FUNCTION
Participate filter
Iodine and gas filter
Carbon dioxide filter
Water vapor (HTO) filter
Breakthrough indicator
Air pump/flow meter
MATERIAL or ITEM
Glass fiber HEPA
Activated charcoal
Sodium zeolite (NaX)
Grade 05 silica gel 6-16 mesh
Grade 42 silica gel 6-16 mesh
Radeco model HD-28/B
-------
5-9
silica gel was weiahed. mixed, and a 3-aram sample was added to 4 cc of
liquid scintillation solution. The mixture was counted on a Beckman LS-150
liquid scintillation detector discriminated for tritium beta radiation.
Finally, selected C02 samples (collected on a molecular sieve) were analyzed
for C. This was accomplished by driving off the trapped gases with heat,
identifying the off-gases with a gas chromatograph, and isolating'the C02
driven off. The fraction was then counted in a 100 cc gas proportional counter.
5.2.3 Results
The largest amount of radioactivity produced at the cyclotron was detected
on February 6. On this date, the machine was delivering 100 MeV protons
to a beryllium/aluminium target, ultimately yielding fast neutrons with
energies distributed around 100 MeV. Beam current was approximately 2yA at
the target.
Nitrogen-13 was detected on the activated charcoal sampler. The physical
form is gaseous since particulate activity was not detected. The chemical
form of the N is still in doubt since many nitrogen compounds (even Ng)
can be trapped in activated charcoal. A backup charcoal cartridge also
contained N (about 17% of the quantity that was detected on the front
cartridge). Based on the activities detected on the two cartridges, a collection
efficiency of 83% is calculated.* The decay curve of the sample activity is
shown in Figure 5-4.
The molecular sieve sampler, as expected, collected carbon-11 dioxide. The
results of the February 6 sampling are shown in Figure 5-5. No other runs
produced detectable 1]C activity. The collection efficiency was determined to
be 8% at 28 liters/min. Although 150was present in the exhaust gas (see below),
it did not form C( 0)o and was not collected in this sampler.
*Efficiency = 1-
-------
5-10
o
u.
O
25
oc
100
16=10 mln.
-NOTES: ERROR BARS ARE ±1 SIGMA COUNTING ERROR. COUNT
RATE AT t-0 IS 7.6 cps. SAMPLE ACTIVITY IS .017 ^Ci.
COLLECTION EFFICIENCY IS 83%. AIR CONCENTRATION OF
13U IS 3 x
N
' nCi/cc.
J.
200
400 600 800
ELAPSED TIME (sec)
1000
1200
1401
FIGURE 5-4
DECAY OF ACTIVITY ON CHARCOAL-A SAMPLE, FEBRUARY 6
-------
5-11
11,
= 20 mln.
o
o>
in
«
*-
u
3
Q.
>
0>
o
<
NOTES: COUNT RATE AT t=0 IS1.9cps. SAMPLE ACTIVITY IS
0.0072 ^Cl. COLLECTION EFFICIENCY IS 8%. AIR
CONCENTRATION OF
_L
_L
J_
100 200 300 400 500 600 700
ELAPSED TIME (sec)
800
900 1000 1100 1200
FIGURE 5-5
DECAY OF ACTIVITY ON NaX-A SAMPLE, FEBRUARY 6
-------
5-12
One of the two gas samples taken on February 6 contained activity decaying
with the two-minute half-life of 0 (see Figure 5-6). The second gas sample,
taken with the beam off, showed no residual activity. The concentration of
13N and ]1C were too low to be detected in this manner. The relative con-
centrations of these isotopes agree fairly well with those predicted in
1 o
Section 4.3; however, N was in relatively lower concentrations than predicted.
The only other activity detected at the cyclotron was trace N activity on
_q
January 27, 30, and 31. The concentrations fluctuated around 10 yCi/cc.
No tritium was detected at this location. Results of the gamma ray and tritium
measurements are given in Appendix A.
The measurement at the Van de Graaff accelerator had a different focus.
At this location, the positron emitters 0, N, and C were not detected,
but tritium was found. The source of the tritium was a target assembly con-
taining ^130 curies of tritium on a uranium "getter." The target cell held
1.5 curies and was separated from the beam tube by a molybdenum foil.
Several pathways exist from the gas cell into the ventilation air. In the
first, tritium travels into the beam tube through minute holes in the foil.
In fact, tritium contamination had been found to exist within the beam tube.
Gases in the beam tube would be vented at the top of the cavity by the vacuum
pumps. When holes in the foil are detected, the tritium in the gas cell is
reabsorbed on the uranium and the target assembly is removed. Some tritium
may escape during this operation. Finally, the assembly is carried to a fume
hood where the foil is replaced and leak tested with helium. Tritium may also
escape during this operation.
As shown in Table A-2 , tritium concentrations were highest in the fume hood
exhaust (10 yCi/cc); the next highest concentrations were in the vacuum pump
exhaust (10 yCi/cc). Tritium concentrations in cavity ventilation air ranged
_o _y
from 10 to 10 yCi/cc; the highest levels were detected during bombardment of
the tritium target.
-------
5-13
100
o
o
(0
at
a.
9
122 sec
110
NOTES: COUNT RATE AT 1 =0 is .70 cps. 15 CONCENTRATION IN
SAMPLE IS 2 • 10*
1001
200
300 400
ELAPSED TIME (sec)
500
600
FIGURE 5-6
DECAY OF ACTIVITY OF 2 LITER GAS SAMPLE #1, FEBRUARY 6
-------
5-14
Several molecular sieve cartridges from each facility that had collected
CO were analyzed for 14C. The results, presented in Table A-3, indicate
14C was present in one of the two samples taken at the Van de Graaff
facility and in none of the samples from the cyclotron. Carbon-14 is oro-
duced by the N(n,p) reaction. The thermal neutron cross section is
1.8 barns;that for fast neutrons is about a factor of 10 lower. Using the
o
thermal neutron flux measured in the Van de Graaff cavity (50 n/cm /sec),
a 14C concentration (after 24 hours of operation) of 0.1 pCi/m is calculated.
The expected concentration on March 6 would then be 1.1 pCi/m including
o
background; however, 100 pCi/m was measured. This result is even more suspect
because it was found in only one of two C00 samples that were taken a week
14
apart, under similar neutron beam conditions. The high C concentration
measured on March 6 remains an anomaly.
5.3 Release Estimates.
As discussed above, operations at the cyclotron involving protons and heavy
13 Q
ions produce small amounts of N. Assuming exhaust concentrations of 10
yCi/cc and a work factor of about 1000 hrs/yr,* approximately 38 mCi/yr
are released to the atmosphere. However, when high energy neutrons are
produced, 0, N, and C are present in the exhaust air from (n,2n) reactions.
Release rates for these nuclides are 16 yCi/sec, 0.27 yCi/sec, and 0.43 yCi/sec,
respectively. The high energy neutron experiment observed lasted about four
hours.
As indicated earlier, two exhaust fans at the Van de Graaff facility actually
exhaust to the atmosphere. These are the fume hood exhaust and the small
ventilation fan at the top of the cavity (see Figure 5-2). The majority of the
cavity exhaust air is recycled. Given the concentration of tritium in the
fume hood exhaust, the release rate during foil changing was 0.43 curies per 24
hours. Since the cell capacity during operation is 1.5 curies, this 24-hour release
* Determined from facility operating schedules.
-------
5-15
constitutes nearly one-third of the tritium used in the cell. However, it is
highly unlikely that this quantity of tritium failed to return to the uranium
getter during cell evacuation. Either there existed another source of tritium
in the hood, or heating of contaminated surfaces occurred releasing tritium,
or an equipment failure and/or operational error occurred in the hood. A more
definitive description of the source of the 0.43 curies of tritium is not pos-
sible without further measurements. Estimating annual releases of tritium
from this source is difficult for the same reason. Nevertheless, arbitrarily
assuming semiannual releases of this magnitude, the annual release rate would
be on the order of 1 curie per year.
The estimated release from the cavity is due to one small (%23,000 L/min esti-
mated) ventilation fan near the end of the vacuum pump exhaust line. Using the
tritium concentration detected 15 meters below the exhaust fan (sample #11), a
release rate of 0.022 yCi/sec is calculated. A release rate from the vacuum pump
exhaust line was not calculable since flow from this line was too low to be mea-
surable. It is assumed that the above release rate is typical of all accelerator
operations using the tritium target; however, this is probably conservative
since leakage in the target cell occurred during the study. Using the work
factor of the target of 0.36, the annual tritium release from the cavity is es-
timated at 0.3 curies. Estimates of the tritium inventory on the uranium getter
were not available as a function of time.
-------
6-1
6.0 EFFLUENT TREATMENT
6-1 Existing Standards for Air Treatment
The existing standards limiting the amount of radioactive gases released
to the atmosphere were established in 1961 by the Atomic Energy Commission
under the recommendations of the ICRP and the NCRP. The Maximum Per-
missible Concentrations in air for radionuclides in unrestricted areas
were published in the Code of Federal Regulations Title 10, part 20,
and are currently enforceable by the Nuclear Regulatory Commission.
However, they apply only to radioactive materials subject to licensing
and not to radiation-producing machines such as accelerators.
The Suggested State Regulations for Control of Radiation^21' were initially
published in 1962 by the Council of State Governments with the Assistance
of the AEC and the BRH. Part D of these model regulations contains standards
for protection against radiation based on the 10 CFR 20 limits for un-
restricted areas. In 1974, this publication was revised to reflect amend-
ments, new information, and other changes in guidelines, making it com-
patible with Federal regulations. A section on particle accelerators was
added as Part I, "Radiation Safety Requirements for Particle Accelerators."
A subsection, Part 1.12, "Ventilation Systems," states:
(a) Adequate ventilation shall be provided in areas where
airborne radioactivity may be produced.
(_b) A registrant [or licensee], as required by D.106, shall not
vent, release or otherwise discharge airborne radioactive material
to an uncontrolled area which exceed the limits specified in Part
D, Appendix A - Table II, except as authorized pursuant to D.302
or D.106(b). For purposes of this paragraph, concentrations may
be averaged over a period not greater than one year. Every reasonable
effort should be made to maintain releases of radioactive material
to uncontrolled areas, as far below these limits as practicable.
Since the Agreement States have promulgated regulations for their own licensing
and/or registration of particle accelerators, it is assumed that the states
also follow these suggested requirements for ventilation.
-------
6-2
6.2 Treatment Systems
Guidelines for the treatment of air from accelerator facilities are
based mainly on dilution and containment at the source prior to exhaustion.
These ventilation and exhaust procedures and their effect on the concen-
trations of gaseous emissions released to the atmosphere will be dis-
cussed. Air cleaning devices and their cost-effectiveness will be dis-
cussed separately.
6.2.1 Ventilation
Ventilation and exhaust procedures have been recommended by several govern-
mental agencies and private groups with the intent of reducing radiation
levels in the immediate vicinity of the accelerator. The prevailing
philosophy has been that safe levels obtained internally would insure
minimal public exppsure.
The American National Standards Institute, with the cooperation of the
National Bureau of Standards, established in 1969 the standard, "Radiological
Safety in the Design and Operation of Particle Accelerators." This
standard requires adequate ventilation of areas where airborne radioactivity
is expected to exceed the Maximum Permissible Concentration* and dispersal
of the vented air in a manner compatible with existing air pollution laws.
Also, precautions should be taken to prevent the possibility of exhaust air
being drawn into neighboring air intakes.^ '
The National Center for Radiological Health under the Department of Health,
Education and Welfare published at about the same time the "Particle
Accelerator Safety Manual." Among its suggested practices for the control
of airborne radioactive materials is the provision of a ventilation system
apart from the building ventilation and so designed that the accelerator
areas are at a lower pressure than the other parts of the building.
*10 CFR 20, App. B. , Table II.
-------
6-3
Additionally, air flow rates should be variable so that air flow can be
increased during periods of high beam intensity.^33)
The National Accelerator Safety Committee prepared a similar report
for high energy accelerators, but with emphasis on vent system design
and maintenance as precautions against the release of hazardous gases in
the event of accidental ignition or other equipment failure.^34^
More recently, the National Council on Radiation Protection published
"Radiation Protection Guidelines for 0.1-100 MeV Particle Accelerator
Facilities." Included in this report is a discussion of ventilation
and ducting. Exhausting the air nearest the electron window and the
use of high capacity blowers are among the suggestions for the reduction
of hazardous gases that accumulate. The necessity for controlling dust
and other particulates by filtration is also mentioned. '
Ventilation practices vary widely according to the mode of operation of the
accelerator and the physical layout of the facility. The proper balance
between ventilation of the work area and decay of radiogases must be deter-
mined prior to release. A high ventilation rate too quickly after operation
would allow the discharge of unpermissible amounts of radioactive gases to the
atmosphere. In cases where there are adsorption filters, the reduced
residence time would affect their efficiency. While a comfort-index would
require a minimum of 2-3 air changes per hour, there is a maximum above
which the operational efficiency of sensitive instruments may be affected.
Several papers have been found in the literature regarding induced radio-
activity in accelerator operations at beam energies greater than 10 MeV.
Some of these references are to non-U.S. or Federally-owned machines and
thus outside the scope of this report. Their review here is for instruc-
tive purposes, especially when ventilation or exhaust conditions are
described.
-------
6-4
In the exhaust from the operation of a 20-40 MeV linear accelerator, the
150 and 13N air-activation components were determined to be less than 1% of
the Maximum Permissible Concentration* for airborne emissions. The accelerator
room was constantly ventilated and exhausted at a rate of 1.1 x 10 L/min
(40,000 ft3/min) from a 20-meter stack.^ ' Two other medium energy linac
facilities were monitored for the occurrence of 0 and ^1 off-site;
the concentrations were within the limits for nonocci'pational exposure.
At one linac, air supply and exhaust configurations are as follows: two
fans, operating in parallel, move target and accelerator room air to
the base of the 20-meter stack from an exhaust filter bank 3.7 meters off
the floor and forward of the target. Air treatment consists of a set of
prefilters, followed by a set of HEPA filters. The exhaust flow rate was
378,000 L/min (13,500 ft /min) providing 11 air changes per hour which is
almost half the design capacity. The author attributed this to dust loading
on the filters.
For the other linac described, the target is located in a caveroom, unventilated
but exhausted through the roof at ground level at a rate of 70,000 L/min
2
(2,500 ft /min) or 6 air changes pet
ports accounts for the make-up air.
13 15
A method described by Kase to predict N and 0 concentrations in air from
a 100 MeV linac is important to mention because it can be applied at the design
stage of similar facilities to determine ventilation criteria. For a known
beam length, for example, activity concentrations can be calculated for varying
room sizes and ventilation rates.' '
In the operation of a 550 MeV linear accelerator located underground, it was
found that concentrations of the radiogases from the stack exceeded the MPC ,
a
especially during operation. The author recommended confining the target room
in a^closed air circuit for a decay time and reducing the exhaust volume by
50%.""
2
(2,500 ft /min) or 6 air changes per hour; leakage from around doors and
* 10 CFR 20, App. B., Table II.
-------
6-5
The operations of the CERN 600 MeV synchrocyclotron and the 28 GeV synchrotron
were examined for gaseous emissions prompted by an anticipated 10-fold
increase in beam intensity. Modifications of the ventilation system,
increased stack height (to 25 meters), and the continued use of high efficiency
filters were indicated/ ^
The combined effects of decay and dilution necessary to reduce the concen-
tration of emissions from the stack of a 200 GeV machine tunnel have been
reported by Thomas/ 3 The activity of the air confined in the tunnel after
machine turn-off was calculated to be 360 curies. Assuming a leakage rate
of 10% per hour to the stack at a travel time of 6 minutes, Thomas determined
that as many as 4 curies per hour would be released to the atmosphere. Since
90% of the total specific activity is found close to the target and drops
by a factor of 10 after 10 minutes, the author recommends enclosure of the
target area and exhausting after 10 minutes only those areas to which access
is required.
Special problems arise with the use of tritium targets for neutron generators.
o
Gaseous tritium ( H), the radioactive isotope of hydrogen, has been detected
in the exhaust gas from a Cockcroft-Walton neutron generator used to bombard
a tritium target. It can be considered hazardous owing to its long half-life
and its low-energy beta radiation. In studies of tritium contamination in particle
accelerators, it has been shown that the tritium was readily absorbed in the
lucite insulators of the ionization chambers/ '
Several references mention ozone contamination of air during electron beam
operation. According to the American Conference of Governmental Industrial
Hygienists, ozone is a highly injurious and lethal gas with a Threshold
Limit Value of .1 ppm> ' One source indicates that hazardous levels of
ozone can be produced even at low energies. An example is given for a 1-mA
external beam of 10-MeV electrons with a path length of 300 cm which is con-
tained in a unventilated room 4 m by 4 m by 3 m. After an irradiation time of
about 15 minutes, the concentration of ozone was 15 ppm. The use of ducts
or hoods with air intakes located as near the electron-permeable window as
possible and the use of'high capacity blowers are recommended/ ' In air
-------
6-6
activation studies, Slabak confirmed the predominance of ozone production
in the 13 to 20 MeV range. Also, it was determined that the decomposition
rate of ozone is sufficiently long that its disappearance is governed pri-
marily by ventilation.' ' The effects of ventilation on radiogas and
ozone concentrations from 40 to 50 MeV linacs were studied in detail by
George and his coworkers, who found initial concentrations of ozone up
to .8 ppm after beam-off. With ventilation, reduction by a factor of 10
was possible after 15 to 20 minutes. In the absence of ventilation, the
reduction was only half over a period of 35 minutes. The use of a local
exhaust system (moveable hoods) that could be adapted to the target con-
(8)
figurations was suggested in this report/ '
6.2.2 Air Cleaning Devices
The air cleaning technology that exists for conventional processes can be
adapted to the treatment of gaseous effluents from accelerators with a few
basic exceptions:
1. The lowest threshold limit for chemical pollutants is two
orders of magnitude higher than the MPC of any radioactive
gas.
2. Airborne particulates, which are irradiated by adsorption of
or association with radioactive gases, ar>e generally in the
submicron range. Also, the dust loading or amount of par-
ti cul ate matter in accelerator exhaust is low due to the
comparatively clean conditions of these facilities. These
factors would limit the selection of filters to those with
a high removal efficiency for submicron particles.
Another departure from conventional processes is that the function of
adsorptive filters is simply to trap the gases for a decay-time; recovery
of regenerative techniques are of no interest. In fact, the replacement
of spent filters might present special problems if there is residual long-
lived activity on the filters.
-------
6-7
The Nuclear Air Cleaning Handbook, published by ERDA in 1976, is considered
a reliable guide for the design and installation of high efficiency air
cleaning systems for the nuclear power industry with some attention given
to radiochemical operations.^ Extensive research in this area has also
been the subject of biannual Air Cleaning Conferences, the latest of which
was held in August 1978.
In the absence of an authoritative guide specific to accelerators, these
sources can be useful in establishing the principal features of the mechanical
or hardware phase of an air cleaning system, but wide differences limit
their applicability. For example, accelerator users are primarily institutions
(hospitals, universities, processing plants) with only a portion of their
facilities restricted for these activities. In some cases, existing ven-
tilation and air conditioning configurations have to be modified to accommodate
the air flow to and from the accelerator room. The volume of air to be
treated and the concentration of contaminants are usually less than that of
most nuclear processing systems. Elaborate methods of trapping iodine and
krypton gases are not of primary concern and rigorous provisions for high
temperatures, moisture, and other emergency conditions are not necessarily
applicable.
The principal control methods for radioactive effluents can be discussed as
two distinct types, employed singly or in combination. The first type
consists of the ventilation controls that direct the contaminated air into
an exhaust system to a controlled release point. The second type of controls
remove contaminants from the exhaust stream prior to release either per-
manently, in the case of particulates, or temporarily to permit decay to
safe levels.
Pressure Zonation
This ventilation principle consists of providing sufficient pressure differen-
tials between the zoned-area and the rest of the building so that there is
no backflow of contaminated air to- spaces of uncontaminated air. Examples of
-------
6-8
this are hoods, wall fans, and vent stacks. The airflow induced by hoods
and wall fans permit dilution of radioactive effluents within the occupied
area of the facility, but the vent stack provides the most rapid dispersion
of effluents due to its elevated release point and high effluent velocity.
Filtration
The filtration of exhaust air prior to release is usually accomplished by
HEPA filters, preceded by pre-filters. In some cases, adsorptive
filters are necessary for the removal of gases. Examples of these are
activated charcoal and molecular sieves, but these are usually preceded
in line by a particulate filter.
HEPA filters, described as dry-type extended medium filters, consist of
pleated webs of fiber glass papers. Their particle removal efficiency
is 99.97% for .3 micron particles. The most frequently used HEPA filter
is rectangular and is designed for a nominal airflow of 28,000 L/min.
For larger air volumes in central systems, the filters can be arranged
5
in banks or in multiple single-units in series. Above 8.5 x 10 L/min
the bank systems are segmented into parallel systems. Self-enclosed
single-stage filters, or caissons, for in-duct installation are commercially
available. One type is designed with a bag-out feature, permitting the
replacement of filters with minimum personnel exposure.
Roughing filters are required in advance of the HEPA filters to protect them
from larger particulates. The extended-medium dry-type units that have
replaceable mediums or consist of throw-away cartridges are desirable.
Adsorptive filters of activated carbon are used to trap hazardous gases
and their reaction products from the air stream. Most conmonly used are
modular tray adsorbers, which are arranged in banks for multi-stage air
cleaning. Their dimensions correspond to standardized openings and con-
tain around 23 kg of carbon in two 5-cm beds separated by an air space.
-------
6-9
Each tray is designed for a gas residence time of .25 sec and an air flow
rate of 9424 L/min. Replacement of these filters is usually necessary
once a year. Also available are "caisson" housings, with the bag-out
filter changing feature. They can be used singly or arranged in groups
within a single housing. They are designed to accommodate not only the
carbon filters described above, but also standard pre-filters and HEPA
filters.
In the case of tritium contamination of air, a more elaborate system
of air cleaning is necessary. Tritium exhaust systems described for
neutron generation consist of a preheater, a catalytic recombiner, and a
regenerable molecular sieve bed. Although these components are commercially
available, one manufacturer was found to market a packaged adsorption unit,
complete with mechanical filters for dust removal. The desirable feature
of this device, it is claimed, is that the service life of the filters
is unaffected by high concentrations of tritium/ '
Pretreatment of Ventilated Air
A major portion of dust and particulate matter collected by filters actually
consists of atmospheric dust Introduced by supply air, or "infiltration,"
as opposed to internally-generated particulates. Installation of fiber
filters on the inlet side of ventilation air would reduce the irradiation
of ambient dust particles. Treatment of ducts to minimize corrosion and
:ri-! flaking is another measure that could be taken.
6.3 Costs and Effectiveness
The cost of air cleaning systems varies according to the size of the
facility, the ease of installation , and the degree of purification desired.
The initial expenditures include the following plus the cost of labor for
each:
• The major control equipment (hoods, filters, housings, mounting
frames, etc.)
-------
6-10
• Ancillary components (fans, blowers, duct work, electrical
connections, etc.)
Operating costs will vary with the volume of air, the pressure drop,
operating time, and power use by fans, motors and pumps. Maintenance
costs will consist mainly of the labor associated with filter replace-
ment, inspection and repairs, in addition to the costs of filters. For
example, the pressure drop obtained with prefilters of increasing efficiency
will affect the costs of power, installation and replacement, and should be
balanced against the costs related to frequency of HEPA filter exchange.
A wide variety of complete systems and components is readily available
from vendors. Table 6-1 shows cost estimates for air treatment systems
and components based on information obtained by Teknekron from various
sources. Except where noted, these estimates were based on a nominal air-
flow of 28,300 L/min, which is the standard sizing of most air filtration
equipment. These figures also represent modifications to existing ventilation
systems, so that the wide ranges in estimates are accounted for by varying
site-specific installation factors. For comparison, several cost figrues
were obtained from the literature for building ventilation exhaust air
cleaning systems in the nuclear industry.
Installation figures include the price of ancillary components, such as
duct work, blowers, wiring, and associated labor costs. The annual operating
costs for HEPA and HEPA/Activated carbon systems include the replacement
costs of the filters. The activated carbon filters are usually replaced
once a year at a cost of about $330 per 28,300 L/min for an unimpregnated
nuclear grade carbon/ 6' Replacement frequency for HEPA filters is
partially determined by the dust-holding capacity, and for design purposes
is considered to be 1809 g for the standard-sized unit.
In the tritium clean-up system, described in reference 36, the annual
operating cost includes the processing and shipment off-site of the spent
molecular sieve beds.
-------
TABLE 6-1
COST OF AIR TREATMENT CONTROLS (1978 DOLLARS)
SYSTEM
Wall Fan
Hood
Vent Stack
(2-3M)
HEPA Filter
Prefilter
Housing, etc.
Vent Stack/HEPA
HEPA/Activated
Carbon
Tritium Absorption
(56.6 L/min.)
Nukem Tritium
Absorber
(200 L/hr holding
tanks)
EQUIPMENT INSTALLATION
$300
2000
10,000
150
75
1500
3,500
6,000
**(9,
***121
120
- $1000 +
- 4000 $6,000 - $20,000
- 60,000 +
3,000 - 10,000
*(4,900 - 14,160)
- 10,000 10,000 - 60,000
- 12,000 +
700)
,000 +
,000 120,000
ANNUAL OPERATING AND MAINTENANCE
$30 - $300
+
+
300 - 500
+
700 - 1400
**(1200)
***(36,300)
+
*Source, Reference 23, p. 251. 1978 dollars obtained by multiplying 1975 costs by 1.211 based on the
Consumer Price Index Listing.
**Source, Reference 25, 1978 dollars obtained by multiplying 1975 dollars by 1.211 as above.
***Source, Reference 36.
+Variable and/or not available
cr>
-------
6-12
In summary, the existing control technology for radioactive effluents consists
of any one or combinations of the following: confinement, dilution, and
removal of contaminants by filtration. Costs vary widely according to the
air flow to be treated, the desired efficiency, and the modifications entailed
for existing ventilation systems.
-------
7-1
7.0 GENERIC FACILITIES
In this section, the information obtained is generalized to the various
classes of accelerators (constant field, incremental acceleration-, and
magnetic field). It is recognized that machines can vary greatly within
a given class. For instance, a high energy research cyclotron is very
different from a high current, isotope production cyclotron. Nevertheless,
some generalizations can be made.
7-1 Characteristics
Within the constant potential classification, '/an de Graaff accelerators are
the most prevalent. As shown in Section 3.2, about 18% of all registered
machines in 1977 belonged in this category. Furthermore, 78% of the Van
de Graaffs were in the energy range of 1 to 19 MeV. This suggests an
intensity ranging from 1 uA to 1 mA (see Figure 7-1). Therefore, the Van
de Graaff monitored in this study and discussed in Section 5.2 (6 MeV, 1 to
10 yA) can be considered a generic constant field machine.
The most popular accelerator currently in use belongs in the second
class, incremental acceleration or cyclic machines. This accelerator,
the linac, represents about half of the accelerators in the United States.
About 90% of these are electron linacs in the energy range of 1 to 19 MeV
and are between 1 and 10 kilowatts in power. These characteristics are
typical of the cancer therapy machines widely used in U.S. hospitals. For
this reason, we have selected an 18 MeV medical linac as a generic cyclic
machine.
As discussed in Section 2.0, the betatron machine is in a class of its own.
However, its radiation characteristics and even its applications are similar
to the electron linac described above. The calculated production of air-
borne radioactivity by the generic linac can also be applied to the betatron,
-------
7-2
- LOW VOLTAGE TRANSFORMERS
2-ELECTROMAGNETIC SYSTEMS
3-VAN DEGRAAFFS
4- TANDEM ACCELERATORS
5-BETATRONS
6-ELECTRON UNACS
7-ION LINACS
8-CYCLOTRONS-
9-ELECTRON SYNCHROTRONS
O-PROTON SYNCHROTRONS
10 10" IOJ 10"
Porticle Energy (MeV)
FIGURE 7-1
Particle-beam intensity versus particle energy, for several
types of accelerators. (Dark-toned areas relate to accelerators
both electrons and ions; middle tones relate to ion accelerators
only; light tones relate to electron accelerators only.)
-------
7-3
and therefore the magnetic field machine will not be discussed generically.
Instead, we have chosen to discuss another cyclic machine, the cyclotron.
Our survey shows that these machines are generally higher in energy and
higher in power^than other machines in use, thus potentially producing
greater quantities of airborne radioactivity.
7.2 Release Estimates
The following release estimates for the generic facilities are developed
from the field studies described in Section 5 and from calculations dis-
cussed in Section 4.
7.2.1 Constant Field Accelerator
The generic constant field machine delivers a beam in the energy range of
1 to 19 MeV, which is, for the most part, below the threshold energies for
air activation shown in Table 4-1. The machine produces a small thermal
neutron flux,* but this flux is not large enough to produce a measurable
amount of activation product.
The field study has shown, however, that measurable quantities of tritium
can be released from a facility using a gaseous tritium target. The litera-
ture indicates that this is also true of facilities using solid tritium/
titanium targets. Our data show that an annual release on the order of
1 curie of tritium is possible due to target maintenance work. Releases due
to tritium escaping from the target and beam tube are dependent on the
facility design and operation but can be as high as several curies annually.
Therefore, the annual release of tritium from the generic Van de Graaff is
estimated at about 1 curie. There are about 100 ion Van de Graaff machines
in the United States.
*The thermal neutron flux at the study Van de Graaff was measured to be
about 50 n/crn2/sec.
-------
7-4
7.2.2 Electron Linac
This machine was selected because of its frequent use in medicine. The great
majority of linacs fall into the energy range from 1 to 19 MeV. However,
machines of this energy are not capable of producing significant airborne
radioactivity for the reasons discussed immediately above. Nevertheless, cal-
41 14
culations were made of Ar and C production rates based on available data.
Given the thermal neutron flux around an 18 MeV linac obtained from the litera-
41 -3
ture, annual releases of 100 yCi Ar and 10 j
are about 440 such linacs in the United States.
41 -3 14
ture, annual releases of 100 yCi Ar and 10 yCi C were estimated. There
7.2.3 Cyclotron
The release estimates for a generic cyclotron apply to a high energy, low
current machine, such as the one monitored as part of this study (see Section
5.2). The acceleration of ions at about 100 MeV into scattering targets
produced secondary radiation in the form of hard gamma rays (>10 MeV). These
13
gammas were presumably responsible for producing the N (via the (y>n) reaction)
15 11
detected in our study. The absence of 0 and C indicates that the gamma fluxes
probably dropped off markedly at the higher energies necessary to produce these
nuclides. This, in turn, suggests that lower energy cyclotrons would probably
fail to produce even N.
In the case of neutron beam production at a cyclotron, relatively large amounts
of 0, N, and C are produced via the (n,2n) reaction. During the 4-hour
1C 10
run with 100 MeV neutrons observed in our study, 0.23 Ci 0, 342 yCi N,
and 457 yCi C were released to the atmosphere. It is difficult to estimate
the annual utilization of a cyclotron for neutron generation except to say that
increased research is being performed in this area for medical applications.
Assuming, say, 20 hours of neutron beam time, and adding 38 mCi N produced as
described in Section 5.3, the following annual releases are calculated: 1 Ci
0, 40 mCi N, and 2 mCi C. There are approximately 12 such machines in
the United States.
-------
7-5
7.3 Dispersion Estimates
Gaseous effluents from a vent or a stack undergo dispersion in ambient air upon
release. Estimates of effluent dispersion for a typical accelerator facility
were made based on annual average conditions by use of the following expres-
sion:^ '
c(x,e) = 2-
u(h + xR (x))
where: C(x,e) = Effluent concentration downwind of the facility (Ci/m3)
Q = Emission rate (Ci/sec)
u = Average wind speed (m/sec)
h = Building height (m)
x = Downwind distance (m)
R2 = Plume vertical half width (m)
f(0) = Frequency of the wind toward one of the 16 directional
sectors, 0
The normalized concentrations (C/Q) in sec/m3 calculated for a typical facility
are presented in Table 7-1. The average wind speed for this site is 5.6 m/sec
and the building height is 16.8 m.* The calculations also assume the climatic
condition of D-stability since this condition, which occurs roughly half the
time, is associated with the building downwash effect. For this reason, the
calculated annual concentrations represent conservative estimates.
To estimate dispersion of short-term effluent releases, the building downwash
effect must be analysed. Plume downwash into the building wake will occur if
the height of the release point above the ground is less than 1.5 times the
height of the building. Since this is the condition found at most accelerator
facilities, plume downwash can be expected to occur. Furthermore, the highest
concentrations of released radioisotopes are predicted to occur in the building
wake due to the downwash effect. A conservative estimate of the ground level
*These values correspond to the actual values associated with the cyclotron
facility discussed in Section 5.2.
-------
7-6
(29)
effluent concentration within the wake is given by the following expression: '
At greater distances downwind of the facility, the emissions may be treated
as coming from a ground level source with an initial plume cross sectional area
of h2:
C(x) =
u(IT + 2RyRz)
The half widths R , R are equal to /IT/2 times the more conventional horizontal
and vertical dispersion, coefficients ay and o^. The normalized concentrations,
Cu/Q, out to 5 kilometers are given in Table 7-2 for the typical facility described
earlier. These concentrations represent averaging times in the range of 30 minutes
to 1 hour.
7.4 Demography
The survey of accelerators (see Section 3.4) revealed that 76% were located
in urban areas. We have selected three actual accelerator facilities located
in urban areas as a basis for developing the demography for the generic facility.
Two of the facilities were monitored in this study; the third is an isotope-
producing accelerator facility located in an urban area.
Demographic data for the generic facility are based on an average of the 1970
census data for the facilities discussed above. The distances for the popu-
lation count are measured from the facility. The AREA POP computer code at
the U.S. EPA, Office of Radiation Programs (Las Vegas Facility), was run to
obtain the population distribution. The number of people within the annular
rings of 0-1 km, 1-2 km, 2-3 km, 3-4 km, and 4-5 km has been computed
for each of the three facilities and then averaged to obtain the population
distribution for the generic facility.
-------
TABLE 7-1 NORMALIZED CONCENTRATIONS, C/Q (sec/m3)
DOWNWIND OF AN ACCELERATOR FACILITY
Distance Downwind (km)
ECTOR
N
NNE
NE
ENE
E
ESE
SE
SSE
S
ssw
sw
WSW
w
WNW
NW
NNU
0
9.9xlO"5
g.ixio"5
9.8xlO"5
8.6xlO"5
1.4xlO~4
2.1xlO~4
1.6xlO"4
8.6x10-^
7.4xlO"5
5.6xlO"5
9.9xlO"5
l.lxlO'4
8.8xlO"5
4.6xlO"5
7.0xlO~5
7.5xlO'5
1
6.2x10~7
5.7xlO"7
6.1xlO"7
5.4xlO"7
9.0xlO"7
1.3xlO"6
l.OxlO"6
5.4xlO"7
4.6xlO"7
3.5xlO"7
6.2xlO"7
7.0xlO"7
5.5xlO"7
2.9xlO"7
4.4xlO"7
4.7xlO"7
2
2.0xlO"7
1.8xlO"7
2.0xlO"7
1.7xlO"7
2.9xlO'7
4.3xlO"7
3.2xlO"7
1.7xlO"7
1.5xlO"7
l.lxlO'7
2.0xlO"7
3.0xlO"7
1.7xlO"7
9.3xlO"7
1.4xlO"7
1.5xlO"7
3
l.lxlO"7
9.7xlO"8
l.OxlO"7
9.1xlO"8
l.BxlO"7
2.2xlO"7
1.7xlO"7
9.1xlO"8
7.9xlO"8
6.0xlO"8
l.lxlO"7
1.2xlO"7
9.2xlO"8
4.9xlO"8
7.4xlO"8
a.oxio"8
4
7.1xlO"8
6.3xlO"8
6.5xlO"8
5.9xlO"8
9.7xlO"8
1.4xlO"7
l.lxlO"7
5.9xlO"8
5.1xlO"8
3.9xlO"8
7.1xlO"8
7.8xlO"8
6.0xlO"8
3.2xlO"8
4.8xlO"8
5.2xlO"8
5
4.8xlO"8
4.3xlO"8
4.4xlO"8
4.0xlO"8
6.6xlO"8
9.5xlO"8
7.5xlO"8
4.0xlO"8
3.5xlO"8
2.7xlO"8
4.8xlO"8
5.4xlO"8
4.0xlO"8
2.2xlO"8
3.3xlO"8
3.6xlO"8
I
•--J
-------
TABLE 7-2 NORMALIZED CONCENTRATION AS A FUNCTION OF DISTANCE
FOR SHORT TERM RELEASES
SITE
Accelerator
Facility
NORMALIZED CONCENTRATION,
0 km
5.3xlO"3
1 km
l.lxlO"4
2 km
3.8xlO"5
Cu fnf2!
Q (m J
3 km
2.1xlO"5
4 km
1.4xlO"5
5 km
l.OxlO"5
00
-------
7-9
The AREA POP code provided data on the number of census enumeration districts
(CEDS) that were searched to obtain the population distribution at selected
radii from a point of interest. The CEDS are used as a measure of the popu-
lation density and expected accuracy of the data. CEDS of at least 6 are
needed to assure good confidence in the accuracy of the population data. The
data presented in Figure 7-2 are based on CEDS that range from 5 to 104.
Although the generic machine is in an urban area, it is located on a university
campus which is not a high-density location. Typically, it would be adjacent
to offices, and the nearest residence (or dormitory) would be at least a
kilometer away. Land use would be nonagricultural within the city limits,
and partly agricultural and partly residential beyond that point.
-------
7-10
5KM
FIGURE 7-2
POPULATION DISTRIBUTION AROUND A GENERIC ACCELERATOR FACILITY
-------
8-1
8.0 CONCLUSIONS
The following conclusions have been drawn from the results of this study:
1. Approximately 1200 accelerators were in operation in the United States
in 1977 excluding those under Federal control. The growth rate of
accelerator use is put at 65 machines per year.
2. The linear accelerator (linac) is the most commonly used machine in the
United States. More than half of the machines in use are linacs, and 7035
of these linacs are used in medical applications (i.e., cancer therapy) with
energies ranging from 1 to 19 MeV.
3. With respect to airborne effluents, machines operating at less than
100 MeV and electron linacs operating at less than 50 MeV produce negligible
amounts of airborne radioactivity. Machines operating above these energies
15 13
produce sufficiently large fluxes of hard gamma rays to produce 0, N, and
C by the (y»n) reaction. The exceptions to this are machines used to
produce fast neutron fluxes (above 20 MeV) which create the above nuclides by
the (n,2n) reaction.
4. The annual quantities of 0, N, and C released to the atmosphere by
a typical 100 MeV cyclotron are estimated to be 1 Ci, 40 mCi, and 2 mCi,
respectively. There are approximately 12 such machines in the United States.
5. Releases of airborne tritium were found at an accelerator facility
utilizing gaseous tritium targets for neturon production. Annual release
of tritium from a typical facility of this type is put at about 1 Ci.
6. The control of airborne radioactivity around accelerators has two
objectives. One objective is to minimize exposure to personnel working
in the accelerator cavity. The second is to minimize releases of radioactivity
-------
8-2
to the atmosphere. These are somewhat in conflict since high ventilation
rates benefit workers desiring quick access to the machine after shutdown
but tend to increase the activity released to the atmosphere. To achieve
both of these objectives, the following actions are possible:
a. delay discharge of cavity air (and access to the machine) for a
sufficient time after shutdown to allow radioactivity to decay
to innocuous levels,
b. use recycle type air handling systems with isolation capability
for hold-up and for accident situations, and/or
c. use air cleaning systems for effluents or recycled air based on
cost/benefit considerations.
-------
9.0 REFERENCES
Radioloafcal°LRS0l0gical Health' Report of State and Local
^^-^^l-^lliL^
Rolo1cal Hea^h. The Use of Electron Linear
Therapy, Overview Report No. 2.
3' T°m McRae' Varian ^oration,
3a. Cohen, S.C., et al_. Evaluation of Occupational Hazards from Industrial
Radiation: A Survey of Selected States. U.S. Department of Health
Education and Welfare, Publication No. (NIOSH) 77-142, December 1976.
4. Grey, D.E. (ed.). American Institute of Physics Handbook. McGrawHill, 1957.
5. Grey, D.E. (ed.). American Institute of Physics Handbook. McGrawHill, 1953,
6. Patterson, H.W., and R.H. Thomas. Accelerator Health Physics.
New York: Academic Press, 1973.
7. Barbier, M. Induced Radioactivity. New York: John Wiley and Sons,
Inc., 1969.
8. George, A.C., et_ aJL "Evaluation of the Hazard from Radioactive Gas
and Ozone at Linear Electron Accelerators." Proc. U.S. A EC 1st Symp.
Accelerator Radiation Dosimetry and Experience (CONF-6511 ), 1965.
9. Slaback, Lester A. "Health Physics Aspects of 20- to 40-MeV Linac
Operations." Armed Forces Radiobiological Research Institute,
Bethesda, Maryland (to be published).
10. Nellis, Donald 0., et_ aj_. Tritium Contamination in Particle Accelerator
Operation (NTIS PB-189-362), November 1967.
11. Peetermans, A., and J. Baarli. "Radioactive Gas and Aerosol Production
by the CERN High-Energy Accelerators and Evaluation of Their Influence
on Environmental Problems." Proceedings of the Environmental
Surveillance around Nuclear Installations Symposium, Warsaw, Poland,
Nov. 5-9, 1973 (IAEA-SM-180/10) , p. 433.
12. Hoefert, M. "Radiation Hazard of Induced Activity as Produced by
High-Energy Accelerators." Proceedings of the International Conference
on Accelerator Dosimetry Experiments, 2nd 111-20, 1970.
13. Engelke, M.J., and H.I. Israel. "Monitor for Radioactive Gas in the
LAMPF (Los Alamos Meson Physics Facility) Accelerator Beam Channel"
(LA-5351-MS). Los Alamos, New Mexico: Los Alamos Scientific Laboratory,
1973.
-------
9-2
14. Vialettes, H. "Gas and Dust Activation in the Target Room of the Saclay
Electron Linac; Identification of the Produced Radioactive Nuclei and
Determination of the Rejected Activities." Proceedings of the International
Conference on Accelerator Dosimetry Experiments, 2nd 121-38, 1970.
15. National Bureau of Standards. Annual Report - 1975, Health Physics
Section. Office of the Associate Director for Administration.
16. McGinley, P.H., et_ a\_. "Dose Levels Due to Neutrons in the Vicinity
of High-Energy Medical Accelerators." Medical Physics, Vol. 3, No. 6,
Nov/Dec 1976.
17. Deye, J.A., and F.C. Young. "Neutron Production from a 10 MeV Medical
Linac." Physics Medical Biology, Vol. 22, No. 1, 1977, p. 90.
18. Lane, Richard G., ejt al_. "Leakage Radiation Characteristics of an
18 MeV Clinical Linear Accelerator." Health Physics, Vol. 35,
1978, p. 485.
19. Kase, K.R. "Radioactive Gas Production at a 100 MeV Electron Linac
Facility." Health Physics. Vol. 13, 1967, p. 869.
20. U.S. Atomic Energy Commission. Standards for Protection Against
Radiation (10 CFR 20), 1965.
21. Conference of Radiation Control Program Directors, U.S. Atomic Energy
Commission and U.S. Department of Health, Education and Welfare.
Suggested State Regulations for Control of Radiation, 1974.
22. Thomas, R.H. "Rough Estimates of Radiation Hazard from Radioactive
Gas in a Machine Tunnel." Lawrence Radiation Laboratory, Internal
Report 10136, 1964.
23. Burchstead, C.A., A.B. Fuller, J.E. Kahn. Nuclear Air Cleaning
Handbook (ERDA 76-21). Oak Ridge National Laboratory, 1976.
24. U.S. Nuclear Regulatory Commission. Proceedings of the Fifteenth Air
Cleaning Conference, Boston, Mass., August 1978.
25. U.S. Nuclear Regulatory Commision. Guide 1.110: Cost Benefit Analysis
for Radwaste Systems for Light-Hater-Cooled Nuclear Power Reactorst
March 1976.
26. Barnebey, Cheney. "Activated Carbon, Purification and Recovery Treatment"
(trade literature).
27. Nukem, Gmbh. "Tritium Technology," N-753-4.1 (trade literature).
28. Eicholz, G.G. (ed.) Radioisotope Engineering. New York: Marcel Dekker,
Inc., 1972.
-------
9-3
30. Silvester, D.J., "Accelerator Production of Medically Useful Radio-
nuclides," 1AEA-SM-171/6, 1974.
31. Myers, W.G., et al_. "123I for Applications in Diagnosis," 1AEA-SM-
i /1 / ^T" 9 i y / T- •
32. National Bureau of Standards. Radiological Safety in the Design and
Operation of Particle Accelerators. NBS Standard 107, 1969.
33. Brobeck, W.M. and Associates. Particle Accelerator Safety Manual.
U.S. Department of Health, Education and Welfare, MORPH68-12, October
1968.
34. U.S. Atomic Energy Commission. Safety Guidelines for High Energy
Accelerators , 1967.
35. National Council on Radiation Protection. Radiation Protection Guidelines
for 1-100 MeV Particle Accelerator Facilities. NCRP Report No. 51, 1977.
36. U.S. Energy Research and Development Administration. Environmental
Statement, Brookhaven National Laboratory. ERDA-1540, November 1975.
-------
A-l
APPENDIX A
The following tables and figures present the data collected during the course
of this study. Table A-l tabulates the Ge(Li) gamma spectra taken at the two
facilities. The annihilation gamma peak was the only one detected. Facility
operation data are also listed in the table. Figure A-l presents the gamma
ray spectrum of the February 6th activated charcoal sample (#129).
Table A-2 presents the results of the tritium analyses, and Table A-3 presents
the results of the C analyses performed on selected molecular sieve (NaX)
samples.
-------
TABLE A-l
Ge(Li) SPECTROMETRY DATA
ID*
DATE
SAMPLING TIME
SAMPLE (hrs:min)
DECAY TIME
(min)
COUNT TIME
(sees)
PEAKS IDENTIFIED
(other than background)
Cyclotron
93
94
95
96
97
98
99
100
101
102
103
105
106
107
108
109
110
1/27
1/27
1/27
1/27
1/27
1/27
1/27
1/27
1/27
1/27
1/27
1/30
1/30
1/30
1/30
1/30
1/30
NaX-A
NaX-B
NaX-B
HEPA
Gas-1
Gas-1
Gas-2
Gas-2
Charcoal-A
Charcoal-A
Charcoal-B
NaX-A
NaX-A
HEPA
Charcoal-A
Charcoal-A
HEPA
1:0
1:0
1:0
1:0
n/a
n/a
n/a
n/a
2:59
2:59
2:59
2:12
2:12
2:12
2:28
2:28
2:28
10
17
23
29
1
6
0
10
6
16
25
2
7
12
2.5
7.5
14
300
300
300
300
300
300
600
600
600
400
600
300
300
300
300
300
300
.511 MeV
.511 MeV
.511 MeV
.511 MeV
COMMENTS
120 MeV alpha beam at
.02 uA (at cobalt target)
Trace peak - decayed with
10-15 min half-life 13N £
10-9 pCi/cc i 50%*
65 MeV proton beam at .02
uA (at spectrometer)
Trace peak - decayed with
5-10 min half-life 13N £
10-9 pCi/cc + 100%
-------
TABLE A-l (Continued)
Ge(L1) SPECTROMETRY DATA
SAMPLING TIME
ID*
111
112
113
114
115
116
117
119
120
121
122
123
124
125
126
127
128
129
130
131
132
DATE
1/30
1/30
1/31
1/31
1/31
1/31
1/31
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
2/6
SAMPLE
Background
HEPA
Charcoal-A
Charcoal-A
NaX-A
Gas
Background
NaX-A
NaX-A
NaX-A
NaX-A
Background
Gas-1
Gas-1
Gas-2
Gas-2
Gas-2
Charcoal-A
Charcoal-A
Charcoal-A
Charcoal-A
(hrs:m1n)
n/a
2:12
17:32
17:32
17:49
n/a
n/a
0:58
0:58
0:58
0:58
n/a
n/a
n/a
n/a
n/a
n/a
2:57
2:57
2:57
2:57
DECAY TIME COUNT TIME
(min)
n/a
63
2
7
3
.5
n/a
2
7
13
19
n/a
1
6
0
6
12
2
7
12
17
(sees)
10,000
10,000
300
300
300
600
n/a
300
300
300
300
1,000
300
300
300
300
300
300
300
300
300
PEAKS IDENTIFIED
(other than
.511
.511
.511
.511
.511
.511
.511
.511
.511
.511
.511
.511
background)
MeV
MeV
MeV
MeV
MeV
MeV
MeV
MeV
MeV
MeV
MeV
MeV
COMMENTS
Trace peak - decayed with
5-10 min half-life 13u %
10-9 uCi/cc + 100%
100 MeV proton beam at 1.9
uA with Be/Al target produc-
ing high energy neutrons,
about 2x1 Ol 2 n/sec at 100 feV
(see Figure 5-5 )
"C cone. =5x10-8 MCi/cc**
Decayed with 96 sec half-
life - 150 (see Figure 5-4)
Cone. =2x10-6 pC1/cc+
Beam off during sampling
Decayed with 10 min half-
life - 13N (see Figure 5-4 )
Cone. =3x10*8 yCi/cc
I
CO
-------
TABLE A-l (Continued)
Ge(L1) SPECTROMETRY DATA
ID*
133
14
16
17
18
19
58
59
60
61
DATE
2/6
2/22
2/22
2/22
2/22
2/22
3/6
3/6
3/6
3/7
3/7
3/7
3/7
3/7
3/13
SAMPLING TIME DECAY TIME COUNT TIME PEAKS IDENTIFIED
SAMPLE (hrs:min) (min) (sees) (other than background) COMMENTS
Charcoal-B
NaX-A
NaX-A
HEPA
Charcoal-A
HEPA
Charcoal-A
(cavity exh. ',
Charcoal-A
HEPA
NaX-A
(cavity exh. ,
HEPA
Charcoal-A
(vacuum exh. ,
HEPA
Background
Charcoal-A
(vacuum exh. ,'
2:57
1:37
1:37
1:37
2:38
2:38
Van
1:33
1:33
1:33
17:20
17:20
3:57
)
3:57
n/a
1:37
)
22
2
7
13
2
8
de Graaff
2
7
24
5
10
2
15
n/a
2
300
300
300
300
300
300
300
300
300
300
300
300
1000
10,000
500
.511 MeV Assume 13N
80 MeV alpha beam at .2
4.7 MeV proton beam at 1
uA with H-3 target pro-
ducing 4 MeV neutrons
uA
.7
No beam - target download-
ing
3 MeV deuteron beam at
3 uA with deuterium target
producing 6 MeV neutrons
beam off 3/13, 2400 hrs
-------
TABLE A-l (Continued)
Ge(Li) SPECTROMETRY DATA
SAMPLING TIME DECAY TIME COUNT TIME PEAKS IDENTIFIED
ID# DATE SAMPLE (hrs:min) (min) (sees) (other than background) COMMENTS
62
65
3/13
3/14
HEPA
NaX-A
1:37
24:28
12
1
500
500
(cavity exh.)
66 3/14 HEPA 24:28 10 500
Sample Calculations:
*r =
.f /CpS\ m-Ti /j_4- _.r.c \ .. ni / -i i _ _r_r \ .. r- <->_i^O
600
cps) x 12.4 (decay correction)/3.7xlO^ (^4) x .0171 (det. eff.) x .83 (coll. eff.) x 5.2xl06 (cc)
= 1.1 x 10"9 yCi/cc
(decay corr. = Xt/l-e"Xt = .00116(s"1) x 10,740(s)/l-10"6 = 12.4
**C = |^ (cps) x 2.3 (decay corr. )/3.7xl04 x .0171 x .08 x 1.64 x 106 (cc)
= 5.4 x 10"8 yCi/cc
+C = |§! (cps)/3.7x!04 x .0048 (det. eff.) x 2000 (cc) = 1.9xlO"6 yCi/cc
en
-------
1000
100 ~
O
0.
oo
10
.511 MeV, N-13
1.46 MeV, K-40
! i
ENERGY
FIGURE A-l
4096 CHANNEL GAMMA SPECTRUM, #129, 300 sec.
-------
TABLE A-2
RESULTS OF TRITIUM ANALYSES
Satnpl ing Time
Sample No.
•3
O
10
14
20
3
5
6
7
10
11
Date
1/27
1/30-31
2/6-7
2/22
3/6-7
3/7-8
3/7-8
3/7-8
3/13-14
3/13-14
Location
Main Exhaust
Main Exhaust
Main Exhaust
Main Exhaust
Cavity Exhaust
Cavity Exhaust
Hood Exhaust
(hrsrmin)
Cyclotron
3:00
20:00
16:25
2:37
Van de Graaff
16:04
26:37
26:12
Vacuum Pump Exhaust 23:19
Cavity Exhaust
Cavity
24:28
26:15
Tritium Concentration
pCi/cc
<7-10
<1 -10
<8-10
-9
-9
-9
-9
6-10
-8
6-10
"7
1.5-10
6-10
-8
-8
Ventilation Flow
cc/min
4.8-10
4.8-10
4.8-10
4.8-10
1.1-10
1.1-10
1.7-10'
,8(*)
8(*)
1.1-10
2.2-10y
3(*)
NOTE: +Saturation of silica gel with water vapor (HTO) occurred reducing collection efficiency to less than 100%.
*This represents the return air flow of a closed-cycle ventilation system. ^
•»
-------
TABLE A-3
RESULTS OF CARBON-14 ANALYSES
Faci1i ty
Cyclotron
Cyclotron
Van de Graaff
Van de Graaff
Control Sample
(Blank)
Date
1/30
2/6
3/6
3/13
-
Sample Time
(hrs:min)
2:12
16:13
16:04
24:28
-
Vol ume
C02 (cc)
c. ^~^^
1.0
22.5
18.1
7.03
14.4
C-14
pCi/£ C00
<69
<9
320+32
<57
<14
Concentration
pCi/m3 air
0+145
0+2
100+10
0+12
0+4
I
c»
-------
B-l
APPENDIX B
Sampling Techniques
The absorber sampling was performed using three different types of media.
These included sodium zeolite for trapping C02> charcoal for gases N2 and 02>
and silica gel for tritiated HLO.
The charcoal and the silica gel samplers were run in a sample train as
shown in Figure 5-3. The charcoal sampler consisted of five media cups in
a cylindrical body with two end caps. "0" rincs served as seals between the
cups and end caps, forcing all flow through the sample media. Sample media,
in direction of flow, included a blank cup, particulate filter, primary
charcoal, back-up charcoal, blank cup. The sampling rate for this canister
was -28 L/m. After passing through the charcoal, a side stream flow was
passed through a rotameter, followed by a tritium sample cartridge. The
tritium sampler consisted of a 30 cm pipe nipple, two end caps with appropriate
tube fittings, silica gel (Grade 05,6-16 mesh), filters, snap rings, screens,
and tell-tale desiccant (Grade 42, 6-16 mesh). The sampling rate through the
tritium sampler was -7 L/m. The sampling probe consisted of a 9.5 mm
aluminum tube bent in such a way as to facilitate isokinetic air sampling.
An identical type probe was used for sampling air through the sodium zeolite
cartridge. The sampler was the same type as used in the charcoal sampler
described above. Sample media, in direction of flow, included a blank cup,
particulate filter, primary sodium zeolite (NaX), back-up NaX, blank cup.
Air was drawn through this sampler at a rate of = 28 L/m.
The duration of sampling was in the range of 1 to 2 hours. Because of the
short-lived activities of O15, N13, and C , an attempt was made to begin
the gamma analysis as soon as possible after sampling. The delay was nor-
mally less than 2 minutes.
-------
B-2
One or a mixture of the three isotopes mentioned above was indicated by
presence of the annihilation peak at 511 KeV. The sample was recounted
sequentially for 5 minutes with each spectrum being stored on magnetic tape.
This procedure was continued until the 511 KeV peak returned to background.
Subsequent analyses of each spectrum was used to obtain a decay curve and
thus to determine the composition of the isotopes in the air sampled.
The tritium samples were returned to the NES laboratory for processing.
The dry silica gel had been weighed before being loaded into the sampler.
The sampling duration and flow rate were recorded in a log. The outlet
end of the samplers were opened and the snap ring, filter, and screen were
removed to check the condition of .the tell-tale indicator, which was noted
in the sample log book. The dry, blue indicator turns through pink to white
as it becomes saturated with HLO. The contents of the sampler were then
dumped into a wide mouth 4-liter bottle for mixing to obtain a homogeneous
sample for analysis. A 3 gram (+_ 0.1 g) sample of silica gel was
weighed out into each of three liquid scintillation vials. Two milliliters
(jHO.l ml) distilled water and ten milliliters (j^ 0.1 ml) of scintillation
cocktail solution were added to each. The vials were capped and shaken
adsorbing the tritium from the silica gel into the scintillation cocktail.
The samples were then allowed to stand for 4 hours to assure equilibrium
between silica gel and cocktail solution. The samples were then counted
on a liquid scintillation counter. An NBS tritium standard and a blank
were also counted with the samples, the blank being taken as the background
tritium activity.
The second type of sampling that was performed was the direct extraction of
a gas sample into an evacuated glass marinelli container. This sampler has a
cylinder in the middle which fits over the Ge(Li) gamma ray detector. This
geometry optimizes the efficiency of analyzing a gas sample. To collect a
gas sample the container was simply evacuated with a vacuum pump and the
stopcock was then closed. The sampler was then connected to a probe in the
air duct to be sampled and the stopcock opened drawing air into the container.
The stopcock was again closed and the container quickly moved to the gamma-ray
detector for counting.
-------
B-3
Flow measurements in the air ducts sampled were made using a pitot tube.
Air velocity was measured at several penetration depths and an average was
taken. The velocity was .multiplied by the cross-sectional area of the duct
to obtain the flow rate.
U.S GOVERNMENT PRINTING OFFICE. 1979 620-007/3799
------- |