You are on page 1of 535

fourth edition

Physics of
Diagnostic Radiology


Professor of Radiology, University of Texas
Southwestern Medical Center at Dallas
and Parkland Memorial Hospital


Associate Professor of Radiology (Physics),
Southwestern Medical Center at Dallas
and Parkland Memorial Hospital


Associate Professor of Radiology (Physics)
Southwestern Medical Center at Dallas
and Parkland Memorial Hospital

4th Edition

Williams & Wilkins


' 1lliams & Wilkins
=lose Tree Corporate Center, Building II
400 North Providence Road, Suite 5025
Media, PA 19063-2043 USA

Library of Congress Cataloging-in-Publication Data

Curry, Thomas S., 1935-
Christensen's physics of diagnostic radiology.-4th ed. I
Thomas S. Curry III, James E. Dowdey, Robert C. Murry, Jr.
p. em.
Rev. ed. of: Christensen's introduction to the physics of
diagnostic radiology. 3rd ed. I Thomas S. Curry III, James E.
Dowdey, Robert C. Murry, Jr. 1984.
Includes bibliographical references.
ISBN 0-8121-1310-1
l. Diagnosis, Radioscopic . 2. Medical physics. I. Dowdey, James E .
II. Murry, Robert C . III. Christensen, Edward E., 1929-
Introduction to the physics of diagnostic radiology. IV. Title.
V. Title: Physics of diagnostic radiology.
[DNLM: I. Physics. 2. Radiography. 3. Radiology. 4. Technology,
Radiologic. WN 110 C976c]
RC78.C87 1990
for Library of Congress 90-5586

I st Edition, 1972
Reprinted 1973, 1975, 1976, 1977
2nd Edition, 1978
Reprinted 1979, 1981, 1982, 1983
3rd Edition, 1984
Reprinted 1985, 1987
4th Edition, 1990

Copyright © 1990 by Lea & Febiger. Copyright under the International Copyright Union.
All Rights Reserved. This book is protected by copyright. No part of it may be reproduced
in any manner or by any means, without written permission from the publisher.


Print Number: 12 11 10 9 8 7
Dedicated to: --




Why a new edition of a physics book? the amazing technology available, and let
Physics does not change. But our under­ the physicist worry about the technical de­
standing and use of physics continues to tails.
develop. In our quest to include information that
The competent radiologist must know should be known, and reduce the amount
something about the physics of diagnostic that need not be known, we have in this
imaging. The tough question is exactly edition eliminated the material on copying
what physics should the radiologist know radiographs and subtraction techniques.
to function competently in the environ­ We have reduced the amount of informa­
ment in which he finds himself. As we sit tion on cinefluorography and non-image­
in our little room in Dallas and ponder this intensified fluoroscopy, and consolidated
fluoroscopic imaging and recording into a
question we realize that we will fail to in­
single chapter. Several areas, such as atten­
clude information that should be known,
uation and grids and beam restrictors, re­
and undoubtedly we will include some
main essentially unchanged. In other areas
things that need not be known. It is obvious
we have tried to update the material to in­
that one individual is unlikely to become
dicate new technology. These areas include
and remain an expert in physics and tech­
x-ray generators (solid-state devices), xe­
nology, plus diagnostic imaging. There is
rography (liquid toner), CT scanners (fast­
just too much to learn in each field, and
imaging technology) and ultrasound (color
learned information becomes obsolete too Doppler). Obviously, the needed addition
fast. We applaud the developing trend that was in MRI, and a new chapter was added
pairs a radiologist and a radiologic physicist to expand this topic.
as a team that provides the maximum pa­ We certainly appreciate the positive re­
tient benefit from diagnostic imaging stud­ sponse to this text, and hope the current
ies. Thus, our goal is not to turn radiolo­ edition will continue to fill the needs of
gists into physicists, but rather to allow radiology residents and student technolo­
radiologists to understand and appreciate gists.

Dallas, Texas Thomas S. Curry, III

James E. Dowdey
Robert C. Murry


Another edition is about wrapped up, selves from the demands and invasions of
causing us once again to reflect on those other academic and clinical responsibilities
who contributed so much time and talent. for several months. The understanding
Our friends at E.I. duPont de Nemours and support of our friends and colleagues
and Company, Eastman Kodak Company, was absolutely spectacular, and this at­
and Philips Medical Systems continued to tempt to say thinks will fall far short of the
offer support. genuine affection we hold for each of these
Marty Burgin once again prepared all splendid individuals. We can name only a
the new illustrations, and introduced us to few, and chose to single out those who were
the wonders of computer graphics. Marty most directly inconvenienced because of
did much more than draw the pictures. She our absence. In the physics department,
spent time helping us improve and refine graduate students Tom Lane, Tim Black­
many of the illustrations, especially those burn, and Jerome Gonzales provided re­
dealing with electronics and MRI. We rec­ search effort and critical review of manu­
ognize and appreciate her considerable tal­ scripts. Our friend, John Moore, continued
ent and kindness. to come up with ideas and suggestions. Our
An amazing stroke of good fortune senior residents in diagnostic radiology
guided a remarkable individual to our pho­ must be the most understanding residents
tography department. Only a few weeks in the cosmos. The seven who functioned
after she joined our department, Claudia in an exemplary manner during the seven
Wylie had assumed responsibility for al­ months they were being given little atten­
most all our photographic support. Clau­ tion from their assigned staff radiologist
dia took great pictures, but her effort did were Drs. Lori Watumull,Jim Fleckenstein,
not stop there. She organized the entire Tom Fletcher, Christine Page, Diane
book, chapter by chapter, and served as Twickler, Mark Girson, and Brian Brue­
liaison between the authors and artist as ning. All our residents are wonderful peo­
additions, alterations, and corrections re­ ple. Our colleagues on the teaching staff
quired attention. The enormous amount of continued with their enthusiastic support
time and dedication Claudia provided was and encouragement. Drs. Helen Redman
one of the delightful surprises one hopes and George Miller absorbed most of the
for but seldom finds. Thank you, Claudia. increased clinical work caused by the fre­
This textbook is made possible by the quent unavailability of T.S.C. Drs. Geor­
entire Department of Radiology, Univer­ giana Gibson and Bill Erdman helped with
sity of Texas Southwestern Medical Center clinical MRI information and illustrations,
at Dallas. Everything in this edition has as did Drs. Hortono Setiawan and Rebecca
been prepared by a physicist and a radi­ Harrell in CT and Ultrasound. We must
ologist sitting at a long table to compose mention those faculty members who really
and revise the manuscript together. It was helped with encouragement when things
necessary for the authors to distance them- didn't go smoothly. These include Drs.


Robert Parkey, George Curry, Peter An­ the manuscript. She accepted this chal­
tich, Geral Dietz, Robert Epstein, Mary lenge, and has shown great skill and ded­
Gaulden, William Kilman, Michael Lan­ ication as work has progressed. Equally im­
day, and Jack Reynolds, who has suffered portant, she has been friendly and patient
through four editions with us. A special as we return page after page marked up
thanks goes to Geoffrey Clarke, Ph.D., who with changes after we had promised that
spent many hours guiding us through the the final version was done. Eula has been
complexities of MRI. a genuine pleasure to work with.
Our search for help with word process­ As always, the final thanks must go to
ing provided a second delightful surprise. our wives and families for continuing to
Eula Stephens had been with us only a few support us and take care of us.
days when we approached her about typing


Radiation.. . . . .. .. . .. . . .. .. . .. . .. . . . . . . .. . .. . . .. . . .. .. . .. . . . . .. . .. . . . . .. . . . . . . . .. .. .. .. 1

2 Production of X Rays................................................................ 10

3 X-Ray Generators.................................................................... 36

4 Basic Interactions between X Rays and Matter.. . . .. . . . . .. . . .. . .. . . . .. . . . . . .. .. . . . 61

5 Attenuation .. . . .. . . .. . .. . . . . . . . . . .. .. . . . . . . . . . . . .. . . . .. .. . . .. . . . . . .. . . . . . . . . . .. . . . .. .. 70

6 Filters.................................................................................. 87

7 X-Ray Beam Restrictors . . . . . . . . . . . . . . . . . .. . . . . .. . .. . . .. . . . . . . . . . . . . . . . . . .. . .. . . . . . . . 93

8 Grids .................................................................................. 99

9 Luminescent Screens................................................................. 118

10 Physical Characteristics of X-Ray Film and Film Processing ..................... 137

11 Photographic Characteristics of X-Ray Film....................................... 148

12 Fluoroscopic Imaging................................................................ 165

13 Viewing and Recording The Fluoroscopic Image................................. 175

14 The Radiographic Image.. . . . . . . . . . . . .. . . .. . . . . . .. .. . . . .. . . . . . . . . . .. . . .. .. . .. . . . . . . . 196

15 Geometry of the Radiographic Image.............................................. 219

16 Body Section Radiography.......................................................... 242

17 Stereoscopy . .. .. .. . . .. . .. .. . .. .. . .. . .. . . . .. . . . . . . .. . .. . . .. . . . . . . . .. . .. . .. .. . . . . . . . . . . . 257

18 Xeroradiography..................................................................... 266

19 Computed Tomography............................................................. 289

20 Ultrasound............................................................................ 323

21 Protection . . . . . . .. .. .. . .. . .. .. . .. . .. .. . .. . . . . . .. . .. . .. . . .. . .. . . . . . . . . . . .. . .. .. . . . .. . . . . 372

22 Digital Radiography.................................................................. 392

23 Nuclear Magnetic Resonance ....................................................... 432

24 Magnetic Resonance Imaging....................................................... 470

Index.................................................................................. 505


1 Radiation

Wilhelm Conrad Roentgen, a German ers. Then he held his hand between the
physicist, discovered x rays on November tube and the screen and, to his surprise,
8, 1895. Several fortunate coincidences set the outline of his skeleton appeared on the
the stage for the discovery. Roentgen was screen. By December 28, 1895, he had
investigating the behavior of cathode rays thoroughly investigated the properties of
(electrons) in high energy cathode ray the rays and had prepared a manuscript
tubes, which consisted of a glass envelope describing his experiments. In recognition
from which as much air as possible had of his outstanding contribution to science,
been evacuated. A short platinum elec­ Wilhelm Conrad Roentgen was awarded
trode was fitted into each end and when a the first Nobel Prize for Physics in 1901.
high-voltage discharge was passed through
this tube, ionization of the remaining gas ELECTROMAGNETIC SPECTRUM

produced a faint light. Roentgen had en­ X rays belong to a group of radiations
closed his cathode ray tube in black card­ called electromagnetic radiation. Electro­
board to prevent this light from escaping magnetic radiation is the transport of en­
to block any effect the light might have on ergy through space as a combination of
experiments he was conducting. He then electric and magnetic fields (hence the
darkened his laboratory room to be sure name electromagnetic). Familiar members
there were no light leaks in the cardboard of the family of electromagnetic radiation
cover. On passing a high-voltage discharge include radio waves, radiant heat, visible
through the tube, he noticed a faint light light, and gamma radiation.
glowing on a work bench about 3 ft away. Electromagnetic (EM) radiation is pro­
He discovered that the source of the light duced by a charge (usually a charged par­
was the fluorescence of a small piece of ticle) being accelerated. The converse is
paper coated with barium platinocyanide. also true; that is, a charge being accelerated
Because electrons could not escape the will emit EM radiation. Right here at the
glass envelope of the tube to produce flu­ beginning we run into our first problem.
orescence, and because the cardboard per­ Physics, our beloved exact science, presents
mitted no light to escape from the tube, he a contradiction . The problem is just this:
concluded that some unknown type of ray in our discussion of atomic structure (see
was produced when the tube was ener­ Chap. 2) we will discuss electrons (charged
gized. We can imagine his excitement as he particles) revolving around the nucleus in
investigated the mysterious new ray. He be­ circular orbits while maintaining a precise
gan placing objects between the tube and energy. (This is a result of the Bohr theory
the fluorescent screen: a book, a block of of the hydrogen atom.) This picture of
wood, and a sheet of aluminum. The atomic structure violates the converse state­
brightness of the fluorescence differed ment above. First, if an electron moves in
with each, indicating that the ray penc�­ a circular orbit, it must have centripetal ac­
trated some objects more easily than oth- celeration (an acceleration toward the cen-

ter of the circular path); therefore, it accelerating charge. Any accelerating

should emit EM radiation. The loss of en­ charge not bound to an atom (including the
ergy would require the electron to change nucleus) will emit EM radiation.
its orbit and energy. As a matter of fact, Some time should be spent here discuss­
there was some heated discussion when ing the production and structure of EM
Bohr first introduced his theory that elec­ radiation. To do that, we will start with a
trons could not possibly be in circular orbits single small charged ball. (Because we rec­
because, being accelerated, they would ognize the electron as a charged particle,
emit energy and spiral into the nucleus. we can put a charge on a ball by adding or
Shifty-eyed physicists easily get around this taking away electrons. If we add electrons,
argument by saying that electrons, after all, we will charge the ball negatively. Subtrac­
are standing waves about the nucleus and tion of electrons results in a positively
therefore do not represent accelerating charged ball.) We can only determine if we
charges. It is fair to say, however, that out­ are successful in placing a charge on the
side the atom a charge being accelerated ball by observing its interactions with the
will emit EM radiation. The energy that world around it. Coulomb studied the
charged particles obtain in circular particle forces that exist between two charged balls,
accelerators is limited by the energy loss to and today we call the forces between
EM radiation as the particles move about charged objects "Coulomb forces." The
the accelerator. A cyclotron is a good ex­ force between two small balls having
ample of an accelerator type that is EM­ charges q1 and q2 is
radiation-limited .

But we haven't finished. In the same r2

atomic structure discussion, we allow elec­
tronic transitions from one energy state to F
force (a vector)
another with emission of EM radiation en­ k = a constant whose value depends
ergy, but with no mention of acceleration. on the system of units
In fact, we really think of instantaneous
..q, and q2 =
charge on balls
r =

distance between balls.

transitions across regions in which elec­
trons cannot possibly exist. This concept of This force is always along the line joining
energy level transitions seems to contradict the two balls. (More accurately, it is the line
the first statement that EM radiation is pro­ joining the centers of the two charge dis­
duced by an accelerated charge. tributions. If we use "point" charges we
There are a couple of observations that don't have to worry about the charge dis­
must be made. First, the world of quantum tribution.) If q1 and q2 are like charges
physics (represented here by Bohr's the­ (same sign; positive and positive or nega­
ory) does not always behave as we, living tive and negative), the force is a repelling
in a somewhat larger world, might expect. force. If q1 and q2 are unlike charges, the
Because x rays are produced in the quan­ force is attractive. Gravitational force also
tum physics world, we must understand has this mathematical form, but it is always
some of the laws governing that world. Sec­ attractive.
ond, in this book we have endeavored to We nearly always introduce the electric
be as physically accurate as possible while field (E) to describe the possible interactive
emphasizing those points that we feel cli­ forces. We define the electric field for a
nicians should understand. charge distribution (q1) as the force that q1
Perhaps we should then rephrase the would exert on a positive unit charge (q2
statement: EM radiation, except for that equals 1). Note that E has a unique value
produced in energy level transitions (in­ and direction at each point surrounding a
cluding nuclear transitions), is produced by charge. Figure 1-1 shows the E field sur-

Figure 1-2 Electric and magnetic fields sur­

rounding a positive charge moving with con­
stant velocity

Figure 1-1 Electric field surrounding a posi­ produces a "kink" in the electric field lines
tive charge at rest
that moves outward from the charge with
a finite (but large) velocity. We can think of
rounding a point positive charge. Note that this kink as being the EM radiation. We
E is radially directed and falls off (de­ don't know any better way to describe this
creases in size) as llr2• The electric field for rather complex concept.
more complicated charge distributions EM radiation is made up of an electric
could be more or even less complicated. field and a magnetic field that mutually
(The electric field for an infinite plane uni­ support each other. Figure 1-3 shows the
formly charged is constant everywhere.) E and H fields and the direction of prop­
This electric field is sometimes called the agation of the EM radiation. Note that E
static (electric is implied) field because the and H are perpendicular, that they reverse
charge is at rest. together, and that both E and H are per­
If the charge moves with a constant ve­ pendicular to the direction of propagation.
locity, we not only see an electric field (E) The concept to be visualized is that E and
moving with the charge, but also a mag­ H interact to build each other up to some
netic field (H) surrounding the line (path) value., then collapse together and build
along which the charge is moving. (A more each other up in the opposite direction.
detailed discussion of magnetic fields will Energy is transmitted through space by the
be found in Chapter 23.) Figure 1-2 shows EM radiation.
the electric field radially directed and the The radiation depicted in Figure 1-3
circular magnetic field (H). Here E and H was produced by a charge oscillating back
(both vectors) are perpendicular at any one and forth. Suppose we consider a radio
The next thing to do is to let the charge
accelerate. Here we have conceptual prob­
lems. At constant velocity, the electric field
moves along with the charge. But, with ac­
celeration, the charge moves to a new lo­
cation before the outer regions of the elec­
tric field realize that the charge is at a
different place than it should be. The elec­
tric field lags behind the charge, as does Figure 1-3 Representation of electromag­
the magnetic field. The lagging behind netic radiation

transmission tower. The purpose of the which makes the electrons oscillate back
tower is to transmit radio signals that, as and forth in the antenna at the frequency
you might have guessed, are EM radiation. of the radio wave . Consequently, the an­
The transmission is accomplished by ac­ tenna detects the EM radiation by the elec­
celerating charge up and down a conductor trons moving under the influence of the E
in the tower. As the charge is accelerated field in the EM radiation. In this type of
up the tower, we might get one section of detection, the EM radiation looks and be­
the EM radiation shown in Figure 1-3. As haves as a wave. As the frequency of the
the charge moves down the tower, the fol­ EM radiation increases, however, a point is
lowing section of the EM radiation is pro­ reached in the frequency range (not a sin­
duced. When the charge changes direction, gle precise frequency) at which the electron
so do the E and H fields. Thus, a radio can no longer follow the electric field. At
wave can be produced by forcing charge to frequencies in this range and higher, the
move up and down the tower by applying electrons interact with the EM radiation as
an alternating voltage to the tower. The if the EM radiation were an energy bundle,
frequency of the radio wave is just how rather than made up of waves. Later, we
many times per second the charge changes will see that coherent scattering of x rays
direction. Of course, we tune our radios at is by an interaction of the wave type, while
home to a given frequency to find a par­ photoelectric absorption is an interaction
ticular station. The information we obtain of the energy bundle (called a photon)
from the radio, perhaps music, is super­ type . To stay away from exotic physics, it
imposed on the EM radiation generated by is necessary to discuss EM radiation as if it
the tower. How that is done is another story. were comprised of both particles (photons).
We would be in excellent shape if EM and waves . Let us hasten to add that EM
radiation didn't interact with the particles radiation will behave as only one of the two,
of our old world. Obviously, without such and that this behavior is always the same
interaction we wouldn't be here to wonder for a given interaction (or experimental
about EM radiation anyway. Life-giving en­ measurement).
ergy from the sun gets here by EM radiac To finish the radio detection, we note
tion. one problem: there are a number of radio
Maybe it would be instructive to see how stations broadcasting at the same time. We
radio waves interact with your radio. Just must pick one frequency (the station we
a little while ago we talked about EM ra­ want to listen to) and discard the rest. This
diation in the form of radio waves being selection is done by a "tank" circuit on the
emitted from a radio tower. (We didn't say end of the antenna. We tune the tank cir­
there that the energy emitted was emitted cuit (by the station selection knob) to keep
outward from the tower in something of a only one frequency. What we hear on the
doughnut-shaped pattern. The maximum radio is the second part of the other story
intensity is emitted perpendicularly to the introduced in the paragraph on radio
tower, while there is no intensity directly transmission.
above the tower. That pattern is best for In the first chapters in this text, we will
us, because most of us don't listen to a radio be interested in how the E part of the EM
while directly above the tower.) What we radiation reacts with electrons. Later,
need to detect the radio wave is some an­ when discussing nuclear magnetic reso­
tenna (an electrical conductor) in the radio nance, we will consider reactions with the
radiation pattern. When the radio wave H (magnetic) part ofthe field that are pro­
(the E and H fields) passes the antenna, the duced by oscillating electrons.
E-field part of the EM radio wave exerts a The interactions of different kinds of
force on the electrons in the antenna, EM radiation are difficult to understand.

Some are explained only if they are as­ Because all types of electromagnetic ra­
sumed to be particles, while others are ex­ diation have the same velocity, the fre­
plained only by theories of wave propa­ quency of the radiation must be inversely
gation. It is necessary to discuss proportional to its wavelength. All types of
electromagnetic radiation as if it were radiation in the electromagnetic spectrum
composed of both particles and waves. differ basically only in wavelength. The
wavelength of a radio wave may be 5 miles
Wave Concept of Electromagnetic
long, while a typical x ray is only 1 billionth
of an inch. The wavelength of diagnostic
Electromagnetic radiation is propagated
x rays is extremely short, and it is usually
through space in the form of waves. They
expressed in angstrom units (A) or nano­
may be compared to waves traveling down
meters (nm). An angstrom is 10-10 m, while
a stretched rope when one end is moved
a nanometer is 10-9m. Therefore, one nm
up and down in a rhythmic motion. While
is equal to 10 A. Or, if you prefer, one A
the waves with which we are familiar must
is equal to 0.1 nm. You may wish to refresh
be propagated in a medium (such as the
your memory of various prefixes by refer­
example of the rope, waves traveling in wa­
ring to Table 1-1. The wavelength of most
ter, or sound waves traveling in air), elec­
diagnostic x rays is between 1 and 0. 1 A.
tromagnetic waves need no such medium;
The wavelength of an electromagnetic
that is, they can be propagated through a
wave determines how it interacts with mat­
vacuum. Waves of all types have an asso­
ter. For example, an electromagnetic wave
ciated wavelength and frequency. The dis­
7000 A (700 nm) long can be seen by the
tance between two successive crests, or
human exe as red light, and a wavelength
troughs, is the wavelength of the wave, and
of 4000 A is seen as blue light. The fre­
is given the symbol A (the Greek letter
quency of blue light may be calculated by
lambda, the initial for length). The number
knowing its wavelength (4000 A = 4 X 10-7
of waves passing a particular point in a unit
of time is called the frequency, and is given
the symbol v (the Greek letter nu, the initial c
c = Av or v = -

for number). If each wave has a length A, A

3 x 10• m/sec
and v waves pass a given point in unit time, v =
4 x 10-7m
the velocity of the wave is given by
v = 7.5 x 1014/sec
V =A X v

For example, if the wavelength is 4 ft and Blue light, with a wavelength of 4000 A,
the frequency is 60 waves/min, then has a frequency of 7.5 x 1014 vibrations
per second. Similarly calculated, the fre­
V = 4 ft x 60/min
V = 240 ft/min quency of an x ray of wavelength 0.1 A is
3 X 1019vibrations per second.
EM radiation always travels at the same ve­
The complete spectrum of electromag-
locity in a vacuum. This velocity is 186,000
miles per second (3 X 108 meters per sec­
Table 1-1. Prefixes
ond), which is usually referred to as the
velocity of light and given the symbol c.
109 gig a G
Therefore, we may express the relation­
1 os mega M
ship between velocity, wavelength, and fre­
103 kilo K
quency as 10-1 deci d
c = AV 10-3 milli m
10-6 micro j.L
c = velocity of light (m/sec)
10-• nano n
A = wavelength (m)
v = frequency (per sec).
10-12 pi co p

netic radiation covers a wide range of wave­ called Planck's constant. (Planck's constant
lengths and frequencies. The various parts in SI units is 6.62 X 10-34 joules seconds
of the spectrum are named according to U·s ]). Planck's constant is normally given
the manner in which the type of radiation in the SI units. We made the conversion to
is generated or detected. Some members keV·sec to make the calculation of the en­
of the group, listed in order of decreasing ergy of x-ray photons have the units of kilo­
wavelength, are electron volts.) The mathematical expres­

Radio, television,
sion is written as follows:
radar: 3 x 10s to 1 em
E = hv
Infrared radiation: 0.01 to 0.00008 em
(8ooo A) E = photon energy
Visible light: 7500 (0.000075 em) to h = Planck's constant
3900 A v = frequency.
Ultraviolet radiation: 3900 to 20 A
Soft x rays: 100 to 1 A The ability to visualize the dual charac­
Diagnostic x rays: 1 to 0.1 A teristics of electromagnetic radiation pre­
Therapeutic x ray and sents a true challenge. But we must un­
gamma rays: 0.1 to 1Q-4 A
avoidably reach the conclusion that EM
There is considerable overlap in the wave­ radiation sometimes behaves as a wave and
lengths of the various members of the elec­ other times as a particle. The particle con­
tromagnetic spectrum; the numbers listed cept is used to describe the interactions be­
are rough guides. It is again stressed that tween radiation and matter. Because we
the great differences in properties of these will be concerned principally with inter­
different types of radiation are attributable actions, such as the photoelectric effect and
to their differences in wav�length (or fre­ Compton scatter, we will use the photon
quency). (or quantum) concept in this text.
The wave concept of electromagnetic ra­ The unit used to measure the energy of
diation explains why it may be reflected, photons is the electron volt (eV). An elec­
refracted, diffracted, and polarized. There tron volt is the amount of energy that an
are some phenomena, however, that can­ electron gains as it is accelerated by a po­
not be explained by the wave concept. tential difference of 1 V. Because the elec­
tron volt is a small unit, x-ray energies are
Particle Concept of Electromagnetic usually measured in terms of the kilo­
Radiation electron volt (keV), which is 1000 electron
Short electromagnetic waves, such as volts. We will usually discuss x rays in terms
x rays, may react with matter as if they were of their energy rather than their wave­
particles rather than waves. These particles lengths, but the two are related as follows:
are actually discrete bundles of energy, and
each of these bundles of energy is called a c = ll.v or v = -

quantum, or photon. Photons travel at the
speed of light. The amount of energy car­ and
ried by each quantum, or photon, depends E = hv
on the frequency (v) of the radiation. If the
frequency (number of vibrations per sec­
ond) is doubled, the energy of the photon
Substituting � for v,
is doubled. The actual amount of energy
of the photon may be calculated by mul­ E =

tiplying its frequency by a constant. The
constant has been determined experimen­ The product of the velocity of light (c) and
tally to be 4.13 X 10-18 ke V·sec, and is Planck's constant (h) is 12.4 when the unit

of energy is keV and the wavelength is in All others are derived from these nine
angstroms. The final equation showing the units, although some are given special
relationship between energy and wave­ names. Table 1-3 gives the seven funda­
length is mental quantities to which the seven base
SI units refer. The MKS units for these
= 12.4
E seven quantities are identical to the SI
units. For comparison, Table 1-3 also lists
E = energy (in keV) those common cgs (centimeter-gram-sec­
A = wavelength (in A).
ond) units that have been used elsewhere
Table 1-2 shows the relationship between in this book. A blank entry does not mean
energy and wavelength for various pho­ that there is no cgs equivalent, but that we
tons. have not used those units elsewhere and do
If a photon has 15 eV or more of energy not wish to add complications. Note that
it is capable of ionizing atoms and mole­ the two supplementary units at the bottom
cules, and it is called "ionizing radiation." of Table 1-3 (radians and steradians)
An atom is ionized when it loses an elec­ merely formalize the use of radians for an­
tron. Gamma rays, x rays, and some ultra­ gular measurements. There are 21T radians
violet rays are all types of ionizing radia­ in 360° (i.e., in a complete circle), so one
tion. radian is about 57.3°; there are 41T stera­
dians in a sphere.
UNITS Units for any physically known quantity
While writing this text we encountered can be derived from these nine SI units.
a problem regarding the units used to Table 1-4 lists some quantities mentioned
measure various quantities. Whenever pos­ in this book, and all but one also have a
sible we have tried to use SI units. Some­ name in SI units. Again, those common
times this resulted in very large or very units used elsewhere are also listed. Of
small numbers, and often in such cases we course, each unit in the table can be ex­
used cgs system units. A brief description pressed in SI base units only, and a column
of units will help you to follow the various is included to show this. Sometimes ex­
units used in this text. pressing one derived SI unit in terms of
The SI system (Systeme Internationale other derived SI units reveals some un­
d'Unites) is a modernized metric system derlying principles of physics. For instance,
based on the MKS (meter-kilogram-sec­ we note that the SI unit of power is the
ond) system. The SI system was originally watt, which in base SI units equals 1
defined and given official status by the m2kg
Eleventh General Conference on Weights . The meaning of the base SI units
and Measures in 1960. (A complete listing
might be immediately apparent to a phys­
of SI units can be found in the National
icist, but ]Is Uoules per second) is a little
Bureau of Standards Special Publication
easier to comprehend. This is just energy
330, 1977 edition.) There are only seven
per unit time, or power. Electrical power
base units and two supplementary units.
would be even harder to understand in
Table 1-2. Correlation Between Wavelength base SI units, but V · A (volts times amps)
and Energy is our comfortable definition.
WAVELENGTH (A) ENERGY (keV) The SI unit of radionuclide activity is the
0.0005 24,800 becquerel (Becquerel has finally made the
0.08 155 big time, after discovering radioactivity in
0.1 124 1896, the same year that Roentgen discov­
1.24 10
ered x rays). The fact that one Bq is the

Table 1-3. Sl Base and Supplementary Units

UNIT Sl cgs cgs

Sl base units:
Length meter m centimeter em
Mass kilogram kg gram g
Time second- s second s
Electric current ampere A
Temperature kelvin K
Amount of substance mole mol
Luminous intensity candela cd
Sl supplementary units:
Plane angle radian rad
Solid angle steradian sr

Table 1-4. Sl Derived Units with Special Names


Frequency hertz Hz

Force newton N

Energy joule J N·m erg (cgs)

m2kg J
Power watt w -or V ·A
53 s

Charge coulomb c A·s

Radioactivity becquerel Bq curie


m2 J
Absorbed dose gray Gy rad
52 kg

A·s c
kg kg

m2·kg w
Electric potential volt v -

53· A A

A2s• c
Capacitance farad F
m2kg v

Magnetic flux weber Wb V·s
5 2·A

Magnetic flux density � Wb

tesla T gauss
(magnetic induction) 52· A m2

decay of one nucleus per second may be a ingly complex world should uniformly use
little inconvenient, but even simplicity has something like SI units. A sixteenth-cen­
its price. The SI unit of radiation absorbed tury English peasant may not have needed
dose is the gray, which again refers to the to convert a speed of furlongs per fortnight
quantity of ionizing radiation energy ab­ into inches per second, and we should not
sorbed per unit of mass. There is no special have to either.
SI unit corresponding to the familiar ra­
diation exposure unit of the roentgen. The SUMMARY

roentgen was originally defined as the Wilhelm Conrad Roentgen discovered

charge produced in a given mass of air, and x rays on November 8, 1895. X rays are
comparable SI units are C/kg (C = cou­ members of a group of radiations known
lomb). as electromagnetic radiations, of which
A good table is needed to convert the light is the best-known member. They have
older units and United States units (based a dual nature, behaving in some circum­
on the British engineering system) to SI stances as waves and under different con­
units. A foot is about 0.305 m, a joule is 10 ditions as particles. Therefore, two con­
million ergs, and so forth. We will un­ cepts have been postulated to explain their
doubtedly continue to use a combination characteristics. A single particle of radia­
of units from different systems for some tion is called a photon, and we will discuss
time. For instance, we purchase electric x rays in terms of photons.
power in kilowatts (SI units), potatoes in
pounds (British engineering unit), and
l. Glasser, 0.: Wilhelm Conrad Roentgen and the
heat energy in BTUs (British thermal Early History of the Roentgen Rays. Springfield,
units), all different systems. An increas- IL, Charles C Thomas, 1934.

2 Production of X Rays

DIAGNOSTIC X-RAY TUBES present inside the tube, the electrons that
X rays are produced by energy conver­ were being accelerated toward the anode
sion when a fast-moving stream of elec­ (target) would collide with the gas mole­
trons is suddenly decelerated in the "tar­ cules, lose energy, and cause secondary
get" anode of an x-ray tube. The x-ray tube electrons to be ejected from the gas mol­
is made of Pyrex glass that encloses a vac­ ecules. By this process (ionization), addi­
uum containing two electrodes (this is a di­ tional electrons would be available for ac­
ode tube). The electrodes are designed so celeration toward the anode. Obviously,
that electrons produced at the cathode this production of secondary electrons
(negative electrode or filament) can be ac­ could not be satisfactorily controlled. Their
celerated by a high potential difference to­ presence would result in variation in the
ward the anode (positive or target elec�
number and, more strikingly, in the re­
trode). The basic elements of an x-ray tube
duced speed of the electrons impinging on
are shown in Figure 2-1, a diagram of a
the target. This would cause a wide varia­
stationary anode x-ray tube. Electrons are
tion in tube current and in the energy of
produced by the heated tungsten filament
the x rays produced. Actually, this princi­
and accelerated across the tube to hit the
ple was used in the design of the early so­
tungsten target, where x rays are pro­
called "gas" x-ray tubes, which contained
duced. This section will describe the design
small amounts of gas to serve as a source
of the x-ray tube and will review the way
in which x rays are produced. of secondary electrons. The purpose of the
vacuum in the modern x-ray tube is to al­
Glass Enclosure low the number and speed of the acceler­
It is necessary to seal the two electrodes ated electrons to be controlled independ­
of the x-ray tube in .a vacuum. If gas were ently. The shape and size of these x-ray



Figure 2-1 The major components of a stationary anode x-ray tube


tubes are specifically designed to prevent form a vertical spiral about 0.2 em in di­
electric discharge between the electrodes. ameter and 1 em or less in length. When
The connecting wires must be sealed into current flows through this fine tungsten
the glass wall of the x-ray tube. During op­ wire, it becomes heated. When a metal is
eration of the x-ray tube, both the glass and heated its atoms absorb thermal energy,
the connecting wires are heated to high and some of the electrons in the metal ac­
temperatures. Because of differences in quire enough energy to allow them to move
their coefficients of expansion, most metals a small distance from the surface of the
expand more than glass when heated. This metal (normally, electrons can move within
difference in expansion would cause the a metal, but cannot escape from it). Their
glass-metal seal to break and would destroy escape is referred to as the process of
the vacuum in the tube if special precau­ thermionic emission, which may be de­
tions were not taken. Because of this prob­ fined as the emission of electrons resulting
lem, special alloys, having approximately from the absorption of thermal energy.
the same coefficients of linear expansion as The electron cloud surrounding the fila­
Pyrex glass, are generally used in x-ray ment, produced by thermionic emission,
tubes. has been termed the "Edison effect." A
pure tungsten filament must be heated to
Cathode a temperature of at least 2200° Cto emit a
The negative terminal of the x-ray tube useful number of electrons (thermions).
is called the cathode. In referring to an x­ Tungsten is not as efficient an emitting ma­
ray tube, the terms cathode and filament terial as other materials (such as alloys of
may be used interchangeably, a statement tungsten) used in some electron tubes. It
that is not true for other types of diode is chosen for use in x-ray tubes, however,
tubes. In addition to the filament, which is because it can be drawn into a thin wire
the source of electrons for the x-ray tube, that is quite strong, has a high melting
the cathode has two other elements. These point (3370° C), and has little tendency to
are the connecting wires, which supply vaporize; thus, such a filament has a rea­
both the voltage (average about 10 V) and sonably long life expectancy.
the amperage (average about 3 to 5 A) that Electrons emitted from the tungsten fil­
heat the filament, and a metallic focusing ament form a small cloud in the immediate
cup. The number (quantity) of x rays pro­ vicinity of the filament. This collection of
duced depends entirely on the number of negatively charged electrons forms what is
electrons that flow from the filament to the called the space charge. This cloud of neg­
target (anode) of the tube. The x-ray tube ative charges tends to prevent other elec­
current, measured in milliamperes (1 rnA trons from being emitted from the filament
= 0.001 A), refers to the number of elec­ until they have acquired sufficient thermal
trons flowing per second from the fila­ energy to overcome the force caused by the
ment to the target. It is important to un­ space charge. The tendency of the space
derstand where these electrons come from, charge to limit the emission of more elec­
and to remember that the number of elec­ trons from the filament is called the space
trons determines x-ray tube current. For charge effect. When electrons leave the fil­
example, in a given unit of time, a tube ament the loss of negative charges causes
current of 200 rnA is produced by twice as the filament to acquire a positive charge.
many electrons as a current of 100 rnA, and The filament then attracts some emitted
200 rnA produces twice as many x rays as electrons back to itself. When a filament is
100 rnA. heated to its emission temperature, a state
The filament is made of tungsten wire, of equilibrium is quickly reached. In equi­
about 0.2 mm in diameter, that is coiled to librium the number of electrons returning

to the filament is equal to the number of FOCUSING

electrons being emitted. As a result, the
number of electrons in the space charge
remains constant, with the actual number
depending on filament temperature.
The high currents that can be produced
by the use of thermionic emission are pos­
sible because large numbers of electrons
can be accelerated from the cathode (neg­
ative electrode) to the anode (positive elec­
trode) of the x-ray tube. The number of Figure 2-2 A double filament contained in a
focusing cup
electrons involved is enormous. The unit
of electric current is the ampere, which
may be defined as the rate of "flow" when served by looking into the beam exit port
l coulomb of electricity flows through a of an x-ray tube housing (do not forget to
conductor in l sec. The coulomb is the remove the filter).
equivalent of the amount of electric charge Two additional filament arrangements
carried by 6.25 X 1018 electrons. There­ may be seen in highly specialized x-ray
fore, an x-ray tube current of l 00 rnA (0.1 tubes. A tube with three filaments (triple­
A) may be considered as the "flow" of 6.25 focus) is available. Another special appli­
x 1017 electrons from the cathode to the cation is a stereoscopic angiographic tube.
anode in l sec. Electron current across an In this tube the two focal spots are widely
x-ray tube is in one direction only (always separated (about a 4-cm separation). When
cathode to anode). Because of the forces of two films are exposed, using a different
mutual repulsion and the large number of focal spot for each film, a stereoscopic film
electrons, this electron stream would tend pair is produced. This tube is useful in an­
to spread itself out and result in bombard­ giography when rapid exposure of multi­
ment of an unacceptably large area on the ple stereoscopic film pairs is desired. In­
anode of the x-ray tube. This is prevented tervals as short as 0.1 sec between
by a structure called the cathode focusing exposures can be obtained with stereo­
cup, which surrounds the filament (Figs. scopic tubes.
2-2 and 2 4). When the x-ray tube is con­
- Vaporization of the filament when it is
ducting, the focusing cup is maintained at heated acts to shorten the life of an x-ray
the same negative potential as the filament. tube, because the filament will break if it
The focusing cup is designed so that its becomes too thin. The filament should
electrical forces cause the electron stream never be heated for longer periods than
to converge onto the target anode in the necessary. Many modern x-ray circuits con­
required size and shape. The focusing cup tain an automatic filament-boosting circuit.
is usually made of nickel. Modern x-ray When the x-ray circuit is turned on, but no
tubes may be supplied with a single or, exposure is being made, a "standby" cur­
more commonly, a double filament. Each rent heats the filament to a value corre­
filament consists of a spiral of wire, and sponding to low current, commonly about
they are mounted side by side or one above 5 rnA. This amount of filament heating is
the other, with one being longer than the all that is required for fluoroscopy. When
other (Fig. 2-2). It is important to under­ exposures requiring larger tube currents
stand that only one filament is used for any are desired, an automatic filament-boost­
given x-ray exposure; the larger filament ing circuit will raise the filament current
is generally used for larger exposures. The from the standby value to the required
heated filament glows and can be easily ob- value before the exposure is made, and

lower it to the standby value immediately (t) (- )
after the exposure.
Tungsten that is vaporized from the fil­ 20° CATHODE
ament (and occasionally from the anode) is ANODE
deposited as an extremely thin coating on

the inner surface of the glass wall of the x­ 1 I
ray tube. It produces a color that becomes

deeper as the tube ages. Aging tubes ac­ APPARENT
quire a bronze-colored "sunburn." This FOCAL SPOT
tungsten coat has two effects: it tends to
filter the x-ray beam, gradually changing
Figure 2-3 The line focus principle
the quality of the beam, and the presence
of the metal on the glass increases the pos­
sibility of arcing between the glass and the focusing cup. The electron stream bom­
electrodes at higher peak kilovoltage (kVp) bards the target, the surface of which is
values, which may result in puncture of the inclined so that it forms an angle with the
tube. One of the reasons metal, as opposed plane perpendicular to the incident beam.
to glass, x-ray tube enclosures have been The anode angle differs according to in­
developed is to minimize the effect of dep­ dividual tube design and may vary from 6
osition of tungsten on the tube wall. We to 20°. Because of this angulation, when the
will discuss metal tubes later in this chapter. slanted surface of the focal spot is viewed
from the direction in which x-rays emerge
Line Focus Principle from the x-ray tube, the surface is fore­
The focal spot is the area of the tungsten shortened and appears small. It is evident,
target (anode) that is bombarded by elec­ therefore, that the side of the effective, or
trons from the cathode. Most of the energy apparent, focal spot is considerably smaller
of the electrons is converted into heat, than that of the actual focal spot. If the
with less than 1% being converted into x decrease in projected focal spot size is cal­
rays. Because the heat is uniformly distrib­ culated, it is found that the size of the pro­
uted over the focal spot, a large focal spot jected focal spot is directly related to the
allows the accumulation of larger amounts sine of the angle of the anode. Because sine
of heat before damage to the tungsten tar­ 20° = 0.342 and sine l6.SO = 0.284, an
get occurs. The melting point of tungsten anode angle of 16.5° will produce a smaller
is about 3370° C, but it is best to keep the focal spot size than an angle of 20°. Thus,
temperature below 3000° C. The problems as the angle of the anode is made smaller,
posed by the need for.a large focal spot to the apparent focal spot also becomes
allow greater heat loading, and the con­ smaller.
flicting need for a small focal area to pro­ Some newer 0.3-mm focal spot tubes
duce good radiographic detail, were re­ may use an anode angle of only 6°. Such
solved in 1918 with the development of the small angles permit the use of larger areas
line focus principle. The theory of line of the target for electron bombardment
focus is illustrated in Figure 2-3. The size (and heat dissipation), yet achieve a small
and shape of the focal spot are determined apparent focal spot size. For practical pur­
by the size and shape of the electron stream poses, however, there is a limit to which the
when it hits the anode. The size and shape anode angle can be decreased as dictated
of the electron stream are determined by by the heel effect (the point of anode cut­
the dimensions of the filament tungsten off). For general diagnostic radiography
wire coil, the construction of the focusing done at a 40-in. focus-film distance, the
cup, and the position of the filament in the anode angle is usually no smaller than 15°.

Tungsten Focusing
Focal spot size is expressed in terms of
Target Cup
the apparent or projected focal spot; sizes
of 0.3, 0.6, 1.0, and 1.2 mm are commonly
employed. � TRON
Anode (+) ----sTR � (-)
Anodes (positive electrodes) of x-ray
- c, _-_ -_- ___

tubes are of two types, stationary or rotat­

ing. The stationary anode will be discussed
first because many of its basic principles
also apply to the rotating anode.
Stationary Anode. The anode of a sta­
tionary anode x-ray tube consists of a small Figure 2-4 Lateral view of the cathode and
anode of a stationary anode x-ray tube
plate of tungsten, 2- or 3-mm thick, that is
embedded in a large mass of copper. The
tungsten plate is square or rectangular in high temperature is reached by any metal
shape, with each dimension usually being in the immediate vicinity of the focal spot.
greater than 1 em. The anode angle is usu­ If the tungsten target were not sufficiently
ally 15 to 20°, as discussed above. large to allow for some cooling around the
Tungsten is chosen as the target material edges of the focal spot, the heat produced
for several reasons. It has a high atomic would melt the copper in the immediate
number (74), which makes it more efficient vicinity of the target.
for the production of x rays. In addition, All the metals expand when heated, but
because of its high melting point, it is able they expand at different rates. The bond­
to withstand the high temperature pro­ ing between the tungsten target and the
duced. Most metals melt between 300 and copper anode provides technical problems
1500° C, whereas tungsten melts at because tungsten and copper have differ­
3370° C. Tungsten is a reasonably good ent coefficients of expansion. If the bond
material for the absorption of heat and for between the tungsten and the copper were
the rapid dissipation of the heat away from not satisfactorily produced, the tungsten
the target area. target would tend to peel away from the
The rather small tungsten target must copper anode.
be bonded to the much larger copper por­ Rotating Anode. With the development
tion of the anode to facilitate heat dissi­ of x-ray generators capable of delivering
pation. In spite of its good thermal char­ large amounts of power, the limiting factor
acteristics, tungsten cannot withstand the in the output of an x-ray circuit became the
heat of repeated exposures. Copper is a x-ray tube itself. The ability of the x-ray
better conductor of heat than tungsten, so tube to achieve high x-ray outputs is limited
the massive copper anode acts to increase by the heat generated at the anode. The
the total thermal capacity of the anode and rotating anode principle is used to produce
to speed its rate of cooling. x-ray tubes capable of withstanding the
The actual size of the tungsten target is heat generated by large exposures.
considerably larger than the area bom­ The anode of a rotating anode tube con­
barded by the electron stream (Fig. 2-4). sists of a large disc of tungsten, or an alloy
This is necessary because of the relatively of tungsten, which theoretically rotates at
low melting point of copper (1070° C). A a speed of about 3600 revolutions per min­
single x-ray exposure may raise the tem­ ute (rpm) when an exposure is being made.
perature of the bombarded area of the In practice, the anode never reaches a
tungsten target by 1000° C or more. This speed of 3600 rpm because of mechanical

factors such as slipping between the rotor however, the electrons will bombard a con­
and bearings; therefore, to calculate the stantly changing area of the target. The
ability of a tube to withstand high loads, a total bombarded area of tungsten is rep­
speed of 3000 rpm is usually assumed. Al­ resented by a track 7-mm wide that extends
though the actual speed of anode rotation around the periphery of the beveled ro­
varies even between new tubes of identical tating tungsten disc. The effective focal
design, it is safe to assume that anode ro­ spot will, of course, appear to remain sta­
tation of any functioning rotating anode tionary. At a speed of 3600 rpm, any given
will never drop below 3000 rpm, and will area on the tungsten disc is found opposite
usually be greater than 3000 rpm, if 60 the electron stream only once every 1160
hertz (Hz) current is used. sec, arid the remainder of the time heat
The tungsten disc has a beveled edge. generated during the exposure can be dis­
The angle of the bevel may vary from 6 to sipated. During a 1160-sec exposure, the
20°. The bevel is used to take advantage of entire circumference of the tungsten disc
the line focus principle previously de­ will be exposed to the electron beam.
scribed. The purpose of the rotating anode A comparison between the total and in­
is to spread the heat produced during an stantaneous target areas will illustrate the
exposure over a large area of the anode. tremendous advantage offered by the ro­
Figure 2-5 illustrates this principle. (The tating anode tube (Fig. 2-5). At any instant
dimensions are not drawn to scale.) If we an area of 14 mm2 is bombarded by the
assume that the filament and focusing cup electron beam in our example. If we as­
of the x-ray tube produce an electron beam sume the average radius of the bombarded
that covers an area of the anode 7-mm high area of the tungsten disc to be 40 mm,
and 2-mm wide, the area of the anode bom­ which is a typical value, the circumference
barded by electrons is represented by a 14- of the disc at a radius equal to 40 mm will
mm2 rectangle. We recognize that a 2-mm be 251 mm. The total target area will then
focal spot is rarely encountered in modern be represented by the height of the elec­
x-ray tubes, but is a useful way to illustrate tron stream (7 mm) times the average cir­
a point. If the bevel of the target is 16.5°, cumference of the disc (251 mm), or a total
the effective or apparent focal spot size in of 1757 mm2• Even though the total load­
our illustration will be about 2 x 2 mm. If ing area has been increased by a factor of
the anode were stationary, the entire heat about 125 (14 versus 1757 mm2), the ap­
load would be delivered to this one small parent or effective focal spot size has re­
14-mm2 area of the target. If the target is mained the same.
made to rotate at a speed of 3600 rpm, The diameter of the tungsten disc de­
termines the total length of the target track,
and obviously affects the maximum per­
missible loading of the anode. Typical disc
diameters measure 75, 100, or 125 mm.
To make the anode rotate, some me­
C=2rrr chanical problems must be overcome, be­
C = 211 40 mm. cause the anode is contained within the vac­
C = 251 mm.
uum of the tube. The power to effect
rotation is provided by a magnetic field
produced by stator coils that surround the
neck of the x-ray tube outside the envelope
7 x 251 = 1757 mm.
(Fig. 2-6). The magnetic field produced
Figure 2-5 A rotating anode mcreases the by the stator coils induces a current in the
total target area copper rotor of the induction motor, and

STEM ANODE absorbed by the bearings of the anode as­
sembly would cause them to expand and
bind. Because of this problem the stem
I 1111 (Fig. 2-6), which connects the tungsten tar­
ROTOR get to the remainder of the anode assem­
INDUCTION bly, is made of molybdenum. Molybdenum
has a high melting point (2600° C) and is

IIII a poor heat conductor. Thus, the molyb­

denum stem provides a partial heat barrier
between the tungsten disc and the bearings

FOCUSING CUP of the anode assembly.
The length of the molybdenum stem is
Figure 2-6 The major components of a ro­
tating anode x-ray tube another important consideration. As a
length of the stem is increased, the inertia
(this is really a gyroscopic, rather than in­
this induced current provides the power to ertial, problem) of the tungsten disc in­
rotate the anode assembly. The clearance creases; this increases the load on the bear­
between the rotor and the neck of the tube ings of the anode assembly. It is desirable
is made as small as possible to ensure max­ to keep the stem as short as possible. This
imum efficiency in utilization of the mag­ problem is reduced in metal tubes by the
netic force supplied by the stator core and use of bearings at each end of the anode
windings. Early in the development of ro­ axle (see Fig. 2-12).
tating anode x-ray tubes, the life of the tube Even if all the factors that affect rotation
was quite short because of the lack of du­ of the relatively heavy anode assembly are
rable bearings on which the anode assem­ optimally controlled, inertia is still a prob­
bly could rotate. Because of the friction lem. Because of this inertia there is a short
produced it was necessary to lubricate the delay between application of force to the
bearings, but commonly available lubri­ anode assembly and the time at which the
cants could not be used. Lubricants such as rotor reaches its full angular velocity. This
oil would vaporize when heated and de­ period usually varies from 0.5 to 1 sec . A
stroy the vacuum in the tube; dry lubricants safety circuit is incorporated into the x-ray
such as graphite would wear off as a pow­ circuit that prevents an x-ray exposure
der and destroy the vacuum. This problem from being made until the rotor has
was solved by the use of metallic lubricants reached its full speed.
(especially silver), which are suitable for use The life of a rotating anode x-ray tube
in a high vacuum. In modern rotating may be limited by roughening and pitting
anode tubes, bearing wear has become a of the surface of the anode exposed to the
negligible factor in overall tube life. electron beam. These physical changes are
Heat dissipation in a rotating anode tube the result of thermal stress, and they act to
presents an additional problem. Heat gen­ diminish the x-ray output of the tube. The
erated in a solid tungsten disc is dissipated decreased output of the x radiation results
by radiating through the vacuum to the from excessive scattering of the x rays
wall of the tube, and then into the sur­ (more radiation is directed away from the
rounding oil and tube housing. Recall that exit window of the tube) and increased ab­
heat is dissipated in a stationary anode by sorption of x rays by the target itself. The
absorption and conductivity is provided by combination of short exposure time and
the massive copper anode. In the rotating fast anode rotation causes very rapid heat­
anode tube, absorption of heat by the ing and cooling of the surface of the anode
anode assembly is undesirable because heat disc. The loss of heat by radiation from the

disc surface occurs before there is time for than a molybdenum disc. There are also
any significant amount of heat to be con­ some technical problems associated with
ducted into the main mass of the tungsten bonding the tungsten-rhenium layer to the
disc. Therefore, thermal expansion of graphite. A laminated disc with a molyb­
metal on the surface is much greater than denum substrate has a moment of inertia
expansion of metal immediately beneath about 35% less than a solid tungsten disc
the surface. This condition causes stresses of equal diameter and heat capacity. A fur­
to develop that distort the target surface of ther reduction in moment of inertia of at
the anode disc. It has been found that an least 50% can be achieved when the mo­
alloy of about 90% tungsten and 10% rhe­ lybdenum is replaced by graphite.
nium (a heavy metal with good thermal ca­ Some anode discs are manufactured with
pacity) produces an anode that is more re­ slits or grooves in the target surface of the
sistant to surface roughening and has a disc. This allows the material in the focal
higher thermal capacity than an anode of track to expand without producing the me­
pure tungsten. With these improved anode chanical tension that arises in a �solid disc.
discs, roughening of the focal track has The back of the anode disc may be coated
ceased to be a major problem. with a black substance, such as carbon, to
The usual speed of anode rotation using aid in heat dissipation from the anode.
60-cycle current varies between 3000 and
3600 rpm. If the speed of rotation is in­ Grid-Controlled X-Ray Tubes
creased, the ability of the anode to with­ Conventional x-ray tubes contain two
stand heat will become greater because any electrodes (cathode and anode). The
given area of the target is exposed to the switches used to initiate and to stop an ex­
electron beam for a shorter period of time posure with these tubes must be able to
during each revolution of the anode. By withstand the large changes in voltage ap­
use of proper circuits, the speed of anode plied between the cathode and anode. The
rotation can be increased to about 10,000 rather involved topic of dealing with the
rpm. Three modifications of the tube help timers and switches used in the x-ray circuit
to overcome the problem associated with will be considered in Chapter 3. A grid­
this increased velocity; the length of the controlled x-ray tube contains its own
anode stem is made as short as possible to "switch," which allows the x-ray tube to be
decrease the inertia of the anode; the turned on and off rapidly, as is required
anode assembly rotates on two sets of bear­ with cinefluorography.
ings, which are placed as far apart as pos­ A third electrode is used in the grid-con­
sible; and finally, the inertia of the anode trolled tube to control the flow of electrons
is reduced by decreasing the weight of the from the filament to the target. The third
anode itself. This is achieved by employing electrode is the focusing cup that sur­
a compound anode disc in which the largest rounds the filament. In conventional x-ray
part of the disc is made of molybdenum tubes a focusing cup is electrically con­
(specific gravity 10.2), which is considera­ nected to the filament. This focusing cup
bly lighter than tungsten (specific gravity helps to focus the electrons on the target.
19.3). A relatively thin layer of tungsten­ Because each electron is negatively
rhenium alloy attached to the disc serves charged, the electrons repel one another
as the actual target for the electron beam. as they travel to the target. As a result, the
Some laminated discs use carbon (graphite) electron beam (tube current) spreads out.
instead of molybdenum to further de­ The focusing cup is designed to counteract
crease disc inertia. Graphite does not con­ the spread of the electron beam, which can
duct heat as well as molybdenum, so a be accomplished even with the cup and fil­
graphite anode disc will become hotter ament electrically connected.

In the grid-controlled tube, the focusing (the saturation voltage) the space charge
cup can be electrically negative relative to effect, theoretically, has no influence on
the filament. The voltage across the fila­ current flowing in the x-ray tube. In this
ment-grid produces an electric field along region the current is determined by the
the path of the electron beam that pushes number of electrons made available by the
the electrons even closer together. If the heated filament and is said to be emission­
voltage is made large enough, the tube cur­ limited or temperature-limited. Reference
rent may be completely pinched off, a con­ to Figure 2-7 will show that, in actual prac­
dition in which no electrons go from the tice, a continued increase of kilovoltage
filament to the target. The voltage applied above 40 kVp will produce a slight increase
between the focusing cup and filament may in tube current because of a small residual
therefore act like a switch to turn the tube space charge effect. In modern x-ray cir­
current on and off. Because the cup and cuits this slight increase in milliamperes ac­
filament are close together, the voltage nec­ companying increased kilovoltage would
essary to cut off the tube current is not be undesirable, because the tube current
extremely large. For example, to pulse could not be precisely controlled. By the
(turn the tube current on and off with the use of resistors the circuit automatically
focusing cup), a 0.3-mm focal spot tube op­ compensates for this change by producing
erating at 105 kVp requires about - 1500 a slight decrease in filament heating as kilo­
V between the filament and the cup. voltage is increased. Note that different x­
ray tubes have different saturation voltages
Saturation Voltage and require different amounts of space
When the filament of an x-ray tube is charge compensation.
heated, a space charge is produced, as pre­ Heel Effect. The intensity of the x-ray
viously discussed. When a potential differ­ beam that leaves the x-ray tube is not uni­
ence is applied between the cathode and form throughout all portions of the beam.
anode, electrons flow from the filament to The intensity of the beam depends on the
the anode to produce the tube current. If angle at which the x rays are emitted from
the potential applied across the tube is in­ the focal spot. This variation is termed the
sufficient to cause almost all electrons to be "heel effect."
pulled away from the filament the instant Figure 2-8 shows that the intensity of
they are emitted, a residual space charge the beam toward the anode side of the tube
will exist about the filament. As we de­ is less than that which angles toward the
scribed earlier in this chapter, this residual cathode. The decreased intensity of the x­
space charge acts to limit the number of ray beam that is emitted more nearly par­
electrons available, and thus it limits the allel to the surface of the angled target is
current flowing in the x-ray tube. From the
chart shown in Figure 2-7 it can be seen
that, up to about 40 kVp, an increase in 1-
kilovoltage_ produces a significant increase
in x-ray tube current even though filament
� .. .
. . . . . ... .
.. . ..... .:.!.�.�.��.;.r.�����1.�.:. � � � :: .
::::> I
heating remains the same. Above 40 kVp, u I
however, further increases in kilovoltage
produce very little change in tube current.
I- I
In our example, 40 kVp defines the loca­ <! I
tion of the saturation point of this x-ray
tube. Below 40 kVp, the current flowing in kVp--
the tube is limited by the space charge ef­
fect (space-charge-limited). Above 40 kVp Figure 2-7 Saturation voltage

toward the cathode (filament) side of the

ANODE x-ray tube. This is usually illustrated by
pointing out that an anteroposterior (AP)
film of the thoracic spine should be made
with the tube oriented so that its anode end
is over the upper thoracic spine where the
body is less thick; the cathode end of the
tube is over the lower thoracic spine where
/ thicker body structures will receive the in­

/ creased exposure, which they require. Sec­

ond, it can be seen that the heel effect is
/ less noticeable when larger focus-film dis­
tances are used. Third, for equal target­
film distances, the heel effect will be less
31 56 73 85 95 100 103 104 105 102 95
for smaller films. This is because the in­
tensity of the x-ray beam nearest the cen­
Figure 2-8 The heel effect tral ray is more uniform than that toward
the periphery of the beam.
caused by the absorption of some of the x­ Tube Shielding and High-Voltage Ca­
ray photons by the target itself. Beam in­ bles. Although we usually think of x rays
tensity, as related to the angle of emission, as being confined to the beam emerging
varies depending on the physical charac­ from the tube, they are, in fact, emitted
teristics of individual x-ray tubes. Figure with more or less equal intensity in every
2-8 contains average values taken from direction from the target. In addition, the
charts published in several textbooks and x rays are scattered in all directions follow­
is used for purposes of illustration only. If, ing collisions with various structures in and
for example, a 14- X 17-in. film is oriented around the tube. The tube housing is lined
so that its long axis corresponds with that with lead and serves to absorb primary and
of the tube, the heel effect can be studied. secondary x rays that would otherwise pro­
At a 40-in. target-film distance, the anode duce a high intensity of radiation around
end of the film will receive a relative ex­ the tube, resulting in needless exposure of
posure of 73% and the cathode end will patients and personnel, as well as excessive
receive a relative exposure of 105%. Thus, film fogging. The effectiveness of the tube
there is about a 30% difference in the in­ housing in limiting leakage radiation must
tensity of exposure between the two ends meet the specifications listed in the Na­
of the film. If the target-film distance is tional Council on Radiation Protection and
increased to 72 in., it can be determined Measurements Report No. 49, which states
from the chart that the difference in ex­ that: "The leakage radiation measured at
posure intensities will be considerably less, a distance of 1 meter from the source shall
roughly 87 to 104%. not exceed 100 mR (milliroentgens) in an
Three clinically important aspects of the hour when the tube is operated at its max­
heel effect are illustrated with this exam­ imum continuous rated current for the
ple. First, the intensity of film exposure maximum rated tube potential."
on the anode side of the x-ray tube is sig­ Another function of the tube housing is
nificantly less than that on the cathode to provide shielding for the high voltages
side of the tube. This factor can be used required to produce x rays. The high-volt­
in obtaining balanced densities in radio­ age cables, which are connected to the tube
graphs of body parts of different thick­ through appropriate receptacles in the
nesses. The thicker parts should be placed tube housing, contain a grounding sheath

of wires to provide proper grounding of age and current and exposure time. This
the tube to the earth. To prevent short­ energy is currently expressed in two dif­
circuiting between the grounding wires ferent systems:
and the tube, the space between them is Heat units (an artificial system)
filled with extremely thick mineral oil. Sl units (the watt-second or joule)
Thus, the x-ray tube is contained within the It is our task to present both systems and
tube housing, and oil inside the housing
relate their values. Heat units will be elim­
surrounds the tube. The housing is then
inated in the near future but are still en­
carefully sealed to exclude all air, because
countered in manufacturer's charts and
air would expand excessively when heated
the literature.
and rupture the housing. The oil has good
The heat unit (HU) is defined as the
electrical insulating and thermal cooling
product of current (rnA) and kVp and time
properties. Because of its insulating prop­
(sec) for single-phase power supplies-a
erties, the oil allows more compact tubes
very artificial and unfortunate definition.
and housings to be used, because it permits
Recall that in a single-phase generator the
points of high potential difference to be
peak voltage (kVp) is not the average volt­
placed closer to each other. In addition,
age; kVp is actually 1.4 (or 1.35) times the
convection currents set up in the oil help
average voltage (this is more precisely
to carry heat away from the tube. Heat
called the "root mean square," or "r.m.s.,"
from the oil is absorbed through the metal
voltage; we will call it "average" and use
of the tube shield to be dissipated into the
the conversion factor of 1.4). Modern gen­
erators contain a three-phase power supply
As the oil in the tube housing is heated
with an almost constant potential voltage
it will, of course, expand. A metal bellows
(we will assume constant potential) in which
within the tube shield allows the oil to ex­
kVp and average voltage are the same.
pand without increasing pressures on the
What a mess! The HU is defined using kVp
tube and shield, thus averting possible
of a single-phase unit, which is 1.4 times
damage. In addition, the expanded bellows
the average voltage. To find HU for a con­
can be made to operate a microswitch,
stant potential generator, we are required
which will automatically prevent further
(since kVp and average voltage are the
exposures if maximum heating of the oil
same) to multiply kVp by 1.4 because HU
has occurred.
are defined in a system in which kVp =

1.4 times average voltage. An example will

help. Using a technique of 70 kVp, 100
It is customary to speak of the total load
rnA, and 0.1 sec, calculate HU applied to
that can be applied to an x-ray tube in
the x-ray tube for a single-phase and a con­
terms of kilovoltage, milliamperes, and ex­
stant-potential generator:
posure time. The limit on the load that can
Single phase
be safely accepted by an x-ray tube is a
70 kVp x 100 mAx 0.1 sec = 700 HU
function of the heat energy produced dur­
Constant potential
ing the exposure. The maximum tem� er­
70 kVp x 1.4 x 100 mAx 0.1 sec= 980 HU
ature to which tungsten can be safely raised
is generally considered to be C. For identical factors, HU of a constant po­
Above this level considerable vaponzauon tential generator are higher, but the x-ray
of the tungsten target occurs. The rate at tube produces many more x-rays because
which heat is generated by an electric cur­ of the constant voltage.
rent is proportional to the product of the A watt (W) is a unit of power, and is
voltage (kV) and the current (rnA). Thus, defined as the product of 1 V times 1 A:
the total heat produced is a product of volt- 1W=1Vx1A

Note that 1 KV (1000 V) times 1 A Heat Units Joules

(111000 A) is also 1 W: Single phase 700 500

Constant potential 980 700
1 w = 1000 v X 1/1000 A
See what a mess the artificial term "heat
The volt in this definition is the average unit" causes? Do not memorize conversion
volt. We realize that power is the instan­ factors. Remember the definition of a HU,
taneous voltage times current, but using ajoule, and the relationship of average and
the average voltage will keep us from using peak kilovoltage. Then figure out the an­
integral calculus. Power must do some­ swer in the system of units required.
thing to produce energy (you have power The term kilowatt rating (of an x-ray
if you are strong enough to lift a weight; tube) is commonly used to express the abil­
you give the weight energy when you pick ity of the tube to make a single exposure
it up). The unit of energy in the SI system of a reasonable duration. The reasonable
is the joule 0), and is equal to a watt-sec­ exposure time is defined as 0.1 sec. Kilo­
ond. watt ratings are always calculated for an x­
ray tube used with a constant potential
watt x second = watt-second = joule generator and high-speed rotation. For ex­
ample, what is the maximum milliamper­
A HU is a unit of energy, and a watt­ age that can be used at 70 kVp for a single
second or joule is a unit of energy, so we exposure with a 30-kW (30,000-W) x-ray
may compare the HU with the joule. Dif­ tube?
ficulty arises because the HU is defined by
70 kVp x ? mA 30,000 W
kVp of a generator (where kVp and aver­ =

age voltage are not the same), while the ? mA = -- = 429 mA
watt and joule are defined using average
voltage. With a constant potential genera­ Similarly, a 150-kW tube could accept a
tor it is easy to calculate energy in the SI tube current of 2140 rnA. Remember,
system, because joules are simply kVp x these are defined for 0.1-sec exposures.
rnA X time (sec). In a single-phase system One can use an x-ray tube rating chart to
a conversion is required because kVp and calculate the approximate k W rating. Look
average voltage are not the same. One con­ at the theoretical tube rating chart in Fig­
verts kV p to average voltage by dividing by ure 2-9. Find the 0.1 sec line and notice
1.4 (or multiplying by 0.7). In an x-ray tube where it cuts the 70 kVp curve. These two
the current (rnA) is fairly constant so that lines cross at the 500 rnA line, indicating
we do not need to calculate an average that this tube could accept an exposure of
value for rnA. Another example: calculate 70 kVp and 500 rnA at 0.1 sec. The ap­
the joules of energy generated by a tech­ proximate kW rating is:
nique of 70 kVp, 100 rnA, and 0.1 sec for
70 kVp x 500 mA = 35,000 W = 35 kW
a single phase and a constant potential gen­
erator: A similar calculation at 1 00 kVp yields:

100 kVp x 350 mA = 35 kW

Single phase:
-;::; kVp x 100 mA x 0.1 sec = 500 J So this is a 35-kW tube.
To summarize, this section has thus far
Constant potential:
70 kVp x 100 mA x 0.1 sec = 700 J introduced the heat unit, watt-second
Qoule), and kilowatt rating as parameters
Now, let's combine our examples of HU that measure x-ray tube loading.
and joule calculations for an exposure of The amount of heat that can be accepted
70 kVp, 100 rnA, and 0.1 sec: by an x-ray tube without excessive damage

is in part determined by: the type of rec­ kilovoltage to about 70 kVp. In similar
tification and type of power supply (see fashion, the safe loading for any combi­
Chap. 3); the surface area of tungsten bom­ nation of exposure factors can be deter­
barded by electrons (focal spot size, anode mined from the tube rating chart. This
diameter, anode mass, target angle, and chart is used as an example only. The man­
speed of rotation); and the length of the ufacturer of the x-ray tube used in any di­
exposure. agnostic installation always supplies tube
In considering tube rating, three char­ rating charts for the specific circumstances
acteristics are encountered: the ability of under which the tube will be used. For in­
the tube to withstand a single exposure; the stance, tube rating charts for a single-phase
ability of the tube to function despite mul­ power supply are not valid for a three­
tiple rapid exposures (as in angiography); phase power supply. Tube rating charts
and the ability of the tube to withstand mul­ contain information for many different
tiple exposures during several hours of kilovoltages.
heavy use. Angiographic techniques requiring mul­
The safe limit within which an x-ray tube tiple exposures during a short period of
can be operated for a single exposure can time produce large amounts of heat. For
be easily determined by the tube rating any tube used for angiography, appropri­
chart supplied with all x-ray tubes. An ex­ ate ratings will be made available in the
ample of such a chart is given in Figure form of graphs or charts. Several special
2-9. For example, if it is determined that considerations must be kept in mind when
an exposure will require 50 mAs (500 rnA a tube is subjected to the stresses of rapid
at 0.1 sec), reference to the rating chart will sequence radiography:
show that the lines signifying 500 rnA and 1. The surface of the target can be over­
0.1 sec cross at a point that limits maximum heated by repeating exposures before


- Jok��
;;(500 80kVp .......
E r--....
90kVp ...... �r--
� 400 1"--o..
-100kVp ...... r--.l'o-
z �
w -110kVo .......
a: 300 =125kVp -...r--. . �'l'o-
.. r--. '""
� �� �·'
(.) 200 :-0��"'
w """"'�t"��
co I"'
::> 100 r::�

� .01 0.1 1 10
TIME (sec)
Figure 2-9 An x-ray tube rating chart

the surface heat has had time to dis­ Angiographic rating charts are now fre­
sipate into the body of the anode. quently available. Such a chart is illustrated
2. The entire anode can be overheated in Table 2-1 (this is the angiographic rat­
by repeating exposures before the ing chart of a Machlett Dynamax 69 tube
heat in the anode has had time to ra­ used with three-phase rectification, high
diate into the surrounding oil and speed anode rotation, and a 1.2-mm focal
tube housing. spot). Notice the asterisk, which reminds
3. The tube housing can be overheated you to correlate this chart with the graph
by making too many exposures be­ for single-exposure ratings. Let us assume
fore the housing has had time to lose we wish to make 20 exposures at a rate of
its heat to the surrounding air. 2 exposures per second; the chart indicates
4. The total HU of a series of exposures we can use 8800 J per exposure. This re­
made in rapid sequence must not ex­ sults in heat input of 176,000 J in 10 sec.
ceed the HU permissible for a single Reference to the single-exposure graph for
exposure of a duration equal to the this tube (not illustrated) would show that
total elapsed time required to com­ the tube could accept a 10-sec continuous
plete the series of exposures. exposure at 70 kVp and 250 rnA, or an
The chart used to calculate the maxi­ input of 175,000 J (10 x 70 x 250) in 10
mum loading of the x-ray tube for a single sec. The single-exposure graph would also
exposure can also be used to determine the be checked to ensure that the tube could
ability of the tube to withstand multiple accept each individual exposure in the se­
rapid exposures (Fig. 2-9). For example, ries. Both the angiographic rating .chart
assume exposure factors of 500 rnA, 0.1 and the single-exposure rating graph must
sec, and 70 kVp, with 10 exposures to be be used to ensure that planned exposure
made in 10 sec. First, reference to the chart factors will not damage the tube.
will show that the single exposure is within The ability of an x-ray tube to withstand
the capability of the tube. Each exposure heat loading over a period of several hours
will produce 5000 HU (500 x 70 x 0.1 x depends on the anode heat-storage char­
1.4 = 5000). Ten exposures will produce acteristics. Figure 2-10 illustrates an anode
50,000 HU in the 1 O-see interval. The max­ capable of storing 110,000 J. Note that
imum number of heat units the x-ray tube anode cooling is much more rapid when
can accept during a 10-sec period can be the anode has accumulated large amounts
calculated from the chart. Actually, for an of heat. This is a practical application of
exposure time as long as 10 sec, the num­ the physical law stating that heat loss by
ber of heat units produced is relatively in­ radiation is proportional to the fourth
dependent of the kilovoltage used, and the power of the temperature. The rate at
permissible loading of the tube during the which an anode heats during a period of
1 O-see period can be calculated for any con­ constant exposure is important in fluoros­
venient kilovoltage curve. On our chart, copy. For example, Figure 2-10 shows that
this calculation could be made using 70 a continuous heat input of 500 J per second
kVp or 100 kVp. For example, 70 kVp ap­ will produce an accumulation of the max­
plied for 10 sec could be used with a max­ imum number of J in the anode in about
imum milliamperage of 160 rnA, and 70 X 7 min. Therefore, if the tube is used during
160 X 10 X 1.4 would allow 156,000 HU fluoroscopy at 100 kVp and 5 rnA (100 X
input in 10 sec. The same calculation for 5 x 1 = 500 J each second), continuous
100 kVp for 10 sec at approximately 110 fluoroscopy beginning with a cold tube will
rnA would allow 154,000 HU input. In our be limited to 7 min. On the other hand, if
example, the input of 50,000 HU in 10 sec fluoroscopy is carried out at 90 kVp and
is well within the capability of the tube. 3.8 rnA (about 340 J each second), fluo-

Table 2-1. Dynamix 69 Angiographic Rating Chart

OF EXPOSURES 2 5 10 20 30 40 50 60


1 45,000 27,000 17,500 10,800 7,400 5,500 4,400 3,700

2 30,000 20,000 13,600 8,800 6,600 5,400 4,400 3,700
3 24,000 16,500 11,600 7,600 5,800 4,700 4,100 3,600
4 19,500 14,000 10,000 6,800 5,200 4,300 3,700 3,300
5 17,000 12,200 9,000 6,200 4,800 4,000 3,400 3,000
6 14,800 11,000 8,200 5,700 4,500 3,700 3,200 2,900

8 11,800 9,200 7,000 5,000 4,000 3,400 2,950 2,650

10 10,000 7,800 6,200 4,500 3,600 3,100 2,750 2,400
12 8,600 7,000 5,500 4,000 3,300 2,900 2,550 2,250
*Effective focal-spot size 1.20 mm; stator frequency 180 Hz
Used with permission of Machlett Laboratories, Inc., Stamford, CT, 06907

roscopy can be continued indefinitely with­ represented the anode heat storage chart
out causing excessive anode heating. for the tube in question, it could be deter­
One of the most important uses of the mined from the chart that it would require
anode heat-storage chart is to determine approximately 6 min for the tube to cool
the length of time the tube must be allowed from 100,000 to 10,000 J. Thus, if repeat
to cool before additional exposures can be rapid filming were necessary, a delay of 6
made. For example, assume that an angi­ min would be needed because of the lim­
ographic procedure produces 5000 ]/ex­ itation imposed by the ability of the anode
posure. If a rapid filming procedure re­ to store heat.
quired 20 exposures, the total anode Additional consideration must be given
heating would be I 00,000 J. If Figure 2-l0 to the ability of the entire tube housing to

LJ.J 90,000
a::- 80,000
o ·
f-g 70,000
<tE 50,000
w �
o- 30,000
z 20,000

2 3 4 5 6 7

Figure 2-10 A chart of the anode heat storage capacity of an x-ray tube using a constant potential

withstand heat. The tube housing can ab­ (instead of glass) offers several advantages,
sorb an enormous amount of heat (a figure the three most important being:
of 1,500,000 J is not unreasonable). Al­ 1. Less off-focus radiation
though the tube housing can absorb large 2. Longer tube life with high tube cur­
amounts of heat, it also requires consid­ rents
erable time for cooling. Reference to some 3. Higher tube loading
typical charts will show that it takes, in gen­ The metal envelope is grounded; this
eral, about 30 min for a tube housing to grounding plus use of ceramic insulation
dissipate 600,000]. If more rapid cooling allows adequate electrical safety despite the
of a tube housing is required, the use of tube's small size.
air circulators will usually double the cool­ Off-Focus Radiation. Off-focus radia­
ing rate. Cooling rates exceeding 96,000 J tion is produced by an x-ray tube when
per minute are possible. high-speed electrons interact with metal
surfaces other than the focal track of the
Metal/Ceramic X-Ray Tubes anode (usually other parts of the anode).
A high-performance x-ray tube intro­ The main source of off-focus electrons is
duced by Philips Medical Systems* bears electron backscatter from the anode. These
the trade name of a ceramic Super Rolatix scattered electrons may then strike the
tube. This tube has a metal casing instead anode a second time and produce x-rays
of the usual glass envelope, and has three from areas other than the anode focal
ceramic insulators. Two insulators provide track.
insulation for the two (positive and nega­ Off-focus radiation may be partly con­
tive) high-voltage cables, and one supports trolled by placing the collimator, or a lead
the anode stem. We shall explore the ad­ diaphragm, as close to the x-ray tube as
vantages of such construction. A picture of possible. A small port close to the anode
this tube is shown in Figure 2-11, and a will stop more of the widely spread off­
schematic diagram in Figure 2-12. focus beam than will a similar sized port
Notice in Figure 2-12 that the anode ro­ placed farther from the anode.
tates on an axle which has bearings at each The metal enclosure decreases off-focus
end to provide greater stability and reduce radiation by attracting off-focus electrons
stress on the shaft. This additional support to the grounded metal wall of the x-ray
allows use of a more massive anode, with tube. Since the metal enclosure is at zero
anode weights of 2000 g possible. Anodes potential (i.e., it is grounded), the enclo­
in conventional x-ray tubes are generally sure is relatively positive as compared to
limited to no more than 700 g. the electrons, which are at a negative po­
Ceramic insulators are used to insulate tential. Off-focus electrons may be at­
the high voltage parts of the x-ray tube tracted to the anode or to the grounded
from the metal tube envelope. Aluminum x-ray tube enclosure. Where the electron
oxide is a commonly used ceramic insula­ goes is in part a function of its distance

tor. Ceramic insulators have long been used from the anode or metal wall. Electrons
to attach high-voltage transmission lines to striking the metal wall may produce x-rays,
overhead supporting towers. The three ce­ but the low atomic number metal will pro­

ramic insulators are labeled 4 (high-voltage duce few and low-energy x-rays. The metal
connectors) and 8 (anode shaft) in Figure enclosure thus decreases off-focus radia­
2-12. The use of ceramic insulators allows tion by removing many off-focus electrons.

a more compact tube design. Longer Tube Life. We have already

Using metal as the x-ray tube enclosure mentioned the problem of deposition of
tungsten (from the anode) on the glass wall
*Philips Medical Systems, Inc., Shelton, CT. of x-ray tubes. This tungsten eventually

Figure 2-11 Photograph of a Super Rolatix ceramic x-ray tube. (Courtesy of Philips Medical
Systems, Inc., Shelton, CT)

builds up enough to act as an electrode and INTERACTION OF ELECTRON

cause arcing between the glass and fila­ BEAM WITH X-RAY TUBE TARGET
ment. Tungsten deposition is of greatest
Atomic Structure
concern when high tube currents (high
rnA) are used. A metal enclosed x-ray tube Before attempting to describe the phe­
has its metal enclosure connected to nomena that occur when fast-moving elec­
ground, and deposition of tungsten will not trons encounter the tungsten target in an
alter this grounding. For this reason, the x-ray tube, we must briefly examine the
useful life of a metal-enclosed x-ray tube structure of an atom. In 1897, J.J. Thom­
will be greater than that of a glass tube, son discovered a negatively charged par­
especially when used for high tube cur­ ticle much smaller than any atom, which
rents, such as in angiography. came to be called the electron. Based on
High Tube Loading. The more massive the work of Rutherford and Bohr, a simple
metal anode of this tube allows significantly model of an atom may be visualized as a
higher tube currents (more mAs) to be massive positively charged nucleus sur­
used because of the larger heat storage ca­ rounded by electrons in orbits of specific
pacity of the anode. This allows a higher diameters. The solar system is organized in
mAs setting for single exposures. In ad­ similar fashion.
dition, there is increased capacity for serial Nucleus. The nucleus of an atom is
exposures because of better cooling that made up of several types of elementary
results from more efficient transfer of heat particles, termed nucleons. Of the nucle­
to the oil through the metal enclosure, as ons, only protons and neutrons will be con­
compared to a glass enclosure (metal is a sidered in this discussion, because they are
much better heat conductor than glass). the only ones of importance outside the

In summary, the nucleus of an atom con­

tains protons and neutrons, has a positive
electrical charge, and contains almost all
the mass of an atom.
Electron Orbits and Energy Levels. The
electrons are negative charges revolving
around the nucleus. Because an atom is
always electrically neutral in its normal
state, it must contain an equal number of
protons and electrons. The simplest way to
describe an atom is to visualize a central
positive nucleus with electrons in circular
orbits about the nucleus. In this descrip­
tion, the atom resembles a tiny planetary
system, with the nucleus as the sun and the
2 ANODE DISK electrons as the orbiting planets. Unlike the
3/6 BALL-BEARINGS solar system, with one planet in each orbit,
4/8 CERAMIC INSULATORS the atomic system allows 2 electrons in the
5 CATHODE WITH FILAMENT first orbit, up to 8 in the second, up to 18
7 STATOR WINDINGS in the third, up to 32 in the fourth, and up
9 ANODE SHAFT to 50 in the fifth. The electron orbits are
10 BERYLLIUM WINDOW designated by letters: K, L, M, N, 0, and
so on.
Figure 2-12 Schematic diagram of a Super
Rolatix ceramic x-ray tube. (Courtesy of Philips The planets of the solar system are all
Medical Systems, Inc., Shelton, CT) nearly in the same plane, whereas in the
description we have given the electrons
field of nuclear physics. The proton has a move about the nucleus in spherical shells.
positive electric charge numerically equal An electron in the shell closest to the nu­
to the charge of the electron, while the neu­ cleus is in the K shell and the electron is
tron has zero electrical charge. The neu­ called a K electron. L electrons are in the
tron and proton have about the same mass L shell, the second nearest shell to the nu­
(1.66 X I0-24 g), which is approximately cleus. The diameters of the electronic shells
1836 times greater than the mass of an elec­ are determined by the nuclear force on the
tron. The number of protons in the nu­ electron, and by the angular momentum
cleus is called the atomic number of the and energy of the electron. Angular mo­
atoms, and is given the symbol Z. The total mentum simply indicates that the electron
number of protons and neutrons in the nu­ is moving in a curved path. The attractive
cleus of an atom is called the mass number force between the positively charged nu­
and is symbolized by the letter A. For ex­ cleus and the negatively charged electron
ample, stable gold (Au) has a nucleus with is the force that keeps the electrons in the
79 protons (Z =79) and 118 neutrons (A atom. This force is called the "binding
= 197). All atoms of an elemen.t have the force" of the electron, and is inversely pro­
same atomic number (Z), but an element portional to the square of the distance be­
may have several isotopes. All the isotopes tween the nucleus and electron. Therefore,
of an element have the same number of a K electron has a larger binding force than
protons in the nucleus; that is, they have an L electron. Of course, the binding force
the same atomic number but have different of the electron is directed toward the nu­
numbers of neutrons and therefore differ­ cleus, and the electron would move toward
ent atomic masses. the nudeus if it were not moving in a

curved path. The attractive force caused it jumps is not already filled. An electron
by the nucleus keeps the electrons moving can move to either a higher or lower energy
in a circular path. The earth has a gravi­ shell. Electron movement to a lower energy
tational force toward the sun, but it con­ shell (for example, from an L to a K shell)
tinues to move around the sun in a stable results in the emission of energy. The
orbit. Similarly , electrons are in stable or­ amount of energy is equal to the difference
bital shells. in the binding energy between the two
The earth has energy that is the sum of shells. The energy may take the form of a
its potential and kinetic energies. The photon. If the quantity of energy is suffi­
earth's potential energy is a result of the cient, the photon may be called an x ray.
gravitational effect of the sun; its kinetic Electron movement to a higher energy (for
energy is a result of its motion. By defini­ example, from a K to an L shell, or from
tion, kinetic energy must be positive, but the K shell to a free electron) requires the
potential energy can be either positive or addition of energy to the electron. One
negative. Bound particles always have source of this addition may be the absorp­
negative energy. The earth is in a negative tion of an x-ray photon.
energy state, and electrons in an atom are Each of the atomic energy shells, except
also in negative energy states. To free an the K shell, has subshells of slightly differ­
electron from an atom, the energy must be ent energies. For example, the L shell has
raised to zero or to a positive value. The three subshells and an electron in the L
energy that an electron in a shell must be shell may have one of three energy values.
given to raise the energy value to zero is Of course, a single electron will have only
called the binding energy of the electron. one energy value, but the eight electrons
This is also the energy value designated for needed to fill the L shell will be divided
the atomic shell that houses the electron. with two electrons in the lowest, four in the
The atomic shells are also called energy middle, and two in the highest subshells.
shells. Tungsten has a K-shell energy of In a transition from the L shell to the K
70 keV and an L-shell energy of 11 keV. shell, an electron will emit energy precisely
Remember, these are negative energy val­ equal to the difference in the energy of the
ues. To free a K electron from tungsten, shells. If an L electron leaves the highest
the electron must be given 70 keV of en­ subshell and goes to the K shell, it will emit
ergy , while only 11 keV are required to free some value of energy. If it leaves the middle
an L electron. The L electron has 59 keV shell and goes to the K shell, however, it
more energy than the K electron. The will emit a slightly different value. An L
binding energy of the electron shells varies electron in the lowest subshell may not go
from one element to another. For example, to the K shell. This is called a forbidden
tungsten, with an atomic number of74, has transition.
a K-shell binding energy of 70 keV, while There are more elaborate models of
copper, with an atomic number of 29, has atomic structure that suggest that the orbits
a K-shell binding energy of only 9 keV. are elliptic (so are the orbits of the solar
If a small amount of energy were taken sy stem) or that the electron is just a stand­
from the earth, it would move a bit closer ing wave about the nucleus. For diagnostic
to the sun. Atomic structural laws, how­ radiology the circular orbit model is suffi­
ever, prohibit small additions or subtrac­ ciently accurate.
tions of energy from the bound electrons. The diameter of the nucleus of an atom
An electron cannot have any more, or less, is about 5 X 10-15 m and the diameter of
energy than that associated with its energy the entire atom is about 5 X 10-10 m. This
shell, but an electron may jump from one means that the diameter of the atom is
energy shell to another if the shell to which about 100,000 times larger than the di-

ameter of its nucleus. Most of an atom is ferent units. We previously encountered a

an empty space, and this explains why a joule as a watt-second. We leave it as an
high-speed electron may go through many exercise in unit management to determine
atoms before colliding with any of the com­ that a watt-second is equal to a coulomb­
ponents of an atom. It is interesting to vi­ volt (hint: W = V ·A, A = C/sec). Voltage
sualize the hydrogen atom in terms of balls. is expressed as the peak kilovoltage (kVp)
Suppose the hydrogen nucleus (a proton) applied across the x-ray tube (100 kVp =

were a ball 3 in. in diameter. The electron 100,000 peak volts). We must clearly dis­
(assuming electron density to be equal to tinguish between kVp and keV (kiloelec­
proton density) would be a ball � in. in tron volts). The expression 100 kVp means
diameter. The K shell would have a di­ that the maximum voltage across the tube
ameter of 1.5 miles. A normal hydrogen causing acceleration of the electrons is
atom would be visualized as a 3-inch ball 100,000 V. The expression keV denotes the
with a �-in. ball moving on a spherical shell energy of any individual electron in the
% of a mile away. More fun, if all the elec­ beam (100 keV = 100,000 electron volts).
trons in the atoms of the world could be When the x-ray tube is operated at 100
removed and the nuclei packed together (a kVp, few electrons acquire a kinetic energy
condition that exists in the white dwarf of 100 keV because the applied voltage pul­
stars), the diameter of the earth would be sates between some lower value and the
reduced to about 0.1 mile. maximum value (kVp) selected. It takes the
The production of x rays makes use of electrons something like l 0-10 sec to go
three properties of the tungsten atoms in from the cathode to the anode, separated
the target of the x-ray tube: the electric by 1 in., when the applied voltage is 100
field of the nucleus; the binding energy of kilovolts (kV). Each electron will acquire a
orbital electrons; and the need of the atom kinetic energy (eV), where V is the instan­
to exist in its lowest energy state. taneous voltage across the tube. In a single­
phase, full-wave rectified circuit, the volt­
Processes of X-Ray Generation age varies from 0 to the maximum (kVp)
X rays are produced by energy conver­ value selected, at a rate of 120 times per
sion when fast-moving electrons from the second. This will be discussed in more de­
filament of the x-ray tube interact with the tail in Chapter 3. For the present discus­
tungsten anode (target). The kinetic en­ sion, it is adequate to state that the voltage
ergy (E) of the electron in passing across (V) providing the potential to accelerate
the voltage (V) is increased by the electrons is pulsating, so that the energy
(eV) of electrons that encounter the target
E eV
(anode of the x-ray tube) covers a broad

where e is the electronic charge. Because range. In other words, the high-speed
the electric charge (e) of the electron does electrons striking the target do not have
not change (e = 1.60 x 10-19 C), it is ap­ the same energy.
parent that increasing the voltage across X ray s are generated by two different
the x-ray tube will increase the kinetic en­ processes when the high-speed electrons
ergy of the electron. We define the electron lose energy in the target of the x-ray tube.
volt as the energy a single electron obtains One involves reaction of the electrons with
when crossing one volt. We can write: the nucleus of the tungsten atoms, pro­
ducing x rays that are termed general ra­
E 1.6 X 1Q-19 c X 1 v
diation, or bremsstrahlung. The second

= 1.6 X 1Q-19 J
involves collision between the high-speed
Therefore, one electron volt equals 1.6 x electrons and the electrons in the shell of
10-19 J. A joule may be expressed in dif- the target tungsten atoms, producing

x rays that are called characteristic radia­ of collision all the energy of the electron
tion. To repeat, when high-speed electrons appears as a single x-ray photon.
lose energy in the target of an x-ray tube, Usually an electron will undergo many
x rays are produced by two different proc­ reactions before coming to rest, and the
esses: (1) general radiation (bremsstrah­ energy it loses with each reaction is small.
lung), and (2) characteristic radiation. In addition, the electrons in the beam strik­
General Radiation (Bremsstrahlung). ing the target have widely different ener­
When an electron passes near the nucleus gies. These two factors cause a wide dis­
of a tungsten atom, the positive charge of tribution in the energy of the radiation
the nucleus acts on the negative charge of produced by this braking phenomenon.
the electron. The electron is attracted to­ Most of the radiation will have little energy,
ward the nucleus and is thus deflected from and will appear as heat. Few x rays will ap­
its original direction. The electron may lose pear because over 99% of all reactions pro­
energy and be slowed down when its di­ duce heat. The energy of the radiation is
rection changes. The kinetic energy lost by the amount of energy lost by the electrons.
the electron is emitted directly in the form As discussed in Chapter 1, the energy of a
of a photon of radiation. The radiation photon of radiation is inversely related to
produced by this process is called general its wavelength. The wavelength of x-ray
radiation or bremsstrahlung (from the photons produced when the electron is
German for "braking radiation"). Figure braked by the tungsten nuclei in the target
2-13 is a schematic representation of this is related to the energy (keV) of the elec­
production of bremsstrahlung. tron. The energy of the electron is related
Most electrons that strike the target give to the potential difference (kVp) across the
up their energy by interactions with a num­ x-ray tube. Consider the case of a head-on
ber of atoms. The electron gives up only collision between the electron and nucleus.
part of its energy in the form of radiation All the energy of the electron is given to
each time it is "braked." Electrons pene­ the resulting x-ray photon. The minimum
trate through many atomic layers before wavelength (in angstroms) of this x-ray
giving up all their energy; therefore, not photon can be calculated:
all x-rays are produced on the surface of
the target. Occasionally, the electron will Am;n

collide head-on with a nucleus. In this type

For example, using 100-kVp x-ray tube po­
tential, the maximum energy (eV) that an
electron can acquire is 100 keV. An elec­
tron with this energy can produce an x-ray
photon with a minimum wavelength of
0.124 A:
\ \ Am;n =


) I+ Remember, 0.124 A is the shortest wave­

length (highest energy) x-ray photon that
I can be produced with an x-ray tube poten­
tial of 100 kVp. Another way to say this is
that 100-keV electrons that strike a target
can produce x-ray photons with 100 keV
Figure 2-13 The production of general ra­ of energy (at most), but this a rather rare
diation (bremsstrahlung) event. Most of the x rays produced will

have wavelengths longer than 0.124 A. In mum wavelength x-ray photon is dictated
fact, the wavelength of over 99% of the by the x-ray tube voltage (kVp). The max­
radiations will be so long that the radiation imum wavelength (lowest energy) x rays es­
will produce only heat. caping the tube will depend on the filtering
The energy of the emitted x-ray photon action of the enclosure of the x-ray tube
resulting from deceleration of electrons in and on any added filtration.
the electric field of a nucleus depends on Characteristic Radiation. Characteris­
how close the electron passes to the nu­ tic radiation results when the electrons
cleus, the energy of the electron, and the bombarding the target eject electrons
charge of the nucleus. Figure 2-14 is a from the inner orbits of the target atoms.
graph of the distribution, or "continuous Removal of an electron from a tungsten
spectrum," of the wavelengths of x rays re­ atom causes the atom to have an excess
sulting from bombardment of the x-ray positive charge, and the atom thus becomes
tube target by electrons. Notice that there a positive ion. In the process of returning
will be well defined minimum wavelength to its normal state, the ionized atom of
(X.min) of x rays produced; this "-min will, of tungsten may get rid of excess energy in
course, depend on the kVp used. The x-ray one of two ways. An additional electron
beam will also contain all wavelengths of (called an Auger electron) may be expelled
x rays longer than the minimum wave­ by the atom and carry off the excess energy.
length. Filters (see Chap. 6) are used to The ejection of Auger electrons does not
remove the long wavelength low energy produce x rays, and so is not of much in­
x rays from the beam. Therefore, the high­ terest for this discussion. An alternative
est energy x-ray photon leaving the x-ray way to get rid of excess energy is for the
tube depends on the kVp used; the lowest atom to emit radiation that has wavelengths
energy x-ray photon leaving the x-ray tube within the x-ray range. A tungsten atom
does not depend on kVp, but is determined with an inner shell vacancy is much more
by the filter used (or by the absorption of likely to produce an x ray than to expel an
low energy x rays by the envelope of the electron. X rays produced in this manner
tube if no filter is used). are called characteristic x rays because the
the wavelength of x rays in
To review, wavelengths of the x rays produced are
the continuous spectrum varies. The var­ characteristic of the atom that has been ion­
iation is produced by the different ener­ ized. For the purposes of illustration we will
gies with which the electrons reach the discuss ejection of an electron from the K
target, and by the fact that most electrons shell of tungsten and then apply the prin­
give up their energy in stages. The mini- ciples to electrons in other shells of the
tungsten atom.
The binding energy of an electron in the
K shell of tungsten is about 70 keV. There­
fore, a cathode electron must have energy
>­ of more than 70 keV to eject the K-shell
iii electron from its orbit. A 60-kVp electron
beam will not contain any electrons with
� enough energy to eject a K shell electron
from tungsten. After an impinging elec­
A min
tron uses a 70 keV of its energy to eject the
0 .2 .6
K-shell electron, the remaining energy is
WAVELENGTH (A) shared between the initial electron and the
Figure 2-14 The continuous spectrum of ejected electron. Both these electrons leave
x-rays produced by bremsstrahlung the atom (Fig. 2-15). The ionized tungsten

its excess energy. The energy lost by the L­

shell electron is radiated as a single x-ray
photon. In tungsten, the energy of this
x-ray photon is approximately 59 keV,
which is the difference between the bind­
ing energy in the K shell (about 70 keV)
and that in the L shell (about 11 keV). For
tungsten, the energy of this x-ray photon
will always be the same, regardless of the
energy of the electron that ejected the K­
shell electron. Thus, the x-ray photon en­
ergy is a "characteristic" of the K shell of
a tungsten atom. This process is illustrated
in Figure 2-15.
When the L-shell electron moves into the
K shell, a vacancy is created. The vacancy
may be filled from the M shell, and another
x-ray photon will be produced. The energy
of the L-characteristic radiation, however,
will be much less than that of the K-char­
acteristic radiation. In tungsten, L-char­
acteristic x-ray photons have an energy of
about 9 keV (L-shell binding energy is

• about 11 keV, and that of the M shell about

2 ke V). Characteristic radiations will also
be generated from transitions involving the
outer electron shells of tungsten. The en­
ergy of this radiation is small, and ioniza­
tion in these outer shells produces mostly
heat, or x rays that are absorbed by the
walls of the x-ray tube.
Figure 2-16 diagrams the energy of the
K-characteristic x rays of tungsten, super­
imposed on the continuous spectrum
(bremsstrahlung). There is actually more
than one binding energy for each inner
shell in the tungsten atom except the K
Figure 2-15 The production of characteristic shell, and this causes the appearance of sev­
radiation eral different energies of characteristic ra­
diation. In Figure 2-16 the a1 (59.3 keV)
atom is unstable, and the K-shell electron and a2 (57.9 keV) characteristic x rays arise
is rapidly replaced, usually with an electron from transition of L-shell electrons to the
from the L shell. The replacement electron K shell. The [31 (67.2 keV) results from an
may, however, come from other shells in M-shell to K-shell transition, and the [32 (69
the atom. The electron in the L shell has keV) from anN-shell to K-shell transition.
more energy than the K-shell electron, as The dashed line in Figure 2-16 represents
discussed earlier. In its transition from the the low energy x rays produced by brems­
L to the K shell, the electron must give up strahlung that are removed from the x ray


', �""
1-- '
(/) '
z ' �I
w C32



0 50 100 150 200


Figure 2-16 The spectrum of bremsstrahlung and K-characteristic radiation

beam by the enclosure of the x-ray tube used determines how much radiation (the
and the added filtration. quantity) will be produced by a given ap­
What contribution does the characteris­ plied voltage. (The next section will define
tic radiation make to the total production the effect of this voltage on the radiation.)
of x rays by a standard x-ray tube? Below The higher the atomic number of the tar­
70 kVp there is no K-shell characteristic get atoms, the greater will be the efficiency
radiation. Between 80 and 150 kVp, char­ of the production of x rays. For example,
acteristic radiation (K-shell characteristic) tungsten (Z 74) would produce much

contributes about 10% (80 kVp) to 28% more bremsstrahlung than tin (Z 50) if=

(150 kVp) of the useful x-ray beam. Above both were used as the target of an x-ray
150 kVp the contribution of characteristic tube and compared at identical tube po­
radiation decreases, and it becomes negli­ tential (kVp) and current (rnA). Earlier in
gible above 300 kV p. this chapter we pointed out that tungsten
is used as the target material because of its
Intensity of X-Ray Beams relatively high atomic number (74) and its
The intensity of an x-ray beam is defined high melting point (3370° C). Platinum,
as the number of photons in the beam mul­ with a more favorable atomic number of
tiplied by the energy of each photon. The 78, has a melting point of 1770° C, and
intensity is commonly measured in roent­ stable gold (Z=79) melts at 1063° C. Thus,
gens per minute (R/min, or C/kg in the SI for the continuous spectrum, the atomic
system). The intensity of the x-ray beam number of the target material partly de­
varies with the kilovoltage, x-ray tube cur­ termines the quantity of x rays produced.
rent, target material, and filtration. The relationship between atomic num­
Target Material. The target material ber and the production of characteristic ra-

diation is quite different. The atomic num­ voltage for mammography is approxi­
ber of the target material determines the mately 40 kVp. At this voltage the 17.5-
energy, or quality, of characteristic x rays keV K-alpha and 19.6 keV K-beta char­
produced. For example, the K-shell char­ acteristic radiation of molybdenum makes
acteristic x rays for tungsten (Z = 74) vary up a significant portion of the total radia­
from 57 to 69 keV; those of tin (Z = 50) tion output of a molybdenum target x-ray
vary from 25 to 29 keV; and those of lead tube. In Chapter 6 we will discuss use of a
(Z =81) have energies between 72 and 88 molybdenum filter to cause the character­
keV. istic radiation to make up an even larger
Molybdenum Target. With a high atomic fraction of the x-ray beam from a molyb­
number anode like tungsten, the X"ray denum tube.
beam consists almost entirely of brems­ To summarize, the atomic number of the
strahlung radiation. The contribution target material determines the quantity
from characteristic radiation varies some­ (number) of bremsstrahlung produced and
what with tube voltage, but it never makes determines the quality (energy) of the char­
up a large percentage of the total beam. acteristic radiation.
With lower atomic number anodes, how­ Voltage (kVp) Applied. We have re­
ever, bremsstrahlung production is less ef­ viewed how the energy of the photons
ficient. Efficiency also diminishes as the emitted from the x-ray tube depends on
tube voltage is decreased. The combination the energy of the electrons in the electron
of a low atomic number anode and low tube stream that bombards the target of the
voltage reduces the efficiency of brems­ x-ray tube. The energy of the electrons is,
strahlung production to the point at which in turn, determined by the peak kilovoltage
characteristic radiation assumes greater (kVp) used. Therefore, the kVp deter­
importance. Molybdenum anode tubes are mines the maximum energy (quality) of
designed to take advantage of this principle the x rays produced. In addition, higher
for breast radiography. Maximum tube kVp techniques will also increase the quan-


( kVp Constant) (mA Constant)

>­ >­
t- HIGH mA f- HIGH kVp
z / C/)
z /
w w
f­ z

Figure 2-17 The effect of kVp and tube current on the quality and intensity of the x-ray beam

tity of x rays produced. The amount of ra­ length (quality) and intensity of the x-ray
diation produced increases as the square of beam is illustrated in Figure 2-17.
the kilovoltage:
Intensity is proportional to (kVp}2 X rays are produced by energy conver­
sion when a fast-moving stream of elec­
The wavelength of the characteristic ra­
trons is suddenly decelerated in the target
diation produced by the target is not
of an x-ray tube. An x-ray tube is a specially
changed by the kVp used. Of course, the
designed vacuum diode tube. The target
applied kilovoltage must be high enough
of an x-ray tube is usually tungsten or an
to excite the characteristic radiation. For
alloy of tungsten. Heat production in the
example, using a tungsten target, at least
x-ray tube is minimized by using the line
70 kVp must be used to cause the K-char­ focus principle and a rotating anode.
acteristic x rays to appear.
X rays are generated by two different
X-Ray Tube Current. The number of processes, resulting in (1) the production
x rays produced obviously depends on the of a continuous spectrum of x rays (brems­
number of electrons that strike the target strahlung) and (2) characteristic x rays.
of the x-ray tube. The number of electrons The quantity (number) of the x rays gen­
depends directly on the tube current (rnA) erated is proportional to the atomic num­
used. The greater the rnA the more elec­ ber of the target material (Z), the square
trons that are produced; consequently, of the kilovoltage [(kVp)2], and the mil­
more x rays will be produced . This prin­ liamperes of x-ray tube current (rnA). The
ciple was reviewed earlier in this chapter. quality (energy) of the x rays generated de­
The effect of x-ray tube potential (kVp) pends almost entirely on the x-ray tube po­
and rnA (x-ray tube current) on the wave- tential (kVp).

3 X-Ray Generators

An x-ray generator is the device that sup­ propriate kVp, rnA, and exposure time for
plies electric power to the x-ray tube. It is a particular radiographic examination. Me­
not an electrical generator in the strict ters measure the actual rnA and kV p dur­
sense of the word, because by definition a ing the exposure. One exposure button
generator converts mechanical energy into (standby) readies the x-ray tube for expo­
electrical energy. An x-ray generator be­ sure by heating the filament and rotating
gins with a source of electrical energy. In the anode, and the other button starts the
the United States, any building will have exposure. The timing mechanism termi­
115- or 230-V, 60-Hz alternating current nates the exposure.
available. Most radiology departments will The second component of the x-ray gen­
have three-phase power available in the erator, the transformer assembly , is a
range of 208 to 230 V. The x-ray generator grounded metal box filled with oil. It con­
modifies this energy to meet the needs of tains a low-voltage transformer for the fil­
the x-ray tube. The tube requires electrical ament circuit and a high-voltage trans­
energy to meet the needs of the x-ray tube. former and a group of rectifiers for the
The tube requires electrical energy for two high-voltage circuit. The potential differ­
purposes: to boil electrons from the fila­ ences in these circuits may be as high as
ment and to accelerate these electrons from 150,000 V, so the transformers and recti­
the cathode to the anode. The x-ray gen­ fiers are immersed in oil. The oil serves as
erator has a circuit for each of these func­ an insulator and prevents sparking be­
tions, and we will refer to them as the fil­ tween the various components. By defini­
ament and high-voltage circuits. Also, the tion, a transformer is a device that either
generator has a timer mechanism, a third increases or decreases the voltage in a cir­
circuit, which regulates the length of the x­ cuit. A rectifier changes alternating cur­
ray exposure. These three circuits are all rent into direct current.
interrelated but, to simplify the discussion,
we will describe them separately. We will TRANSFORMERS

not, however, discuss the overall design of As mentioned earlier, the x-ray genera­
the electric sy stem but will leave the tech­ tor receives 115- or 230-V, 60-Hz (cy cles
nical problems in the capable hands of the per second) alternating current. Filament
x-ray equipment manufacturer. heating requires a potential difference of
The mechanism of an x-ray generator is approximately 10 V, whereas electron ac­
usually continued in two separate com­ celeration requires a potential difference
partments: a control panel or console and that can be varied between 40,000 and
a transformer assembly. Control panels 150,000 V. Transformers are used to
may be very simple or quite complex, and change the potential difference of the in­
any attempt to describe a single panel or coming electric energy to the appropriate
console would be of little value. The con­ level.
trols allow the operator to select the ap- Before we describe transformers, we


must pause briefly to discuss the meaning

of potential and potential difference. Po­
tential is a relative term. For a discussion
of electrical circuits, the earth (ground
state) is considered to be at zero potential.
A point in a circuit with an excess of elec­
trons has a negative potential, while a point SECONDARY COIL
with a deficiency of electrons has a positive
potential. Both potential and potential dif­
ference are measured in volts. If one point
has a negative potential of 10 V, and an­ CLOSE OPEN
other point has a positive potential of 10 I I
V, the potential difference between them 8 V u-----L PRIMARY COIL

is 20 V, and electrons will tend to flow to­ I

ward the positive potential. Now suppose

one point has a negative potential of 30 V C
and another point a negative potential of
l 0 V; the potential difference is still 20 V.
Figure 3-1 Current induction by a trans­
Electrons will flow toward the point with a former
negative potential of 10 V. This flow of
electrons represents a current, and is pro­
duced by a potential difference. A volt­ When current flows through the pri­
meter is used to measure the potential dif­ mary coil, it creates a magnetic field
ference between two points. The terms within the core, and this magnetic field
"potential difference" and "voltage" are induces a current in the secondary coil.
synonymous, and will be used interchange­ Current only flows through the secondary
ably. circuit when the magnetic field is changing
A transformer consists of two wire coils (either increasing or decreasing): no sec­
wrapped around a closed core. The core ondary current flows while the magnetic
may be a simple rectangle with the wind­ field in the core is in a steady state. We can
ings wound around opposite sides of the demonstrate this principle with a simple
rectangle, such as is shown in Figure 3-lA. experiment. In Figure 3-lA the primary
The circuit containing the first coil (which circuit is connected to a battery and the
is connected to the available electric energy secondary circuit to a voltmeter. When the
source) is called the primary circuit, and switch in the primary circuit is closed, the
the circuit containing the second coil (from battery drives current through the primary
which comes the modified electric energy) coil, which creates a magnetic field in the
is called the secondary circuit. Other core iron core. As the magnetic field increases,
configurations may be used, and the sec­ it induces a current through the secondary
ondary windings may be wrapped on top coil. Thus current builds up a potential dif­
of (but insulated from) the primary wind­ ference between the two ends of the coil,
ings. The core of a transformer is lami­ and the voltmeter needles swings to one
nated. It is made up of thin sheets of special side. As soon as the magnetic field stabi­
iron alloys separated from each other by lizes, the potential across the secondary coil
thin insulating layers. These sheets are drops to zero and remains there until the
clamped tightly together. The purpose of switch in the primary coil is opened. When
the laminations is to reduce eddy currents, the switch is opened, the magnetic field de­
which waste power and appear as heat in creases, and again this changing field in­
the transformer core. duces a potential difference across the sec-

ondary coil . The polarity of the potential 1. The voltage In the two circuits is proportional
to the number of turns in the two coils.
is reversed, and the voltmeter needle
moves in the opposite direction. The volt­ Np
= \jy_
age across the primary coil is shown in Fig­ N, V,
ure 3-1B, and that across the secondary
coil in Figure 3-1C. The important fact to NP = number of turns in the primary coil
remember is that a current only flows in N, = number of turns in the secondary coil

the secondary circuit when the magnetic VP = voltage in the primary circuit

field is increasing or decreasing. No cur­ V, = voltage in the secondary circuit

rent flows while the magnetic field is stable.

For example, suppose the primary coil
For this reason, steady direct current (like
has 100 turns and the secondary coil has
that from a battery) in the primary coil can­
30,000 turns . If the potential difference
not be used to produce a continuous cur­
across the primary coil is 100 V, the poten­
rent through the secondary coil.
tial difference across the secondary coil will
Alternating current is used for a trans­
former because it is produced by a poten­
tial difference (voltage) that changes con­ 100 100

tinuously in magnitude and periodically in 30,000 =

polarity (Fig. 3-2). Current flows in one
direction while the voltage is positive and
in the opposite direction while the voltage v, = 30,000 v

is negative. The most important character­

A transformer with more turns in the sec­
istic of alternating current is that its voltage
ondary coil than in the primary coil in­
changes continuously, so it produces a con­
creases the voltage of the secondary circuit
tinuously changing magnetic field. There­
and, appropriately, is called a step-up
fore, an alternating current in the primary
transformer. One with fewer turns in the
coil of a transfer produces an alternating
secondary coil decreases the voltage and is
current in the secondary coil.
called a step-down transformer.

2. The second law of transformers is simply a re­

Laws of Transformers statement of the law of the conservation of en­
ergy . A transformer cannot create energy. An
Two simple laws govern the behavior of increase in voltage must be accompanied by a
a transformer. corresponding decrease in current. The product
of the voltage and current in the two circuits must
be equal.

VP voltage in the primary coil


IP = current in the primary coil

V, = voltage in the secondary coil
I, = current in the secondary coil

In our previous example the voltage

across the primary coil was 100 V, and that

• TIME t
across the secondary coil 30,000 V. If the
current in the primary coil is 30 A, then
the current in the secondary coil will be

100 X 30 = 30,000 I,
Figure 3-2 Alternating current wave form I, = 0.1A(100mA)

The product of voltage and current is Provides voltage for the x-ray tube fil­
power. If the potential difference is in volts ament circuit
and the current is in amperes, then power Provides voltage for the primary of
will be in watts: the high-voltage transformer
Provides suitable voltage for subsidi­
W =V X I
ary circuits, which we will not con­
W =watts
V =volts sider
I amperes
= Provides a convenient location for the
kVp meter that indicates the volt­
In the last example the power in the trans­
age to be applied across the x-ray
former is 3,000 W; it is the same on both tube
the high-voltage (100 V x 30 A) and low­ An autotransformer consists of a single
voltage (30,000 V x 0.1 A) sides of the
winding wound on a laminated closed core
transformer. The wire in the transformer
(Fig. 3-3). The autotransformer works on
must be large enough to carry the current the principle of self-induction. An alter­
without overheating. As a result, high-volt­
nating current applied between the input
age transformers are both large and heavy,
points (A and B, Fig. 3-3) will induce a
which also makes them very expensive.
flow of magnetic flux around the core. This
In summary, a step-up transformer in­
magnetic flux will link with all the turns
creases the voltage and decreases the cur­
forming the coil, inducing a voltage into
rent, while a step-down transformer de­
each turn of the winding. For example, if
creases the voltage and increases the
230 V are applied between points A and B
current. These laws assume 100% trans­
(Fig. 3-3), and the points A and B connect
former efficiency, which cannot be
to 115 turns of the autotransformer wind­
achieved, but they are sufficiently accurate
ing, the volts per turn will be 2. By a suit­
for our purposes .
able selection of taps one may select the
There are two basic circuits in a diag­
number of turns to supply the necessary
nostic x-ray unit. One circuit contains the
step-up transformer and supplies the high
voltage to the x-ray tube. The other circuit
contains a step-down transformer and sup­
plies the power that heats the filament of
the x-ray tube. A transformer called the
"autotransformer" supplies the primary
voltage for both these circuits. We must
now discuss, in order:

The autotransformer
The x-ray tube filament circuit
The high-voltage circuit

The Autotransformer
The voltage supplied to the x-ray room
connects to the x-ray generator through an
autotransformer in most cases (we will dis­
cuss an arrangement that does not use an
autotransformer when we consider me­
dium-frequency generators). The auto­
transformer has several functions: Figure 3-3 The autotransformer

voltage to the other components of the x­ transformer in the filament circuit has ap­
ray generator. In Figure3-3, a connection proximately 10 to 20 times as many turns
between terminals 0 and P will tap 115 of wire in the primary coil as in the sec­
turns and supply 230 V. There are only 55 ondary coil . The secondary winding of the
turns and 110 V between terminals 0 and filament transformer has only a very small
X, while 160 turns provide 320 V. Notice voltage across it, and is connected to the
that, within a very limited range, an auto­ filament of the x-ray tube . The x-ray tube,
transformer can function as a step-up or of course, has a very high voltage across it.
step-down transformer. This makes it necessary to provide high­
We will discuss the specific connections voltage insulation between the secondary
from the autotransformer (filament circuit, and primary windings of the filament
high tension transformer primary, kV p transformer. The filament transformer is
meter) in the following sections . usually placed in the same oil-filled
grounded metal tank as the high-voltage
Filament Circuit transformer.
The filament circuit regulates current Precise control of filament heating is

flow through the filament of the x-ray tube critical, because a small variation in fila­

(Fig. 3-4 ). The filament is a coiled tungsten ment current results in a large variation

wire that emits electrons when it is heated in x-ray tube current. Remember that x­

by this current flow (thermionic emission, ray tube current is produced by the flow
explained in Chapter 2). Not much power of electrons from their point of origin (the

is needed to heat this filament to the nec­ filament) to the anode (target) of the x-ray
essary high temperature; a current flow of tube.
A change is filament voltage of about
3 to 5 A with an applied voltage of about 5% will result in a 20- to 30-% change in
10 V are typical values. This current merely x-ray tube current. The x-ray filament cur­

heats the filament, and does not represent rent may be controlled by altering the volt­

the current across the x-ray tube. age to the primary of the step-down trans­
The power to heat the x-ray tube fila­ former by addition of resistors connected
ment is provided by a small step-down in series in the circuit leading from the au­

transformer called the "filament trans­ totransformer. The resistors may be a num­
former." The filament is connected directly ber of separate resistors chosen by a switch

to the secondary winding of this trans­ or a push button on the control panel, or

former. The primary winding of the fila­ may be a single variable resistor as shown

ment transformer obtains its voltage by in Figure 3-4. If resistance is increased,

tapping off an appropriate number of more voltage must be used to push current

turns from the autotransformer (Fig.3-5). through the resistance, making less voltage

This voltage will be around 100 to 220 V available to the filament transformer pri­

across the primary winding. To reduce this mary. For example, a current of 4 A and a
to the desired 10 V range, the step-down resistance of 1.5 ohms (D) will reduce volt­
age by 6 V. This is the application of Ohm's
Law, which states:

1 Volts = Current (amperes) x Resistance (ohms)


j ".
Several other components in the fila­
ment circuit are used to stabilize the voltage
to the filament transformer, including a

voltage stabilizer and a frequency stabilizer.
There is also a circuit that automatically
Figure 3-4 Filament circuit compensates for the space charge effect.

These stabilizing and compensating cir­ 600. The potential difference across the
cuits could be shown in Figure 3-4 as boxes secondary coil may be as high as 150,000
in the circuit between the autotransformer V, so the step-up transformer is immersed
and the filament transformer, but we in oil in the transformer assembly for max­
choose to refrain from drawing them. imum insulation.
Two meters are incorporated into the
High-Voltage Circuit high-voltage circuit, one to measure kVp
A simplified schematic of the high-volt­ and the other to measure rnA. The meters
age (cathode-anode) circuit is shown in Fig­ themselves are located on the control
ure 3-5. The circuit has two transformers, panel, but their connections are in the
an autotransformer and a step-up trans­ high-voltage circuit, as shown in Figure
former. We have also shown the x-ray tube 3-5. They indicate the potential across the
filament transformer. The autotrans­ x-ray tube and the actual current flowing
former is actually the kVp selector and is through the tube during an x-ray expo­
located in the control panel. The voltage sure. A voltmeter measures the difference
across the primary coil of the step-up trans­ in electrical potential between two points .
former can be varied by selecting the ap­ Electrons moving through the difference
propriate number of turns in the auto­ in potential constitute an electric current.
transformer. Only five selections are shown In a closed circuit, the same number of
in Figure 3-5, but actually the kVp can be electrons flows through all points. An am­
adjusted in steps from approximately 40 to meter counts the number of electrons flow­
150 kVp. ing past a point per unit time, and it can
The step-up transformer, which is some­ be placed in the circuit wherever it is most
times called the high-voltage transformer, convenient.
has many more turns in the secondary coil The ratio of the voltage across the pri­
than in the primary coil, and it increases mary and secondary coils in a transformer
the voltage by a factor of approximately is proportional to the number of turns in


+110V .. --��------------------
150kVp TO




Figure 3-5 High-voltage (cathode-anode) circuit and x-ray tube filament circuit

the two coils. Some voltage is lost in the Rectification

rectifier circuit, but with the appropriate
Rectification is the process of changing
calibrations the potential difference in the
alternating current into direct current,
high-voltage side of the circuit (i.e., across
and the device that produces the change is
the x-ray tube) can be measured indirectly called a rectifier. The high-voltage trans­
on the low-voltage side of the transformer. former provides an alternating voltage for
Therefore, the kVp meter can be placed in the x-ray tube. The simplest way to use this
the circuit between the autotransformer high voltage is to hook an x-ray tube di­
and step-up transformer, as shown in Fig­ rectly to the secondary windings of the
ure 3-5. The voltage which energizes the step-up transformer, with one side of the
kVp meter is the voltage from the auto­ transformer connected to the cathode (fil­
transformer that will be applied to the pri­ ament) and the other to the anode (target)
mary winding of the high-voltage trans­ of the x-ray tube. Such an arrangement is
former when the exposure begins. Because shown in Figure 3-6. When the cathode is
the kVp meter records the selected kVp negative with respect to the anode, elec­
before the actual exposure begins, it is usu­ trons flow at high speed from the cathode
ally termed the "prereading peak kilovolt­ to the anode and x rays are produced. Dur­
meter." The circuit for the prereading kVp ing the next half of the electrical cycle the
meter shown in Figure 3-5 is greatly sim­ target (anode) of the x-ray tube is negative
plified. The voltage in this circuit is rela­ and the filament positive, so electrons, if
tively small and the meter can be located they are available, would flow away from
on the control panel with a minimum of the target toward the filament. It would be
insulation, and without serious risk of elec­ highly undesirable to have electrons mov­
trical shock. ing from the target to the filament for two
The connections for the rnA meter must reasons: (1) such electrons would not pro-
be in the secondary coil of the high-voltage
transformer to record current flow accu­

rately. Transformers are not 100% effi­
cient, so the current through the primary
coil is not an accurate representation of the
current in the secondary coil. The rnA me­
ter is in a circuit with a potential difference
of up to 150 kVp and, to minimize the risk
of an electric shock, the connections are
made at the point at which the transformer
is grounded, which is the center of the coil.
With a voltage across the coil of 150 kVp,
the potential on one side is + 7 5 kVp and
on the other side -75 kVp. The center of
the coil is at zero potential and, if the meter
is connected at this point, it may be placed
on the control panel without risk of shock
� �
to the operator. Remember, although the FILAMENT TARGET
meter is remote from the x-ray tube, it CATHODE (ANODE)
measures the actual current flow across the
tube, because the same number of elec­
trons flows through all portions of a closed
circuit. Figure 3-6 The circuit for self-rectification

duce useful x rays, and (2) such electrons Rectifiers. A rectifier is a device that
would further heat the filament and reduce allows an electrical current to flow in one
its lifetime. By blocking current flows in the direction but does not allow current to
inverse half of the electrical cycle, the x­ flow in the other direction. Rectifiers are
ray tube changes an alternating current incorporated into the x-ray circuit in series
into a direct current, so it is, in effect, a with the x-ray tube. Exactly the same cur­
rectifier. Because only half of the electrical rent flows through the x-ray tube and the
wave is used to produce x rays, the wave rectifiers.
form is called half-wave rectification. Fig­ High-voltage rectifiers can be of the vac­
ure 3-7A shows the wave form of the in­ uum-tube type (often called "thermionic
coming electrical supply, Figure 3-7B diode tubes") or they can be of solid-state
shows that of half-wave rectification, and composition. In modern equipment, tubes
Figure 3-7C shows that of full-wave rec­ are no longer used and we will not discuss
tification for comparison. Only the upper them. Solid-state rectifiers are smaller,
half of each electrical cycle is used to pro­ more reliable, and have a longer life. Se­
duce x rays. When the x-ray tube itself lenium was the first material used for solid­
serves as a rectifier, the circuit is called state rectifiers. In 1965 high-voltage silicon
"self-rectified." rectifiers were introduced, and today most
Self-rectification has two disadvantages. x-ray generators use silicon rectifiers.
First, half of the available electrical cycle is Semiconductor. The heart of a solid-state
not utilized to produce x rays, so exposure rectifier is a semiconductor, which is usu­
times must be twice as long as they would ally a piece of crystalline silicon. Silicon
be if the whole cycle were utilized. Second, contains four valence electrons. In a solid,
as repeated or prolonged exposures heat such as silicon, there are numerous energy
the anode, it may become hot enough to levels permissible for electrons. Figure 3-8
emit electrons and to produce a current diagrams the last-filled and the first-un­
during the inverse half-cycle. The elec­ filled energy bands for any semiconductor.
trons in this current would bombard the The valence electrons must lose or gain en­
filament and eventually destroy it. There­ ergy to move from one energy level to an­
fore, to protect the x-ray tube and to im­ other. This sounds like the transitions of K
prove the efficiency of the x-ray produc­ and L electrons that produce characteristic
tion, special rectifiers are incorporated into x rays, but in this case energy differences
the high-voltage circuit. are much smaller and the resulting radia­
tion is either heat or light. Electrons in the



B z

c _J

Figure 3-7 Electrical wave forms for full-wave Figure 3-8 Outer electron energy bands in a
and half-wave rectification solid

conduction band (which corresponds to A.

unfilled energy levels) are relatively free
from atomic bonding and may move freely
REG� '

,... ,.... "'"'"
;;., h"'... ,_�
<:' ;;o!i;.<.:t;�"""'".
... ����
-- ·;:-_:: "";<, ._"',:-,

through the semiconductor material.

In the picture of an atom that we present
there are energies that are not available for
electrons to reside in. Obviously, these are
the energies between the allowed energy
levels. As we bring numerous atoms to­
gether to make solid materials, these un­
allowable energies remain to form what we
call the "forbidden energy gap." We can
characterize three types of materials by the
size of the energy gap. If there is no for­
bidden region at normal temperature and c.
pressure, the material is called a "conduc­
tor." If, the forbidden region is in the order
of an electron volt, the material is called a FORBIDDEN GAP
"semiconductor." If the forbidden gap is of
the order of 10 eV the material is an in­
sulator. For a semiconductor at absolute
zero temperature all the electrons are in
Figure 3-9 Electron energy bands for a con­
the valence band (and all the energy states ductor (A), semiconductor (B), and insulator (C)
in that band are filled), and there are no
electrons in the conduction band. At ab­ bond) with silicon. This unbound electron
solute zero temperature, the semiconduc­ can move about in the crystal much easier
tor behaves as an insulator. At room tem­ than one of the bound electrons. The im­
perature, some of the electrons are purity is called a donor since it donates an
thermally raised to the conduction band extra electron . The crystal resulting from
and are available to support a current. So the addition of the donor is called N-type,
the word "semiconductor" describes ma­ with N derived from the negative charge
terial that at low temperatures acts like an of the surplus electron. The most com­
insulator, but at normal room temperature monly used donor materials are arsenic
acts like a conductor. It follows that an in­ and antimony, added in tiny amounts of
sulator will not have electrons in the con­ about one atom of impurity for each 107
duction band at normal temperatures, and
a conductor will always have electrons in COVALENT
the conduction band at normal tempera­ BOND �
tures. This concept is diagrammed in Fig­ • •

ure 3-9.
N-type Semiconductors. Silicon contains
four valence electrons. If a material with

:@ : 8
five valence electrons is added as an im­
;; � ...-- FREE


purity to the silicon lattice (which means ••

crystal), the added atoms will take the place

of some silicon atoms throughout the crys­
o =o=o
· ·

tal. Notice that in Figure 3-10 one of the

SILICON _) • • •

five valence electrons of the impurity is not
utilized in the bonding (this is a covalent Figure 3-10 N-type silicon semiconductor

atoms of silicon. The donor electrons re­ have about the same energy. P-type liD­
quire only about 0.05 eV of energy to raise acceptors.
purities are called
them to the conduction band, compared to P-NJunctions. We will discuss a P-Njunc­
1.1 eV for electrons in the valence band of tion as if it were possible to mechanically
silicon. The .05 eV means this energy level join the N-type and P-type materials. Ac­
is .05 eV below the bottom of the conduc­ tually, a P-N junction can be formed only
tion band, and therefore in the forbidden by a complex process in which the P and
gap. Energy levels that appear in the gap N materials are diffused into a single crys­
are called "traps." tal.
P-type Semiconductors. If an impurity When N-type and P-type crystals are
with only three valence electrons is added joined, a P-N junction is created. The N­
to silicon, the impurity atom will have only type material is rich in electrons (solid cir­
three electrons to share with four sur­ cles in Fig. 3-12A) and the P-type is rich
rounding silicon atoms. This is dia­ in holes (open circles in Fig. 3-12A). When
grammmed in Figure 3-11. One silicon the junction is formed, electrons diffuse
atom now has an electron that is looking across the junction, as shown in Figure
for another electron with which to form a 3-12B. One might expect diffusion to con­
covalent bond. The absence of this electron tinue until there was uniform distribution
is called a "hole." Since the hole is a positive of holes and electrons. This is not the case
"particle," as compared to the negative elec­ because an electrostatic barrier is created
tron, the material is called a P-type semi­
conductor. A hole can migrate through the
lattice structure by an electron from a N p
neighboring bond filling the hole, thus A. 0 0
• • 0
leaving a hole at a new site. Note that the 0
hole moves in a direction opposite that of •
0 0

the electron. This new hole may be filled
• •
by another electron, and so forth. The P­
type traps are about .08 eV above the top
of the valence band and in the forbidden
region. The only other way to form a hole N p
is to raise an electron from the valence B. 0
• •
band to the conduction band, requiring 1.1 0

eV. The P-type impurities used in semi­ 0

• 0
conductors are indium, gallium, and alu­
• •
minum. The trap level is a function of
which impurity is used, but all trap levels

N + p

� :�:�
· 0

• 0


· •

(absence ATOM �
Q o =o =Q
� ·
• • • '--- siLICON LAYER
Figure 3-12 The process of forming a PN
Figure 3-11 P-type silicon semiconductor junction (diode)

that limits diffusion. When electrons leave

the N-type material, the junction area is left
with a net positive charge . Similarly, the P­
type material acquires a negative charge.
This creates what is called a "depletion DIRECTION OF
layer." The depletion layer has a junction
potential that is opposite in sign to the des­
ignation of the materials (i.e., the junction
potential is positive on the side of the N­ DIRECTION OF
type and negative on the side of the P-type �----------------·
material). This is diagrammed in Figure ELECTRON FLOW
3-12C. In silicon, the junction potential is
Figure 3-14 Symbol for a solid-state rectifier
about 0.7 V. The device formed by a P-N (diode)
junction is called a diode. Solid-state rec­
tifiers are diodes. are positive they will contribute to the cur­
If a voltage is applied to a diode, current rent flow.
will flow or not flow depending on the po­ Figure 3-14 shows the symbol for a
larity. If the polarity of the applied voltage solid-state rectifier. Notice that the arrow
is opposite that of the junction, electrons points in the direction of current flow,
will flow from the N-type material across which by convention in physics is opposite
the junction barrier to the P-type material to the flow of electrons. This is confusing,
and current will flow. This is shown in Fig­ but a fact of life when one must deal with
ure 3-13, in which the negative pole of a physics. Figure 3-15 diagrams a solid-state
battery is connected to the N-type material rectifier and the associated terms and sym­
and the positive pole to the P-type. This is bols.
called forward bias. If the polarity of the Silicon Rectifiers. A single silicon rectifier
applied voltage were reversed, with the (called a cell) will resist a reverse voltage of
negative pole of a battery being connected about 1000 V, which is 10 to 20 times
to the P-type material, the junction poten­ higher than a selenium rectifier. Silicon rec­
tial would be augmented and no current tifiers can withstand a temperature of up
would flow. This is called reverse bias. to 392° C, considerably higher than sele­
Since a P-N diode conducts current in a nium at 266° C. A silicon rectifier is made
forward direction only, it meets our defi­ up of a number of cells, or individual di­
nition of a rectifier. Note that holes will odes, connected together to form a cylin-
move in opposite direction to the electrons
under an applied voltage, and since they

N p

,.. - -e ANODE(-) CATHODE(+)

ELECTRON ·------------·

Figure 3-13 Forward bias of a PN diode Figure 3-15 A solid-state rectifier


drical stack that might have dimensions of

20 to 30 em long by 20 mm diameter. Such w
a rectifier can operate up to 150 kVp and
1000 rnA. Modern x-ray equipment uses
solid state silicon rectifiers.
Half-Wave Rectification. We have al­
ready described one form of half-wave rec­
tification, self-rectification by the x-ray '
' R3
tube (see Fig. 3-6). The same wave form �

is produced by two rectifiers connected in

series with the x-ray tube, as shown in Fig­
ure 3-16. With the voltage shown in the
illustration, electrons flow through the
x-ray tube from the cathode to the anode.
When the voltage reverses during the in­
verse half of the alternating cycle, the rec­
tifier stops current flow. When rectifiers are
used in this manner they produce half­
wave rectification . The only advantage of

1------ 1 cycle�

Figure 3-17 The circuit for full-wave rectifi­

+ cation (two pulses per cycle). Combination of
A-=======-B the four diodes is called a diode bridge

the rectifiers is that they protect the x-ray

tube from the full potential of the inverse
e e
Full-Wave Rectification. Modern x-ray
generators employ full-wave rectification,
which utilizes the full potential of the elec­
trical supply. Figure 3-7C shows the wave
form produced by full-wave rectification.
Both halves of the alternating voltage are
..... -----�
used to produce x rays, so the x-ray output
per unit time is twice as large as it is with
I half-wave rectification .
1/60 S9C----+j
Figure 3-17 is a schematic presentation
of the high-voltage circuit for full-wave rec­
tification. The voltage across the circuit is
supplied by the step-up transformer. In
Figure 3-17, if side A of the step-up trans­
) former is negative with respect to B, elec­

\ ,'1 trons will flow from A through rectifier R1

',- ... , I
to the x-ray tube, and return through rec­

r---- 1 cycle -------.; tifier R2 to side B (shown as solid lines

through the rectifiers). Note that electrons
Figure 3-16 The circuit for half-wave recti­ entering the bank of four rectifiers from
fication (one pulse per cycle) side A of the transformer cannot flow

through rectifier R4 to reach the target of

the x-ray tube. This direction of electron
flow produces a reverse bias on the rectifier PHASE I

and current cannot flow. In the following

half-cycle, side B of the transformer be­
comes negative, and side A positive. Elec­ PHASE 2
trons will now reach the filament of the
x-ray tube by flowing from B through rec­
tifier R3 to the filament and return via rec­
tifier R4 to side A of the transformer
(shown as dashed lines through the recti­
fiers). In this manner, the four rectifiers
produce a pulsating direct current (unidi­ PHASES
rectional) through the x-ray tube even 1,2 S 3

though the transformer supplied an alter­

nating input current. The voltage across Three-phase alternating current
Figure 3-19
the tube, however, still fluctuates from zero wave forms
to its maximum level, and x rays are gen­
erated in 120 short bursts each second. Fig­ tage is not shared by three-phase
ure 3-18 shows the intensity of the x-ray generators, which we will discuss next.
beam superimposed on the electrical volt­
age of the x-ray tube during one half-cycle. TYPES OF GENERATORS
As you can see, most of the x rays are gen­ Three-Phase Generators
erated during the central high-voltage por­
Three-phase generators produce an al­
tion of the cycle.
most constant potential difference across
The principal disadvantage of pulsed ra­
the x-ray tube.
diation is that a considerable portion of the
Commercial electric power is usually
exposure time is lost while the voltage is in
produced and delivered by three-phase al­
the valley between two pulses. The time
ternating-current generators. It is easier to
spent bombarding the target with low-en­
understand three-phase current if we think
ergy electrons does little except to produce
of each phase in terms of degrees rather
heat in the target and to produce low en­
than in terms of time. Figure 3-19 shows
ergy x-rays, which are absorbed in the pa­
all three phases separately and superim­
tient and raise patient dose. This disadvan-
posed on one another (bottom). Phase two
lags 120° behind phase one, and phase
three 120° behind phase two. Thus, a
three-phase generator produces an almost
<.9 constant voltage, because there are no deep
<l: >-
1- 1- valleys between pulses.
_.J (f)
0 z Three-phase generators are complex, so
> w
w 1- we will present only a simplified explana­
(!) z
::J tion of three basic types:
1- �
0:: Six pulse, six-rectifier
� X
Six pulse, twelve-rectifier


TIME Three-Phase Transformers. A three­

Figure 3-18 X-ray intensity superimposed on phase transformer has three sets of pri­
the electrical voltage mary and secondary windings. The three

sections of copper windings in the primary DELTA CONNECTED

or secondary are connected in one of two PRIMARY
configurations, termed delta and wye (also WINDING

called star). These are diagrammed in Fig­
ure 3-20. These diagrammatic represen­
tations of the wye and delta windings are GROUNDED
not to be interpreted as the physical struc­
ture of the devices; the arrangement of the c A
coils is suggested because of the voltage re­ WYE (STAR)
lationships. The physical appearance of a CONNECTED
single-phase transformer may look just like
a three-phase transformer except for a few
more wires hanging out of the latter. If the R1 R2
same three-phase voltage is applied to a R3 R4
wye and a delta, the output voltage of the
R5 R6
two are not the same. We find that the out­
put voltage may have the same maximum
value, but there is a 30° shift in the phase
between the two. This phase shift is critical
in the 12-pulse transformer to be discussed
shortly. Generally, the primary windings
Figure 3-21 A six-pulse six-rectifier trans­
are of the delta configuration, and the sec­ former
ondary more often wye or both.
Six-Pulse, Six Rectifier. This design em­
ploys a delta-wound primary transformer
with a wye-wound secondary transformer. put, and the rectifier arrangement will full­
The output of the secondary windings is wave rectify each of the voltages presented
rectified with six solid-state rectifiers. Fig­ in this figure. In the drawing of Figure
ure 3-21 is a diagram of a six-pulse, six­ 3-19, there are three maximum and three
rectifier generator. The wye winding and minimum voltages in one complete cycle
the six rectifiers are essentially three full­ (1160 sec). When rectified, there will be six
wave bridges tied together. Figure 3-19 positive maximum voltages per cycle. Thus
shows the relationship of three-phase out- the term "six pulse." We will follow voltage
through only one set of coils in Figure 3-21
to indicate how full-wave rectification takes
place. Suppose A is negative with respect
to B. Electrons will then flow from A
A DELTA through rectifier R3 to the filament of the
CIRCUIT x-ray tube, then to the target of the tube
and through rectifier R2 to coil B. During
the next half-cycle, B would be negative
with respect to A, and electron flow would
be from B through R1 through R4 to A.
STAR By this method, full-wave rectification of
all three phases will produce six pulses per
cycle (360 pulses per second). Since the
voltage supplied to the x-ray tube never
Figure 3-20 A. Delta winding. B. Wye (or star) falls to zero, the ripple factor is significantly
winding reduced and has a theoretical value of

13.5%. We must now digress a moment and DELTA

explain this term ripple factor. PRIMARY
The ripple factor is the variation in the WINDING
voltage across the x-ray tube expressed as
a percentage of the maximum value. With
a single-phase circuit the ripple factor is FIXED TWO WYE
100% because the voltage goes from zero POTENTIAL SECONDARY
to a maximum value with each cycle (Fig. TO GROUND WINDINGS
3-22). A six-pulse circuit has a ripple factor
of 13.5%, which means that at 100 kV the
voltage fluctuates between 86.5 and 100
kV. A twelve-pulse circuit has a theoretical
ripple factor of 3.5%. When three-phase
generators are operated under load, the
ripple factor is accentuated. This is known
as the load ripple-factor, and is always
greater than the theoretical ripple. The
load ripple-factor of a twelve-pulse, three­
phase system is about 5%.
Six-Pulse Twelve-Rectifier. A six pulse,
twelve-rectifier transformer is dia­
grammed in Figure 3-23. This circuit is Figure 3-23 A six-pulse twelve-rectifier trans-
still a six-pulse circuit with a 13.5% theo­
retical ripple factor. Notice that this circuit
has a fixed potential to ground, an advan­ twelve-rectifier transformer. The differ­
tage over the six-rectifier circuit. This al­ ence is that the secondary is not a double
lows a 150-kV generator to have a trans­ wye connection; it is a wye and a delta con­
former that provides a voltage of -75 kV nection (Fig. 3-24). This is the reason we
to +75 kV to the x-ray tube, simplifying previously mentioned the delta and wye
insulating requirements. types of electrical wiring configuration.
Twelve-Pulse. A twelve-pulse trans­ When a delta and a wye winding are con­
former looks similar to the six-pulse, nected together in the secondary, the out­
put of the delta will lag the wye by 30°. The
result of this is that the output of one wind­
100% 13.5%
ing will fill in the ripple of the other, re­
sulting in a twelve-pulse rather than a six­
pulse output. The theoretical ripple is now
1/60sec-: reduced to 3.5%, with a load ripple factor
A B of about 5%. Modern generators are de­
signed on the principle of one-wye and
one-delta configuration in the windings of
the high-voltage transformer.
Compared to the 100% ripple factor of
1/60sec---l a single-phase system, three-phase gener­
c D ators produce a nearly constant potential.
This nearly constant potential gives three­
Figure 3-22 The ripple factor in single-phase
phase generators a major advantage over
two-pulse (A), three-phase six-pulse (B), and
three-phase twelve-pulse (C) circuits. (D) single-phase generators that produce a
Chopped DC pulsating direct-current potential. Three-

DELTA Capacitor-Discharge Generators. A ca­

PRIMARY pacitor is an electrical device for storing
WINDING charge, or electrons. In a capacitor dis­
charge generator, standard 110- V or 220-
V power is fed into a step-up transformer.
The output of the high-voltage trans­
former is rectified and used to charge a
large capacitor or a bank of capacitors.
Once the capacitor is charged, it can be
discharged through the x-ray tube. This
type of unit uses a grid-controlled x-ray
tube to start and stop the x-ray exposure.
You will recall from Chapter 2 that a grid­
controlled x-ray tube contains its own
"switch" (the grid) that allows the tube to
be turned on and off. The k V across the
x-ray tube is controlled by the kV across
the capacitor or capacitors. The kV across
the capacitor is controlled by the voltage
output of the high-voltage transformer

Figure 3-24 A twelve-pulse three-phase and can be varied by adjusting the kV se­
transformer lector on the autotransformer. Capacitor
discharge units typically provide a very
high milliamperage (as high as 500 rnA) for
phase generators produce x rays efficiently
very short exposure times.
throughout the exposure, and the average
These units are small and easy to move
x-ray energy is somewhat higher because
about. Each exposure starts at the same
no time is spent bombarding the x-ray tube
kVp even if line voltage is undependable,
target with low-energy electrons.
since the capacitors are always charged to
A second major advantage of three­
the same potential. However, the unit has
phase generators is a much higher tube rat­
a significant limitation to its mAs output
ing for extremely short x-ray exposures.
(typically limited to 30 to 50 mAs). When
Three-phase generators are now being
exposure begins, the capacitor begins to
built to deliver a tube current of up to 2000
discharge, so the kV will fall during the
rnA. They make it possible to produce ra­
exposure. As a general rule, there will be
diographs with extremely short exposure
a drop of about l kV for l mAs. For ex­
times and high repetition rates, so they are
ample, an exposure of 30 mAs beginning
excellent for angiography.
at 90 kV would drop to 60 kV by the end
of the exposure. This limits the usefulness
Power Storage Generators
of capacitor discharge generators for use
When mobile radiographic equipment is in radiography of thick body parts such as
taken to a patient's room, the available the abdomen.
power supply is often inadequate. Power The capacitor discharge unit is not a
storage generators provide a means of sup­ cordless mobile unit. The capacitor must
plying power for the x-ray tube independ­ be charged immediately prior to use. It is
ent of an external power supply. There are not practical to charge the unit in the x­
two types of power storage generators: ray department and then take it elsewhere
1. Capacitor discharge generators in the hospital for use, because the charge
2. Battery-powered generators rapidly leaks away from the capacitor. At

the end of each exposure the capacitors ciple of high-frequency current to produce
must be recharged, which in most units is an almost constant potential voltage to the
done automatically. x-ray tube with a transformer of small size.
Battery-Powered Generators. In bat­ The basic principle involved is this: in a
tery-powered generators, a standard transformer, the voltage induced in the
power supply is used to charge large ca­ secondary coil is proportional to the rate
pacity nickel-cadmium batteries. The fully­ of change of current in the primary coil.
charged unit can then operate completely (Induced voltage is proportional to the
independent of connection to an outside time rate of change of magnetic flux. This
power supply. is a statement of Faraday's Law.) Thus far,
The direct current from the batteries we have considered generators in which
does not supply the x-ray tube directly the rate of change of current in the pri­
since high-voltage transformers will not mary coil was determined by 60-Hz power
operate on constant DC power. The output supply except for the 500 HZ of the bat­
from the batteries is fed into a DC chopper, tery-powered generator. A medium-fre­
which interrupts the current many times quency generator converts the 60-Hz
each second; a typical value is 500 hmes . power-line frequency to up to 6500 Hz be­
per second . Figur� 3-22D illustrates what fore it is fed to the primary coil of the trans­
a chopped DC voltage looks like. This pro­ former.
duces a 500-Hz pulsed DC current that is Let us first consider an overview of a me­
supplied to the primary winding of the dium-frequency generator (Fig. 3-25).
high-voltage transformer. The high-volt­ The incoming power supply is standard 60-
age output from the secondary of the trans­ Hz current . The current is rectified and
former is then rectified and supplied to the smoothed. This direct current is then fed
x-ray tube as a 1000-pulse-per-second wave to a device, often called a chopper, which
form. The 1000-pulse wave form, theoret­ converts the smoothed DC into a chopped
ically, has 100% ripple. At this frequency DC with a frequency of about 6500 Hz.
(1000), the wave form can be conveniently (Look at Fig . 3-22D to see what we mean
smoothed by relatively simple circuitry to by chopped DC.) This 6500-Hz chopped
provide nearly constant potential. (The DC supplies the primary of a step-up trans­
same could be done for 120-pulse voltage, former, which steps up the voltage. The
but the circuitry would be very bulky.) high-voltage 6500-Hz output of the trans­
A battery-powered portable unit is heavy former is rectified to produce 13,000 high­
and requires regular battery maintenance. voltage pulses per second, and then
The batteries can be recharged from any smoothed by filters before being applied to
convenient power supply, with full charg­ the x-ray tube. This provides voltage to the
ing requiring about 12 hours. The advan­ x-ray tube that is nearly ripple-free.
tages of the battery-powered unit are its One advantage of the medium-fre­
ability to (1) store considerable energy to quency generator is not apparent: it sup­
generate x rays and (2) to make exposures plies a constant, nearly ripple-free voltage
independent of a power supply. A typical to the x-ray tube regardless of the input
unit can store 10,000 mAs. Unlike a ca­ power. No special power supply or voltage
pacitor-discharge unit, the battery unit regulators are required. Another advan­
supplies a constant output of kV and rnA tage is the very small size of these gener­
throughout the exposure . ators. You may find it hard to believe that
a 30-kW or 50-kW (we will explain kW rat­
Medium-Frequency Generators ings of generators later in this chapter)
A new type of generator, called a "me­ generator can be contained in a single unit
dium-frequency generator," uses the prin- with about the same dimensions and weight



120Hz SMOOTH 6500Hz

DC DC 6500Hz 13000Hz VOLTAGE

Figure 3-25 Block diagram of a medium-frequency generator

as a regular x-ray tube housing. We must ators provide a nearly constant voltage to
explain why increasing the frequency of the x-ray tube that is not dependent on the
voltage input to the primary of a step-up power supply. Operation at a high fre­
transformer allows the transformer to be quency results in a more efficient trans­
made smaller. First, remember that the out­ former that can be made much smaller
put is determined by the rate of change of than conventional transformers . This small
flux, and is proportional to the frequency , size is especially convenient for portable
the number of windings in the secondary, units.
and the cross-sectional area of the core.
V- fnA
The rating of a transformer states the
V = output voltage maximum safe output of its secondary
f = frequency winding. If the rating is exceeded, the
n number of windings
transformer may overheat and burn out its
A core cross-sectional area
insulation and windings. The rating is ex­

For a given transformer, we can main­ pressed as the maximum safe output of its
tain a constant output voltage by increasing secondary winding in kilowatts. Remem­
the frequency and decreasing the number ber that a watt is the unit of electric (as well
of turns or the core cross-sectional area . as mechanical) power, with a kilowatt ob­
(This statement illustrates the point, but it viously being a thousand watts.
is not true that one could indiscriminately For three-phase generators, kilowatt
increase the output of any transformer by (power) ratings are calculated according to
increasing input frequency. Complex de­ the following formula (recall that kV X rnA
sign changes are required.) The result of = watt):
all this is that both the iron core and kV x mA
kW =

amount of wire can be reduced to make a 1000

smaller transformer. For larger generators kW = kilowatts
kV kilovolts
(80 kW and 100 kW) the high-voltage

mA = milliamperes
transformer is not contained in the tube
head, but is still reduced to about one-third Thus, the ratings of a three-phase gener­
the size of a conventional twelve-pulse ator operating at 100 kV and 500 rnA is
transformer. 100 X 500
= SO kW
In summary, medium-frequency gener- 1000

This would then be termed a 50-kW gen­ rating (as determined at 100 kV). The kW
erator. loading a generator can accept may vary
For single-phase generators the formula considerably with different kV and rnA set­
is modified: tings. When comparing generators, one
kV x mAx 0.7 may find the comparison of kW ratings use­
kW ful. But one must also inquire about the

available kW output at the kV and rnA lev­

The factor 0.7 comes from the fact that in
els at which the generators will generally
single-phase generators the voltage varies
be used. This is especially important when
from zero to some peak value. To figure
considering a unit for use at high-kV tech­
the average power we must consider the
average voltage (this is usually called the
root mean square, or R.M.S., voltage). In EXPOSURE SWITCHING
single-phase circuits, this relationship is: A switch is the device that turns the high
peak voltage applied to the x-ray tube on and
R . M .S . = V2 = 0 . 707 peak
off. Switching presents design engineers
with interesting problems. These problems
So, the formula for single-phase-generator
arise from switching off the currents in the
power rating converts the kV to R.M.S.
circuits very rapidly and removing all the
voltage by using the factor v'2' or about energy that is stored in the voltage-smooth­
ing networks. If the current is turned off
0.7. Since the transformer is driving an x­
improperly, high-voltage spikes may be in­
ray tube, we can consider the current to be
troduced that can damage the equipment.
fairly constant, so an average value is not
If the voltages in the smoothing circuits are
needed for current.
not removed, the voltage across the x-ray
In most three-phase circuits, R.M.S. volt­
tube cannot go to zero. Switching circuits,
age is about 0.95 peak, which is close
. then, must solve both of these problems,
enough to l that it can be ignored. This
but how they do this need be of little con­
slight difference actually serves as a safety
cern to the radiologist.
There are two categories of switching for
Kilowatt ratings of x-ray generators are
modern generators. Switching may take
determined when the generator is under
place in the primary circuit of the high­
load, and it is convenient to test at a voltage
voltage transformer where there are high
level of 100 kVp, because calculations are
currents and low voltage. Switching may
simplified. Thus, an 80-kW constant po­
also take place in the secondary circuit
tential generator would be one that could
where there are low currents and high volt­
operate at 100 kV and 800 milliamperes
age. In most general purpose three-phase
(rnA). Be careful! Some constant potential
units, switching occurs in the primary cir­
generators are rated at 150 kVp, and are
cuit and is called primary switching.
called 150-kV generators. The same gen­
Switching in the secondary circuit is gen­
erator may have the capability of produc­
erally used in units designed for rapid, re­
ing 1000 rnA. It is wrong to think of this
petitive exposures or where extremel y
as a 150-kW generator unless it can operate .
short exposure times are needed, and IS
at 150 kVp and 1000 rnA simultaneously.
called secondary switching.
A generator might be able to operate at:
150 kVp and 500 mA (75 kW) Primary Switching
100 kVp and 800 mA (80 kW) There are three types of primary
80 kVp and 1000 mA (80 kW)
switches: electromechanical contractors,
This hypothetical generator has an 80-kW thyratrons, and solid-state silicon-con-

trolled rectifiers. Electromechanical thyrister. The response of the gate is almost

switches and thyratrons are being phased instantaneous, making the thyrister useful
out. when very fast switching is necessary. No­
Silicon-Controlled Rectifiers. The pri­ tice that a thyrister is also a rectifier (this
mary switching found in most modern gen­ is the "R" part of the SCR). Electrons will
erators uses solid-state semiconductors not flow from anode to cathode because
called silicon-controlled rectifiers (SCRs) there are two PN junctions that prevent
or thyristers. SCRs may be thought of as electron flow. We will leave the way SCRs
solid-state thyratrons. Other control recti­ are put into circuits to the engineers.
fiers may come into common use, but we
will talk only about the SCR. A control rec­ Secondary Switching
tifier is a rectifier that can be turned on Secondary switching takes place on the
and off by a logic signal (which in reality high voltage side of the transformer or at
is just a small voltage pulse). the x-ray tube itself. Remember that
Figure 3-26 is a schematic of an SCR; switches in the high-voltage circuit must
its symbol is also shown. This thyrister con­ prevent high-voltage breakdown, so they
sists of a cathode (negative end), an anode must be insulated to withstand high volt­
(positive end), a gate, and three junctions. age. Two types of secondary switching are
Notice that this is similar to a series of di­ used today:
odes in that there are NP and PN junctions. Triode vacuum tubes
If the cathode is made negative and the Grid-controlled x-ray tubes
anode positive, current can flow through It is currently possible, but not yet practical,
the two NP junctions (they are forward bi­ to replace triode vacuum tubes by stacking
ased), but no current can flow through the rectifying devices in a manner similar to
one PN junction (it is reversed biased). Re­ silicon-controlled rectifiers. Grid-con­
member, electrons in a diode will flow trolled x-ray tubes are described in Chap­
freely from N-type material to P-type ma­ ter 2. As with the SCR, we choose not to
terial, but will not flow freely from P to N discuss the way triode tubes are designed
material. If a small positive voltage (about into the circuit.
1 V) is applied to the gate, the reverse bias
at the PN junction will be overcome and Primary Versus Secondary Switching
electrons will flow through the thyrister. While the technical details of switching
This is the way a thyrister functions: a small need not bother us, it will be of some value
positive pulse (the logic signal) to the gate to consider the appropriate practical uses
causes a large current to flow through the or advantages of primary versus secondary
switching. We define "practical" to include
cost and availability in today's technology.
Primary Switching. Almost all general
purpose generators use primary switching.
In today's technology and pricing mecha­
nisms, it is easier and cheaper to put switch­
GATE ing into the primary circuit. Primary

CATHODE switching can produce exposures as short

as 1 or 2 milliseconds (ms), but it cannot
produce these exposures at a high repeti­
tive rate as well as secondary switching can.
A. B.
Secondary Switching. Secondary switch­
Figure 3-26 An SCR, or thyrister, shown ing is used in special-purpose generators,
schematically (A) and as its symbol (B) such as those needed for angiography and

cinefluorography. This technology makes

it easier, compared to primary switching, ......
to have sharp, crisp exposures with rapid w 500 "
on-and-off rates with many repeated ex­
g: 400 [\ �
posures. Triode tubes will allow exposures
as short as 0.5 ms at an exposure rate of

::; 300 1\-f�
up to 80 per second and a power output

::::! 200 1\..

of up to 150 kW. The main application of
grid-controlled x-ray tubes is in cinefluo­

100 "'
rography with maximum rnA of about 400
to 600 rnA. Exposure rates as high as 500 .01 .05 1 .2 .3 .5 1 2 5 10
per second are possible. Grid controlled x­ EXPOSURE TIME (sec)
ray tubes are generally not appropriate for
Figure 3-27 Theoretic x-ray tube rating chart
the high-milliampere loading required for at 70 kVp
angiography using regular rapid filming
techniques. (shown by the shaded areas in Fig. 3-28).
In the final analysis, few physicists and The operator does not set an exposure
no radiologists care how the switching is time, but tells the generator to produce 200
done so long as it meets the specifications mAs at 70 kVp. The exposure will begin
of the generator. at 600 rnA, the highest rnA available in this
example, then will drop the tube current
FALLING LOAD GENERATORS to 500 rnA when the 70-kVp line crosses
The purpose of a falling load generator the 600-mA line . Corresponding reduc­
is to produce an x-ray exposure in the tions of tube current will occur at 400 and
shortest possible exposure time by oper­ 300 rnA until the required 200 mAs has
ating the x-ray tube at its maximum kil­ been accumulated. In our hypothetical sit­
owatt rating during the entire exposure. uation, the 200 mAs will be composed of
Note that falling-load generators are not 600 mA x .05 sec = 30 mAs
used to give very short exposure times; 500 mA x .15 sec = 75 mAs
they give shorter exposure times than 400 mA x .10 sec = 40 mAs
300 mA x .20 sec 60 mAs
would be obtained with a fixed-tube rnA

Total 0.50 sec 205 mAs

Let us go directly to an example to ex­ By operating the x-ray tube at its maximum
plain how this generator functions. Look tolerance, the falling load generator has
at Figure 3-27, the theoretical chart for an
x-ray tube operating at 70 kVp. If an ex­
posure of 70 kVp and 200 mAs is desired,
� 600 F=====--L
this x-ray tube could be operated at 70
kVp, 200 rnA, and 1.0 sec (200 mAs) to
ffi 500
produce such an exposure. Notice that any � 400

attempt to get a shorter exposure would :::::i 300
fail, because the tube is limited to 0.5 sec � 200
at 300 rnA (150 mAs), 0.3 sec at 400 rnA
(120 mAs), etc. Thus, we must accept an
exposure time of at least 1.0 sec to get 200
.05 .2 .3 .5 1
Now let us examine how the falling load
principle allows us to use the higher rnA Figure 3-28 Principle of the falling load gen­
settings to obtain 200 mAs in a shorter time erator

produced 200 mAs in 0.5 sec, as opposed ing circuit. We leave the exact details of
to 1.0 sec for a fixed rnA technique. these structures to the engineers.
There are some problems with the fall­
ing load principle. Operating the tube at Automatic Exposure Control
high rnA causes maximum focal spot (Phototimer)
blooming. Heating the anode to its maxi­ Mechanical and electronic timers are
mum capacity with each exposure shortens subject to human error. The operator se­
x-ray tube life. Manufacturers generally set lects the exposure time that he believes will
the tube to operate at somewhat lower than produce a film of the desired density. He
maximum loading . can measure the patient's thickness, but
One additional consideration must be must estimate the density of the tissues that
mentioned. When rnA is lowered, there is will be included in the radiographic field.
an automatic rise in kV if some form of kV If his estimate is incorrect, the resulting
compensation is not used. Modern falling radiograph will be improperly exposed.
load generators provide an almost constant Automatic exposure controls, also called
kV. phototimers, have been developed to elim­
Falling load generators are expensive. inate human error. They measure the
They find their greatest use with automatic amount of radiation required to produce
exposure generators where simple opera­ the correct exposure for a radiographic ex­
tor controls are desired. amination. It is first necessary to select a
kilovoltage that will produce satisfactory
EXPOSURE TIMERS penetration of the part to be examined.
Once the kV p is selected, either the tech­
A variety of ways to control the length
nologist or a phototimer must select the
of an x-ray exposure have been developed.
mAs that will produce a proper exposure.
We will mention these briefly, and consider
The goal is to produce a satisfactory radi­
only phototimers in any detail.
ograph with each attempt, and to repro­
There are four basic types of exposure
duce this radiograph reliably each time an­
other examination is required.
1. Mechanical timers (rarely used today)
The essential element in phototimers is
2. Electronic timers
a device that can detect radiation and, in
3. Automatic exposure control (photo­
response to this radiation, produce a small
electric current. There are three such de­
4. Pulse-counting timers
1. Photomultiplier detectors
Electronic Timers
2. Ionization chambers
Electronic timers all work on the same 3. Solid-state detectors
principle. The length of the x-ray exposure Phototimers can be located in front of the
is determined by the time required to cassette, and are called entrance types, or
charge a capacitor through a selected re­ behind the cassette as exit types.
sistance. The exposure button starts the ex­ Photomultiplier Phototimers. This is
posure and also starts charging the capac­ the most common type of automatic ex­
itor. The exposure is terminated when the posure control (Fig. 3-29). The detector is
capacitor is charged to a value necessary to made of lucite, which is a material that can
turn on associated electronic circuits. The transmit light. The lucite is coated with one
exposure time is therefore determined by or more (commonly three) areas of a phos­
the length of time for the capacitor to phor that will emit light when irradiated
charge, and this time can be varied by vary­ with x rays (these lucite detectors are usu­
ing the value of the resistance in the charg- ally called lucite paddles). Each phosphor

LUCITE phor absorbs a higher percentage of the x­

PADDLE ray beam than it does at higher kVp. This
may cause the phototimer to terminate a
.....--+ FLUORESCENT low kVp exposure before the film has been
SCREEN adequately exposed. This problem is usu­
ally handled automatically. Sensors detect
the kVp being used, and the phototimer
detector will be decreased in sensitivity if
low kVp is being used.
A photomultiplier tube is a vacuum tube.
It consists of a photocathode, several in­
GENERATOR termediate electrodes (called "dynodes"),
and an anode. When light strikes the pho­
Figure 3-29 Photomultiplier automatic ex­ toemissive layer on the photocathode, the
posure control (phototimer)
photoemissive material emits photoelec­
trons in numbers proportional to the in­
area is about 100 square centimeters ( 100 tensity of the light. The intermediate elec­
cm2) in size. When a phosphor generates trodes are coated with a material that emits
light, the intensity of the light is obviously secondary electrons when struck by an­
proportional to the intensity of x rays that other electron. The electrons are acceler­
reached the phosphor. The lucite transmits ated by a positive potential from one dy­
this light to an output region called a "light node to the next, with each dynode giving
gate." The light gate directs the light to the rise to more electrons. Photomultiplier
photocathode of a photomultiplier tube, tubes come in many sizes and shapes, and
where the light is converted to an electric are used extensively in nuclear medicine
current that is amplified to produce an and physics laboratories. The final number
electrical signal. The electric current gen­ of electrons collected at the anode repre­
erated by the photomultiplier tube may be sents the output current, and this output
used to charge a capacitor. When the ca­ current is proportional to the intensity of
pacitor reaches a predetermined charge it the light that struck the photocathode. In­
can be used to bias the gate of a thyrister herent in the operation of a photomulti­
in the x-ray circuit and cause the exposure plier tube is the requirement for a stable
to terminate. Photomultiplier phototimers high-voltage supply. There is some insta­
may be used as entrance or exit types. The bility in photomultiplier tubes.
lucite paddle is almost completely radio­ Ionization Chamber Autotimers. Ioni­
lucent, so it will not produce a detectable zation chambers are almost always used as
image if placed in front of a cassette. The entrance-type autotimers. An ionization
lucite serves two functions: it is the support chamber is very slightly imaged on film, but
that holds the fluorescent screen or the image is so faint it is lost in the images
screens, and it transmits light to the pho­ of anatomic parts.
tomultiplier tube so that the photomulti­ Ionization chambers come in many
plier tube may be kept out of the x-ray field shapes and sizes; the one we describe is
(the photomultiplier tube would produce used in automatic timers. An ionization
an image on the radiograph). chamber consists of two thin parallel sheets
Photomultiplier phototimer response is of aluminum or lead foil. Gas (usually room
a function of kVp because low energy pho­ air) is present between the plates. The gas
tons are absorbed more readily in the phos­ becomes ionized when struck by radiation,
ppor. This presents a problem. At lower and ionization is in direct proportion to the
peak kilovoltage (about 60 kVp) the phos- amount of radiation. Prior to exposure a

charge is placed on the parallel plates by same areas (e.g., 50 cm2) but with different
applying a voltage across them. This is ex­ shapes. For example, a chest unit might
actly like charging a capacitor. In fact, the have a rectangular-shaped detector in its
structure we have described is a parallel center (in the region of the mediastinum)
plate capacitor. When the gas is ionized by and two more laterally placed square de­
radiation, the negative ion (electron) moves tectors. The technologist can choose to use
toward the positive plate and the positive one, two, or all three detectors. Use of the
ion (ionized gas atom) moves toward the central field may permit a good view of the
negative plate. When the ions arrive at the thoracic spine but overexpose the lung
plates they neutralize part of the charge fields. The lateral detectors will serve the
that has been placed on the plates. This lung fields well unless one lung happens to
reduction in charge reduces the residual be opaque. Obviously, correct alignment of
voltage across the plates. When the voltage body parts and detector fields is important.
has been reduced to a previously deter­ In addition, collimation is important be­
mined value, an electronic circuit is acti­ cause scatter radiation reaching a detector
vated that will terminate the exposure. may cause premature termination of an ex­
Solid-State Autotimers. Solid-state de­ posure.
vices are becoming more common in au­ When phototimers are used it is essential
tomatic timers. A variety of solid-state ra­ that the phototimer be adjusted for the
diation detectors are on the market today film, intensifying screen, and cassette being
that operate on the basis of radiation-pro­ used. Any variation in film or screen speed
ducing ionization in or near a PN structure. or cassette absorption will result in a
Other devices work on the basis of pho­ change in film density. The Department of
toconductivity and could also be used. Our Health and Human Services requires that
mention of these two types is an indication a phototimer show variation of no more
of how fluid solid-state technology is. We than 10% from exposure to exposure.
are told by our equipment engineer that at Automatic timers have density control
the present time most solid-state autotim­ buttons that can increase or decrease the
ers use PN junction techniques. Solid-state sensitivity of the detectors.
devices offer the advantage of small size, Automatic timers are equipped with a
almost no x-ray beam absorption, and a backup timer that will act as a safety factor
consistent, rapid response. An informal to terminate an exposure in case of equip­
poll of several of our local equipment sup­ ment failure. The Department of Health
pliers indicated that photomultiplier tubes and Human Services regulations state that
and ionization chambers are still in the ma­ the generator must automatically termi­
jority at this time. nate an exposure at 600 mAs for 50 kV or
greater, and at 2000 mAs for kV under 50.
Miscellaneous Autotimer Topics
Use of automatic timer devices requires Pulse-Counting Timers
attention to a number of details; we will A technique used extensively in meas­
use this section to discuss some of these. uring time is to measure the occurrence of
The number, size, and position of the a periodic event. For example, a most ac­
radiation detector(s) is important. There curate clock, the atomic clock, counts the
may be one, two, or three detectors; the very stable frequency of atomic oscillation.
three-field automatic timer is most popular. In the electronic world, timing is accom­
A single detector positioned in the center plished by counting pulses of a regular pe­
of the film is commonly used in fluoro­ riodic voltage. In fact, the periodic pulse
scopic spot film devices. Three-field de­ train is called the clock. When it is neces­
vices have three detectors, usually with the sary to time exposures of a few millisec-

onds, a convenient way to measure such a lector, and a step-down transformer. The
time would be to count voltage pulses of cathode-anode circuit, called the high-volt­
high frequency. Obviously, the higher the age circuit, contains an autotransformer
frequency, the more pulses that are and a step-up transformer. The autotrans­
counted for a given exposure period. The former serves as the kVp selector.
higher frequency has a tendency to reduce The incoming electrical supply to the x­
the error in the time period. ray generator has an alternating potential.
Pulse-counting timers use this technique Rectifiers are devices (usually silicon di­
of voltage pulse counting to control the odes) that transmit a current in only one
time of short exposure techniques. High direction. Single-phase generators may
frequencies can be generated by the oscil­ have half-wave rectification (60 pulses per
lation in a quartz crystal whose frequency second). Three-phase generators may be
of oscillation is determined by the size, ori­ six-pulse (360 pulses per second) with a
entation of the cut of the crystal, and the theoretical ripple factor of 1 3.5%, or
mode of oscillation (technical details that twelve-pulse (720 pulses per second) with
we will not investigate further; we will, a ripple factor of 3.5%. Special types of
however, discuss piezoelectric oscillation in generators include capacitor-discharge
Chapter 20 on ultrasound). With a quartz generators, battery-powered generators,
crystal it is quite easy to obtain megahertz medium-frequency generators, and falling
frequencies that are extremely stable. As load generators.
an example, assume the crystal oscillates at Transformers are given a rating that in­
1,000,000 Hz. A 1 O-ms time could be meas­ dicates the maximum safe output of the
ured by counting 10,000 pulses. The large secondary windings. Such ratings are ex­
number of pulses accounts for the accu­ pressed as the kilowatt rating.
racy, because a small error of plus or minus Exposure switching may be primary

a few pulses will make no significant dif­ switching or secondary switching. Almost

ference in exposure time. all general purpose generators use primary

switching. Secondary switching requires
use of triode vacuum tubes or grid-con­
trolled x-ray tubes, and is used for fast,
An x-ray generator supplies electrical repetitive exposures.
energy to the x-ray tube and regulates the Exposure timers include mechanical tim­
length of the radiographic exposure. The ers (obsolete), electronic timers, automatic
x-ray tube requires two sources of energy, exposure controls (phototimers), and
one to heat the filament and the other to pulse-counting timers. Automatic expo­
accelerate electrons between the cathode sure control may be achieved with photo­
and anode. The filament circuit contains a multiplier detectors, ionization chambers,
variable resistance, which is the current se- or solid-state detectors.

4 Basic Interactions
Between X Rays
and Matter

Atoms are bonded into molecules by photons are scattered they are deflected
electrons in the outermost shell. X-ray pho­ into a random course, and no longer carry
tons may interact either with orbital elec­ useful information . Because their direction
trons or with the nucleus of atoms. In the is random, they cannot portray an image,
diagnostic energy range, the interactions and the only thing they produce on a film
are always with orbital electrons. If these is blackness. In the language of the infor­
electrons happen to be the ones bonding mation transfer theorist, scatter radiation
atoms into molecules, the bonds may be adds noise to the system. You can gain
disrupted and the molecular structure al­ some insight into the effect of noise with a
tered . The molecular bonding energies, simple analogy. Imagine yourself listening
however, are too small to influence the type to music on your car radio as you drive past
and number of interactions. A group of an airport. As a plane approaches the mu­
oxygen atoms will stop the same number sic becomes more difficult to hear until fi­
of x-ray photons, regardless of their phys­ nally, when the plane is directly overhead,
ical state. It does not matter if the oxygen it becomes completely inaudible . Even
is free as a gas or is bound to hydrogen as though the music is still there, you cannot
water. The important factor is the atomic hear it. Radiation noise does exactly the
makeup of a tissue and not its molecular same sort of thing. It destroys image qual­
strucque. ity. We refer to the noise created by scatter
There are five basic ways that an x-ray radiation as "film fog ." When an x-ray film
photon can interact with matter. These are: is badly fogged, the image may be com­
1. Coherent scattering pletely obscured. It is still there, but we
2. Photoelectric effect cannot see it. An understanding of the pro­
3. Compton scattering duction of scatter radiation is of vital im­
4. Pair production portance, so pay particular attention to it
5. Photodisintegration as we discuss the basic interactions between
x rays and matter.
As we discuss each of these interactions,
you will see that at times x-ray photons are COHERENT SCATTERING
absorbed, while at other times they are The name "coherent scattering" is given
merely scattered. When photons are ab­ to those interactions in which radiation un­
sorbed, they are completely removed from dergoes a change in direction without a
the x-ray beam and cease to exist. When change in wavelength. For this reason, the


term "unmodified scattering" is sometimes ing is small compared to that of the other
used. There are two types of coherent scat­ basic interactions; in general, it is less than
tering, Thomson scattering and Rayleigh 5%. Some coherent scattering occurs
scattering. In Thomson scattering a single throughout the diagnostic energy range,
electron is involved in the interaction. Ray­ but it never plays a major role. Even
leigh scattering results from a cooperative though it produces scattered radiation,
interaction with all the electrons of an which contributes to film fog, the total
atom. Both types of coherent scattering quantity is too small to be important in di­
may be described in terms of a wave-par­ agnostic radiology.
ticle interaction, and are therefore some­
times called "classical scattering." Rayleigh PHOTOELECTRIC EFFECT
scattering is shown in Figure 4-1. Low-en­ Before beginning a discussion of the
ergy radiation encounters the electrons of photoelectric effect, we are going to digress
an atom and sets them into vibration at the a moment to review a few points about
frequency of the radiation. A vibrating atomic physics. Atomic structure is ex­
electron, because it is a charged particle, tremely complex, but we can greatly sim­
emits radiation. The process may be envi­ plify it and still have a fairly good idea of
sioned as the absorption of radiation, vi­ how these interactions occur. The atom
bration of the atom, and emission of ra­ consists of a central nucleus and orbital
diation as the atom returns to its electrons. The nucleus is important in our
undisturbed state. This is the only type of present discussion only as a means of keep­
interaction between x rays and matter that ing the electrons in the atom. The posi­
does not cause ionization. To produce an tively charged nucleus holds the negatively
ion pair, a photon must transfer energy to charged electrons in specific orbits, or
the atom. No energy is transferred, and no shells. The innermost shell is called the K
ionization occurs with coherent scattering. shell, and the more peripheral shells are
Its only effect is to change the direction of named consecutively L, M, N, and so forth.
the incident radiation. The percentage of These shells have a limited electron capac­
radiation that undergoes coherent scatter- ity. The K-shell can hold only two electrons.
If more electrons are present in the atom
they must move out to the L shell, which
has a capacity of eight electrons. Each shell
has a specific binding energy. The closer
the shell is to the nucleus, the tighter the
electrons in that shell are bound to the nu­
cleus. The electrons in the outermost shell
are loosely bound. They are essentially free
and, appropriately, are called "free elec­
trons." The energy value of electronic
shells is also determined by the atomic
number of the atom. K-shell electrons are
more tightly bound in elements with high
atomic numbers than they are in elements
with low atomic numbers. In fact, the dif­
ferences are sizable. For example, the K­
shell binding energy for lead is 88 ke V,
while the K-shell binding energy for cal­
cium is only 4 keV. Electrons in the K shell
Figure 4-1 Coherent Rayleigh scattering are at a lower energy level than electrons

in the L shell. If we consider the outermost des have little penetrating power. The
electrons as free, than inner shell electrons atom is left with an electron void on the K
are all in energy debt. Their energy debt shell but only for an instant, because an
is greatest when they are close to the nu­ electron immediately drops into the void
cleus in an element with a high atomic to fill the K shell. This electron usually
number. comes from the adjacent L shell, occasion­
The photoelectric effect is shown in Fig­ ally from the M shell and, on rare occa­
ure 4-2. We will quickly run through the sions, from free electrons from the same,
whole interaction and then fill in some de­ or another, atom. As an electron drops into
tails later. An incident photon, with a little the K shell it gives up energy in the form
more energy than the binding energy of a of an x-ray photon. The amount of energy
K-shell electron, encounters one of these is characteristic of each element, and the
electrons and ejects it from its orbit. The radiation produced by the movement of
photon disappears, giving up all its energy electrons within an atom is called "char­
to the electron. Most of the photon's energy acteristic radiation." When the K-shell void
is needed to overcome the binding energy is filled by an outer shell electron from the
of the electron, and the excess gives the same atom, the atom is left with a deficiency
electron kinetic energy. The electron, of one electron, and it remains a positive
which is now free of its energy debt, flies ion. If a free electron from another atom
off into space as a photoelectron. It be­ fills the void, then the other atom becomes
comes a negative ion and is absorbed al­ a positive ion, and the result is the same.
most immediately, because charged parti- The photoelectric effect always yields three
end products: (1) characteristic radiation;
(2) a negative ion (the photoelectron); and
(3) a positive ion (an atom deficient one

Probability of Occurrence

Three simple rules govern the probabil­

it y of a photoelectric reaction.
I. - .
1. The incident photon must have suf­
ficient energy to overcome the binding en­
. .

--- ...... ergy of the electron. For example, the
// ...... - -- .....
.e '
' , , K-shell electrons of iodine have a binding
\ energy of 33.2 keV. An x-ray photon with
''®N ''
I \ ;I an energy of 33.0 keV absolutely cannot
\ ' , ..,.... • I
' ..... --- _,/
eject them from their shell. The photon
/ may interact at the L or M shells but not

• Characteristic at the K shell.

Radiation 2. A photoelectric reaction is most

likely to occur when the photon energy
,•· -.
and electron binding energy are nearly
I ,•
,.---.... , ,,
' \ the same, provided of course that the pho­
II N \
ton's energy is greater. A 34-keV photon is
J 0

\ , ® I'
' e
I I much more likely to react with a K-shell
' ...... __ ......... /
' electron of iodine than a 100-keV photon .
/ 3.

Po sitive lon In fact, the probability of a photoelectric

reaction drops precipitously as photon en­
Figure 4-2 Photoelectric effect ergy increases. It is inversely proportional

to approximately the third pow_er of en­ "free" (not bound to an atom) electron. If
ergy, or, expressed in mathematiCal form: you like, you may call this a "forbidden"
Photoelectric effect -
Characteristic Radiation
3. The tighter an electron is bound in
We first mentioned characteristic radia­
its orbit, the more likely it is to be involved
tion when we discussed the production of
in a photoelectric reaction. Elect�ons �re
x rays in Chapter 2. Characteristic radia­
more tightly bound in elements with h1gh
tion generated by the photoele.ctric effe�t
atomic numbers than in elements with low
is exactly the same. The only difference Is
atomic numbers, as shown in Table 4-1.
in the method used to eject the inner shell
Most interactions occur at the K shell in
electron. In an x-ray tube a high-speed
elements with low atomic numbers, be­
electron ejects the bound electron, while in
a photoelectric interaction an x-ray p �oton
cause the K shell contains the most tightly
bound electrons. In elements with high
does the trick. In both cases the atom IS left
atomic numbers, however, the energy of
with an excess of energy equal to the bind­
the incident photons is frequently insuffi­
ing energy of the ejected electro?. All phys­
cient to eject a K-shell electron, and many
ical systems seek the lowest possible energy
photoelectric reactions take place at the L­
state. For example, water always runs
and M-shell levels. Because elements with
downhill. An atom with an electron defi­
high atomic numbers bind th�ir electrons
ciency in the K shell is in a higher energy
more tightly, they are more hkely to un­
dergo photoelectric reac.tions. T �e proba­
state than an atom with an L-shell electron
deficiency. The atom releases this excess
bility of the photoelectnc effect mcreases
energy in the form of photons . Usuall� an
sharply as the atomic number increases . In
. electron from an adjacent shell drops mto
fact, it is roughly proportional to the third
an inner shell void, as shown for an iodine
power of the atomic number:
atom in Figure 4-3 . The K-shell binding
Photoelectric effect - (atomic number)3 energy for iodine is 33.2 keV, while the L­
shell binding energy is 4.9 keV, and more
In summary, photoelectric reactions are
peripheral shells have even lower binding
most likely to occur with low energy pho­
energies. When an electron from an L shell
tons and elements with high atomic num­
falls to the K shell, a 28.3-keV (33.2 - 4.9
bers provided the photons have sufficient .
= 28.3) photon is released The void in the
energy to overcome the forces binding the
L shell is then filled with a photon from
electrons in their shells. In fact, the pho­
the M shell with the production of a 4.3-
toelectric effect cannot take place with a
Table 4-1. K-Shell Electron Binding E � ergies
of Elements Important in Diagnostic Radiology N M l K
I 1 I I
33.2 keV

6 Carbon 0.284
\' \
, 4.9keV ',
\ '-•

\ \ ', �28.3 keV

7 Nitrogen 0.400 \ 0.6 k�V --•

8 Oxygen 0.532
13 Aluminum 1.56 o\ev ',,,_ � 4.3 keV
f 0.6 keV
20 Calcium 4.04
Tin 29.2 '',,,..._
53 Iodine 33.2
-----! "--J -
TOTAL 33.2 keV
56 Barium 37.4
74 Tungsten 69.5
Figure 4-3 Characteristic radiation from io­
82 Lead 88.0

keV photon. This process is continued un­ pends on the third power of the atomic
til the whole 33.2 keV of energy has been number, the photoelectric effect magnifies
converted into photons. Of course, the the difference in tissues composed of dif­
lower energy photons are of no importance ferent elements, such as bone and soft tis­
in diagnostic radiology, because they are sue. So, from the point of view of film qual­
absorbed immediately. Occasionally an in­ ity, the photoelectric effect is desirable.
ner shell void is filled by an electron from From the point of view of patient exposure,
distant peripheral shells, or even by free though, it is undesirable. Patients receive
electrons. When a free electron moves into more radiation from photoelectric reac­
the K-shell void in iodine, a single photon tions than from any other type of inter­
of 33.2 keV is generated. This is the most action. All the energy of the incident pho­
energetic characteristic radiation that io­ ton is absorbed by the patient in a
dine can produce. photoelectric reaction. Only part of the in­
The K-shell binding energies of some el­ cident photon's energy is absorbed in a
ements important in diagnostic radiology Compton reaction, as you will see in the
are shown in Table 4-1. Calcium, which next section. Because one of our primary
has the highest atomic number of any el­ goals is to keep patient doses at a minimum,
ement found in the body in significant we must keep this point in mind at all times.
quantity, emits a 4-keV maximal energy The importance of the photoelectric effect
characteristic photon, which is little energy can be minimized by using high-energy
by x-ray standards. It is absorbed within a (kVp) techniques. In general, we should
few millimeters of its site of origin. The use radiation of the highest energy consist­
contrast agents iodine and barium are the ent with that of diagnostic quality x-ray
only elements encountered in diagnostic films to minimize patient exposure.
radiology that emit characteristic radiation In summary, the likelihood of a photo­
energetic enough to leave the patient and electric interaction depends on two factors:
fog an x-ray film. the energy of the radiation and the atomic
Characteristic radiation is usually re­ number of the absorber. The reaction is
ferred to as "secondary radiation" to dif­ most common with low energy photons
ferentiate it from scatter radiation, a dis­ and absorbers with high atomic numbers.
tinction that hardly seems necessary Photons usually strike tightly bound K­
because the end result is the same for both, shell electrons, and they must have more
a photon that is deflected from its original energy than the electron binding energy.
path. Photoelectric reactions are desirable from
the point of view of film quality, because
Applications to Diagnostic Radiology they give excellent contrast without gen­
As diagnostic radiologists we can look at erating significant scatter radiation. Un­
the photoelectric effect in two ways, one fortunately, patient exposures are much
good and the other bad. Starting with the higher than with any other type of inter­
good, it produces radiographic images of action.
excellent quality. The quality is good for
two reasons: first, the photoelectric effect COMPTON SCATTERING

does not produce scatter radiation and sec­ Almost all the scatter radiation that we
ond, it enhances natural tissue contrast. X­ encounter in diagnostic radiology comes
ray image contrast depends on some tissues from Compton scattering. Figure 4-4 di­
absorbing more x rays than other tissues. agrams this reaction. An incident photon
Contrast is greatest when the difference in with relatively high energy strikes a free
absorption between adjacent tissues is outer shell electron, ejecting it from its or­
large. Because the number of reactions de- bit. The photon is deflected by the electron

I. Recoil Electron cue ball deflects at a greater angle and loses

more energy. Finally, with a direct hit, max­
imal energy is transferred to the second
ball but, unlike the cue ball in billiards, a
Compton photon never gives up all its en­
ergy. The photon retains some energy and
deflects back along its original path at an
angle of 180°. Figure 4-5 shows the
amount of energy a 75-keV photon retains
2. Scattered Photon
as it is deflected through various angles. As
you can see, more energy is lost with larger
angles of deflection.
To make our comparison between a
3. Positive ion
Compton reaction and billiard balls com­
plete, we must make one last adjustment in
Figure 4-4 Compton scattering the analogy. Billiard balls all have the same
mass, but we want to compare them with
so that it travels in a new direction as scatter photons whose energies vary. We can keep
radiation. The photon always retains part our analogy going by using cue balls of dif­
of its original energy. The reaction pro­ ferent masses. Thus, a lightweight cue ball
duces an ion pair, a positive atom and a simulates a low energy photon, and a heavy
negative electron, which is called a "recoil" cue ball simulates a high energy photon.
electron. Now, imagine both the lightweight and the
The energy of the incident photon is dis­ heavy cue balls striking a second ball a
tributed in two ways. Part of it goes to the glancing blow at an angle of 45°. The light­
recoil electron as kinetic energy, and the weight ball has little momentum, and it will
rest is retained by the deflected photon. deflect at a much greater angle than the
Unlike a photoelectric reaction in which heavy ball, which has more momentum.
most of a photon's energy is expended Photons also have momentum, and the
freeing the photoelectron from its bond, in higher the energy of the photons, the more
a Compton reaction no energy is needed difficult they are to deflect. With x-ray en­
for this purpose, because the recoil elec­ ergies of 1 MeV (million electron volts),
tron is already free. The incident photon most scattered photons deflect in a forward
always retains some of its original energy. direction. In the diagnostic energy range,
Two factors determine the amount of en­ however, the distribution is more symmet-
ergy the photon retains: its initial energy
and its angle of deflection off the recoil

--75 kev -- :L"--o • -- 75 keV­
electron. If you think of the Compton re­ '1''
action as a collision between two billiard 3o·

balls, the energy transfer is easier to un­ 74.3 keV

. 60•
derstand. The cue ball is the incident pho­
ton, and the second ball is a free electron.
go• �
When the cue ball strikes the second ball,
I 70 keV

they both deflect in a predictable manner.
66 keV
With a glancing blow the angle of deflec­
tion of the cue ball is small, and only a small
quantity of energy is transferred to the sec­
ond ball. The cue ball retains almost all its Figure 4-5 Energy of Compton-scattered
initial energy. With a more direct hit the photons

ric. Figure 4-6 shows the pattern of dis­ Table 4-2. Energy of Compton-Scattered
Photons for Various Angles of Deflection
tribution of scattered photons from a 100-
keV x-ray beam. The distance from the
point of the interaction is used to express PHOTON (keV) ENERGY OF SCATTERED PHOTONS (keV)
the percentage of scattered photons at var­ PHOTON DEFLECTION ANGLE
30° 60° 90° 180°
ious angles. Remember, this is a flat rep­
resentation of events that are occurring in 25 24.9 24.4 24 23
50 49.6 47.8 46 42
three dimensions. You can picture the
75 74.3 70 66 58
three-dimensional distribution by rotating
100 98.5 91 84 72
the pattern around its longitudinal axis. 150 146 131 116 95
With lower energy radiation, fewer pho­
tons scatter forward and more scatter back
at an angle of 180°. This distribution is At narrow angles of deflection, scattered

fairly constant throughout the diagnostic photons retain almost all their original en­
energy range. ergy. This creates a serious problem in di­
The classic formula for calculating the agnostic radiology, because photons that
are scattered at narrow angles have an ex­
�hange in wavelength of a scattered photon
lS cellent chance of reaching an x-ray film and
producing fog. They are exceedingly dif­
AX. 0.024 (1 cos 6)
ficult to remove from the x-ray beam. In
= -

AX. = change in wavelength (A) fact, they cannot be removed by filters be­
a = angle of photon deflection cause they are too energetic, and they can­

Most of us think in terms of kiloelectron not be removed by grids because their an­

volts and not wavelength. We can change gle of deflection is too small. Because we

wavelength from the above formula to kil­ have no way of removing them from the

ovoltage energy with the following conver­ useful beam, we must accept them and tol­

sion factor: erate an image of diminished quality.

Scatter radiation from Compton reac­
tions is also a major safety hazard. As you

can see in Table 4-2, even after a photon
A. = wavelength (A) has been deflected 90°, it still retains most
keV = energy of photon of its original energy. This means that the
scatter radiation that arises in the patient
The amount of energy retained by scat­
during a fluoroscopic examination is al­
tered photons after a Compton reaction is
most as energetic as the primary beam. It
shown in Table 4-2. In the diagnostic en­
creates a real safety hazard for the fluo­
ergy range up to 150 keV, the photon re­
roscopist and other personnel who must be
tains most of its original energy, and very
m exposure rooms .
little is transferred to the recoil electron.

Probability of Occurrence
The probability of a Compton reaction
depends on the total number of electrons
100 keV
in an absorber, which in turn depends on
-t- ___ __;
�� :::.------------- · --------------- - --- ----
its density and the number of electrons per
gram. In Chapter 5 we will show that all
elements contain approximately the same
number of electrons per gram, regardless
Figure 4-6 Distribution of Compton-scat­ of their atomic number. Therefore, the
tered photons number of Compton reactions is inde-

pendent of the atomic number of the ab­ ficient energy to overcome nuclear binding
sorber. The likelihood of a reaction, how­ energies of the order of 7 to 15 MeV.
ever, does depend on the energy of the Because pair production does not occur
radiation and the density of the absorber. with photon energies less than 1.02 MeV,
The number of reactions gradually dimin­ and photodisintegration does not occur
ishes as photon energy increases, so that a with energies less than 7 MeV, neither of
high energy photon is more likely to pass these interactions is of any importance in
through the body than a low energy pho­ diagnostic radiology, where we rarely use
ton. energies above 150 keV.
Because Compton reactions occur with
free electrons, we must define this term a RELATIVE FREQUENCY OF BASIC

little more precisely. An electron can be INTERACTIONS

considered free when its binding energy is Figure 4-7 shows the percentage of each
a great deal less than the energy of the type of interaction for water, compact
incident photon. With the photon energies bone, and sodium iodide, with photon en­
used in diagnostic radiology (10 to 150 ergies ranging from 20 to 100 keV. The
keV), only the outer shell electrons are free total number of reactions is always 100%.
in the elements with high atomic numbers. The contribution of each interaction is rep­
For the elements with low atomic numbers, resented by an area in the illustration.
those found in soft tissues, all electrons can Thus, if coherent scattering accounts for
be considered free, because even those on 5% of the interactions, Compton scattering
the K shell are bound with an energy of for 20%, and the photoelectric effect for
less than 1 keV. 75%, the total is 100%. Water is used to
illustrate the behavior of tissues with low
atomic numbers, such as air, fat, and mus­
cle. The total number of reactions is less
for air than for water, but the percentage
The last two basic interactions, pair pro­ of each type is approximately the same.
duction and photodisintegration, do not Compact bone contains a large amount of
occur in the diagnostic energy range. They calcium, and it represents elements with
have no importance in diagnostic radiol­ intermediate atomic numbers. Iodine and
ogy, and we will only discuss them briefly. barium are the elements with the highest
In pair production, a high energy photon atomic numbers that we encounter in di­
interacts with the nucleus of an atom, the agnostic radiology, and they are repre­
photon disappears, and its energy is con­ sented by sodium iodide.
verted into matter in the form of two par­ As you can see in Figure 4-7, coherent
ticles. One is an ordinary electron and the scattering usually contributes around 5%
other is a positron, a particle with the same to the total, and plays a minor role
mass as an electron but with a positive throughout the diagnostic energy range.
charge. Because the mass of one electron In water, Compton scattering is the dom­
is equal to 0.51 MeV, and pair production inant interaction, except at very low photon
produces two electron masses, the inter­ energies (20 to 30 keV). The contrast
action cannot take place with photon en­ agents, because of their high atomic num­
ergies less than 1.02 MeV. bers, are involved almost exclusively in
In photodisintegration, part of the nu­ photoelectric reactions. Bone is interme­
cleus of an atom is ejected by a high energy diate between water and the contrast
photon. The ejected portion may be a neu­ agents. At low energies, photoelectric re­
tron, a proton, an alpha particle, or a clus­ actions are more common, while at high
ter of particles. The photon must have suf- energies, Compton scattering is dominant.



60 100 20 60 100

keV keV keV


Figure 4-7 Percentage of coherent, photoelectric, and Compton reactions in water, compact
bone, and sodium iodide

SUMMARY sorbers. It generates no significant scatter

radiation and produces high contrast in the
Only two interactions are important in x-ray image but, unfortunately, exposes
diagnostic radiology, the photoelectric ef­ the patient to a great deal of radiation. At
fect and Compton scattering. Coherent higher diagnostic energies, Compton scat­
scattering is numerically unimportant, and tering is the most common interaction be­
pair production and photodisintegration tween x rays and body tissues, and is re­
occur at energies above the useful energy sponsible for almost all scatter radiation.
range. The photoelectric effect is the pre­ Radiographic image contrast is less with
dominant interaction with low energy ra­ Compton reactions than with the photo­
diation and with high atomic number ab- electric effect.

5 Attenuation

Quantity and quality are two terms used directed at a water phantom. The intensity
to express the characteristics of an x-ray of the beam is decreased to 800 photons
beam. Quantity refers to the number of by the first centimeter of water, which is an
photons in the beam, and quality refers to attenuation of20%. The second centimeter
their energies. The intensity of a beam is of water decreases the intensity to 640 pho­
the product of the number and energy of tons, which is 20% less than had passed
the photons, so it depends on both quantity through the first centimeter. With each suc­
and quality. A beam of 50-keV photons has ceeding centimeter of water, 20% of the
a greater intensity than a beam made up remaining photons are removed from the
of a comparable number of 20-keV pho­ beam.
tons. Attenuation is the reduction in the The quality of monochromatic radiation
intensity of an x-ray beam as it traverses does not change as it passes through an
matter by either the absorption or deflec­ absorber. Remember, we are only discuss­
tion of photons from the beam. Because it ing primary photons, the photons that, at
is a measure of a change in x-ray intensity, most, have only one interaction. If we begin
attenuation depends on both the quantity with a group of photons with an energy of
and quality of the photons in a beam. For 60 keV, then the transmitted photons will
our purpose, it would be simpler to think all have the same energy. Their numbers
of attenuation only in terms of quantity . will be reduced, but their quality will not
Then a reduction in intensity would merely be changed. So, when we talk about a
be a reduction in the number of photons change in the intensity of a monochromatic
in the beam. We can do this if we limit our beam, we are really talking about a change
discussion to the attenuation of monochro­
matic radiation. Later in the text, we will WATER
clarify why quality does not change with
monochromatic radiation . To further sim­
plify our discussion, we will disregard the __ _ -----1 � � � � f---4_I_
O _uDetector
Photons Photons
attenuation of secondary radiation result­
ing from coherent or Compton scattering
and the characteristic radiation of the pho­
toelectric effect. Once a photon has un­
dergone any of these reactions, it will be
considered completely attenuated-that is,
1000 410
totally removed from the beam . Photons Photons

-20% -20% -20% -20%
The attenuation of a beam of mono­
chromatic radiation is shown schematically Figure 5-1 Attenuation of monochromatic ra­
in Figure 5-l. A beam of 1000 photons is diation


in the number, or quantity, of photons in centage with each increment of absorber,

the beam. A 50% reduction in the number as with monochromatic radiation, the at­
of photons is a 50% reduction in the in­ tenuation is called exponential. Exponen­
tensity of the beam. The attenuation of tial functions plot a straight line on semi­
polychromatic radiation is much more logarithmic graph paper.
complicated, and we will not discuss it until
the end of the chapter. ATTENUATION COEFFICIENTS
When the number of transmitted pho­ An attenuation coefficient is a measure
tons and absorber thickness are plotted on of the quantity of radiation attenuated by
linear graph paper, a curved line results a given thickness of an absorber. The name
(Fig. 5-2A). The initial portion of the curve of the coefficient is determined by the units
is steep, because more photons are re­ used to measure the thickness of the ab­
moved from the beam by the first few cen­ sorber. Four attenuation coefficients are
timeters of absorber. After the beam has described in classic texts on radiologic
passed through many centimeters of water, physics but only two are important in di­
only a few photons remain. Although each agnostic radiology, the linear and mass at­
centimeter continues to remove 20% of the tenuation coefficients.
photons, the total numbers are small, and
the end of the curve is almost flat. The Linear Attenuation Coefficient
same numbers plot a straight line on semi­ The linear attenuation coefficient is the
logarithmic graph paper (Fig. 5-2B). most important coefficient for diagnostic
When the number of photons remaining radiology. It is a quantitative measurement
in the beam decreases by the same per- of attenuation per centimeter of absorber,

A 8
1,000 1,000
Linear Graph Paper Semi-Log Graph Paper

0 I 200
..r:::. 1
Q.. 6oo L
Q) 100 ,_


E 400
so l
.... I



0 1 4 8 12 16 20
4 8 12 16 20

Cm. of Water Cm. of Water

Figure 5-2 Attenuation of monochromatic radiation plotted on linear (A) and semilogarithmic
graph paper (B).

so it tells us how much attenuation we can with 100 units of activity, at the end of8.04
expect from a certain thickness of tissue. days we would have 50 units, and at the
Because we measure our patients in cen­ end of another 8.04 days we would only
timeters, it is a practical and useful atten­ have 25 units. The equation for this ex­
uation coefficient. In the next section, after ponential decay is
we have discussed the mass attenuation co­
efficient, you will understand why the lin­
ear coefficient is more useful. Its symbol is N = activity of remaining radionuclide
N0 = activity of original radionuclide
the Greek letter J.L.
e = base of natural logarithm (2.718)
The unit of the linear attenuation coef­ )1. = decay constant (seconds -•)
ficient is per centimeter, and thus the name t = time (seconds).

linear attenuation coefficient, because a

The decay constant (A.) can be easily related
centimeter is a linear measurement. The
to half-life (T 1�), which is a more familiar
expression per em is the same as 1/cm, and
concept. After the elapsed time of one half­
is usually written cm-1.
life (T1�), the quantity of remaining activity
It is important to realize that the linear
is N0/2. The exponential decay equation
attenuation coefficient (!-1) is for mono­
can be rewritten as
chromatic radiation, and that it is specific
both for the energy of the x-ray beam and
for the type of absorber. Water, fat, bone,
and air all have different linear attenuation
Dividing both sides by N0, the equation be­
coefficients, and the size of the coefficient
changes as the energy of the x-ray beam
changes. When the energy of the radiation
is increased the number of x rays that are
We would then go to a table of exponential
attenuated decreases, and so does the lin­
values and find the value of the exponent
ear attenuation coefficient.
A.T1�, which makes e-AT112 equal to 1;;; or,
The mathematics involved in the deri­
more simply, we could just look up the log­
vation of the exponential equation for x­
arithm of 1;;; in a table of natural logarithms.
ray attenuation is beyond the scope of this
We would find this value to be - 0.693. We
book. The formula is
can now rewrite the exponential decay

N = number of transmitted photons 0·693

N0 = number of incident photons )1, • T,, = 0.693 or A. =

e = base of natural logarithm

f.L = linear attenuation coefficient
x = absorber thickness (in centimeters) The decay constant (A.) for any radio­
nuclide can be calculated if we know its
This is a classic equation, and it has the half-life. For example, the decay constant
same format as the equation for the decay for 131I (Tl.-<2 8.04 days) is

of radionuclides. In fact, most people find

it easier to think in terms of radionuclide )1. = = 0.086/day
8.04 days
decay, so we will digress for a moment. Ra­
dionuclides decay at an exponential rate. The attenuation of monochromatic ra­
This rate is usually measured and defined diation is exactly the same. All we need
in terms of half-life, which is the time that do is substitute thickness (em of absorber)
it takes a material to decay to one half its for time (seconds). The names are
original activity. For example, 131 I has a changed, but the concept remains the
half-life (T1�) of 8.04 days. If we started same. The decay constant (A.) becomes the

linear attenuation coefficient (JJ.), and half­ ments.1 With the linear attenuation coef­
life (T �� becomes the half-value layer ficient from these tables, we can calculate
(HVL). The product of the linear attenu­ the percentage of transmitted photons for
ation coefficient and half-value layer IS a whole variety of photon energies and for
equal to 0.693: whatever thickness of tissue we may
fJ. x HVL = 0.693, or HVL = --

Mass Attenuation Coefficient
The half-value layer is the absorber thick­ Another useful attenuation coefficient
ness required to reduce the intensity of that is basic in the physics literature is the
the original beam by one half. It is a com­ mass attenuation coefficient used to quan­
mon method for expressing the quality of titate the attenuation of materials inde­
an x-ray beam. A beam with a high half­ pendent of their physical state. For ex­
value layer is a more penetrating beam ample, water, ice, and water vapor, the
than one with a low half-value layer. three physical states of H20, all have the
If the values from the attenuation of the same mass attenuation coefficient. It is ob­
x-ray beam in Figure 5-1 are used in the tained by dividing the linear attenuation
exponential attenuation equation, the lin­ coefficient by the density (p), and has the
ear attenuation coefficient can be calcu­ symbol fLip. The unit for the x-ray absorber
lated as is grams per square centimeter (g/cm2),
which contains a mass unit, and thus the
name mass attenuation coefficient. Figure
where N is 800 photons transmitted, N0 is 5-3 illustrates the meaning of a gram per
1000 initial photons, and x is 1 em of water. square centimeter by comparing aluminum
1000 and water. The density of water is 1 g/cmS,
800 1000 e-•1'1 = --

so a square centimeter of water must be 1



1000 em thick to weigh 1 g. The density of alu­

e• = -- = 1.25
800 minum is 2.7 g/cm3, so a square centimeter
fJ. 0.22/cm
of aluminum only has to be 0.37 em thick

The last step in the calculation requires a to weigh 1 g. The arithmetic is simple. The
table of exponential values, or natural log­ thickness is merely the reciprocal of the
arithms, or a calculator. After we have de­ density, or 1 divided by the density. A
termined the linear attenuation coefficient, square centimeter of water 1 em thick and
the half-value layer of the beam can be cal­ a square centimeter of aluminum 0.37 em
culated as thick both contain 1 g/cm2.
The unit of the mass attenuation coef­
0.693 0.693
3.15 em ficient is per g/cm2. It can be expressed in

several ways: per g/cm2 or 11g/cm2 or cm2/
You can see from Figure 5-1 that the beam g; usually it is written cm2/g.
is reduced to one half its original intensity A brief review should help to avoid con­
in the fourth 1-cm block of water, which fusion about units. When we talk about ei­
confirms our calculations. ther the linear or mass attenuation coeffi­
Calculating a linear attenuation coeffi­ cient, there are two separate units for each,
cient is not of much practical value to a one for the absorber and the other for the
diagnostic radiologist, but the equation can coefficient. The units for the absorber are
be used for other purposes. There are ta­ em for the linear attenuation coefficient
bles available that list the linear attenuation and g/cm2 for the mass attenuation coef­
coefficients for substances such as water, ficient. The units for the coefficients them­
bone, sodium iodide, and most of the ele- selves are cm-1 for the linear and cm2/g for

u -- -----
- --:-------!,

T --- ---
" -
-- 1
--- -�- :

I ---

....... :
'/ ------��;.J

I em


Density = l g/cm3 Density = 2. 7 g/cm3

Thickness of l g/cm2 Thickness of l g/cm2

l g/cm2 l g/cm2
= l em --- = 0.37cm
l g/cm3 2.7g/cm3

Figure 5-3 Thickness of 1 g/cm2 of water and of aluminum

the mass attenuation coefficient. The unit three has exactly the same amount of mass.
of the coefficient is the reciprocal of the The thickness of 1 g/cm2 of water is 1 em,
unit of the absorber. ice 1.09 em, and water vapor 1670 em.
A simple way to understand the rela­ These thicknesses will all attenuate the
tionship between the linear and mass at­ same amount of x ray (i.e., 20%) because
tenuation coefficients is to compare the co­ they all contain the same amount of mass.
efficients for water, ice, and water vapor As y ou can see in Figure 5-4, the mass
(Fig. 5-4). The values in the illustration are attenuation coefficient is independent of
for a 50-keV monochromatic beam. The the density of the absorber. The coeffi­
linear attenuation coefficient for water is cients for water, ice, and water vapor are

0.214 cm-1, and a 1-cm thickness of water all the same, but their densities vary con­
absorbs 20% of the incidence beam. The
To say that a gram of water and a gram
same thickness of ice absorbs 18.5%, be­
of water vapor both absorb the same
cause ice is a little less dense than water.
amount of radiation is true, but the state­
Water vapor has very little density , and a
ment has no meaning to a diagnostic ra­
1-cm thickness absorbs almost nothing.
diologist. In fact, it is misleading. We just
Density has a profound effect on x-ray at­
do not deal with g/cm2 of patient. We meas­
tenuation at all energy levels.
ure a patient's thickness in centimeters, and
The mass attenuation coefficient for wa­
1 em of water absorbs a great deal more x
ter in Figure 5-4 is 0.214 cm2/g, which is ray than 1 em of water vapor. When the
the same as the linear attenuation coeffi­ effect of density is removed from the value
cient. They must be the same, because the of an attenuation coefficient, it is mislead­
density of water is 1 g/cm3, and the mass ing in diagnostic radiology.
attenuation coefficient is obtained by divid­
ing the linear attenuation coefficient by the FACTORS AFFECTING
density (J.Lip). The mass attenuation coef­ ATTENUATION

ficient is the same for water, ice, and water Four factors determine the degree of at­
vapor, which is logical because 1 g of all tenuation of an x-ray beam as it passes

50 keV
Linear Mass
Attenuation Attenuation Thickness of I gm /cm2
Coefficient Coefficient Density
1 2
(cm- ) (cm /gm) (gm/cm3 )


0.214 0.214 I

I em

/ !/ ICE
0.196 0.214 0.917

v 1.09 em

/ # /

0.000128 0.214 0.000598 WATER 1670


Figure 5-4 Comparison of linear and mass attenuation coefficients for water, ice, and water

through matter. One involves the nature of we must digress to review the relationships
the radiation, and three involve the com­ between density (p), atomic number (Z),
position of the matter: and electrons per gram (e/g).
Density and Atomic Number. The re­
lationship between density and atomic
Energy Density
number is complex, and no simple rule cov­
Atomic number
Electrons per gram
ers all situations. In general, elements with
high atomic numbers are denser than el­
Increasing the radiation energy increases ements with low atomic numbers, but there
the number of transmitted photons (and are exceptions. Gold and lead provide a
decreases attenuation), while increasing good example:
the density, atomic number, or electrons
per gram of the absorber decreases the NUMBER (g/cm3)
number of transmitted photons (and in­ Gold 79 19.3
creases attenuation). Lead 82 11.0

Relationships Between Density, Atomic There is no relationship between atomic

Number, and Electrons per Gram number and density when different phys­
We will examine each of the four factors ical states of matter are considered . Water
that affect attenuation separately, but first has an effective atomic number of 7.4 re-

gardless of its state (ice, liquid, or vapor), Table 5-2. Percentage of Photoelectric
but its density is different in each of these
three forms.
Density and Electrons per Gram. When­ ENERGY (keV) (Z = 7.4) (Z = 13.8) (Z = 49.8)
ever a factor is expressed in the units per 20 65% 89% 94%
gram, the concept of volume is eliminated. 60 7% 31% 95%
Because density depends on volume 100 2% 9% 88%

(weight per unit volume), there is no re­

lationship between density and electrons
ble. Table 5-2 shows the percentage of
per gram. A gram of water has the same
photoelectric reactions for various radia­
number of electrons, regardless of whether
tion energies in water, bone, and sodium
they are compressed together in a 1-cm
iodide. You can easily calculate the per­
cube as a liquid, or spread out over 1670
centage of Compton reactions by subtract­
cm3 as a vapor.
ing the percentage of photoelectric reac­
Atomic Number and Electrons per
tions from 100 (ignoring the small
Gram. The number of electrons per gram
contribution from coherent scattering). As
is really a function of the number of neu­
the radiation energy increases, the per­
trons in the atom. If the elements did not
centage of photoelectric reactions de­
have neutrons, all materials would have 6.0
creases for water and bone; as the atomic
x 1023 e/g. Table 5-l shows the relation­
number increases, the percentage of pho­
ships between atomic number and elec­
toelectric reactions increases. With ex­
trons per gram for various substances im­
tremely low energy radiation (20 keV),
portant in diagnostic radiology. Hydrogen,
photoelectric attenuation predominates,
which has no neutrons, has 6.0 x 1023 e/g,
regardless of the atomic number of the ab­
which is twice as many as any other ele­
sorber. As the radiation energy is in­
ment. The elements found in soft tissue­
creased, Compton scattering becomes
carbon, nitrogen, and oxygen-all have 3.0
more important until eventually it replaces
x 1023 e/g, and the higher atomic elements
the photoelectric effect as the predominant
have even fewer e/g. In general, elements
interaction. With high atomic number ab­
with low atomic numbers have more elec­
sorbers, such as sodium iodide, the pho­
trons per gram than those with high atomic
toelectric effect is the predominant inter­
action throughout the diagnostic energy
Effects of Energy and Atomic Number range.
The linear attenuation coefficient is the
Energy and atomic number, in combi­
sum of the contributions from coherent
nation, determine the percentage of each
scattering, photoelectric reactions, and
type of basic interaction, so in this sense
Compton scattering:
their effects on attenuation are insepara-
fJ.. = f.J..coherent + f.J..PE + f.J..Compton
Table 5-1. Electrons Per Gram for Several
Elements Important in Diagnostic Radiology For example, at 40 keV, the linear atten­
ATOMIC NUMBER OF ELECTRONS uation coefficient for water is 0.24 cm-1. Of
this total, 0.018 is from coherent scattering,
Hydrogen 1 6.00 X 1023
0.18 is from Compton scattering and the
Oxygen 8 3.01 X 1023
remaining 0.042 is from photoelectric at­
Calcium 20 3.00 X 1023
Copper 29 2.75 X 1023 tenuation. For water, the mass and linear
Iodine 53 2.51 X 1023 attenuation coefficients are the same be­
Barium 56 2.45 X 1023 cause the density of water is 1 g/cm3. For
Lead 82 2.38 X 1023
other materials, however, the mass and lin-

ear attenuation coefficients are different, actions are photoelectric, and few photons
but the principle is exactly the same. are transmitted. As the energy of the ra­
If we know the relative percentage of diation increases, photoelectic attenuation
each type of interaction, we can predict the becomes less important, and completely
total amount of attenuation (or transmis­ ceases at 100 keV. Even when all attenua­
sion). Attenuation is always greater when tion is from Compton scattering, the per­
the photoelectric effect predominates. Per­ centage of transmitted photons continues
haps the difference between photoelectric to increase as the radiation energy in­
and Compton attenuation will be easier to creases. A larger percentage of photons is
understand with an analogy. Suppose we transmitted with a 150-keV than with a
take a flat stone and slide it along the 100-keV beam.
ground at a high speed. If we slide the As the energy of the radiation increases,
stone over a rough surface, such as a con­ photoelectric attenuation decreases
crete sidewalk, friction will stop it in a short sharply, while Compton attenuation de­
distance. If we slide the stone over a frozen creases only slightly. In the low energy
pond, friction is reduced, and the stone will range, a massive quantity of photoelectric
travel a much greater distance. In this anal­ attenuation is superimposed on a back­
ogy, the x-ray beam is the moving stone ground of Compton attenuation. As the
and attenuation is the friction that stops the energy of the radiation increases, photo­
stone. Photoelectric attenuation is like the electric attenuation diminishes until the
concrete sidewalk and Compton attenua­ background of Compton attenuation is all
tion is like the frozen pond. The photo­ that is left. Once this happens, increasing
electric effect attenuates a beam much the beam energy will cause only a slight
more rapidly than Compton scattering. decrease in attenuation (and slight increase
Energy. In addition to its influence on in transmission).
the type of basic interaction, energy has its As a general rule, the higher the energy
own effect on attenuation. Even when all of the radiation, the larger the percentage
the basic interactions are of one type, at­ of transmitted photons, regardless of the
tenuation decreases as the radiation energy type of basic interaction. The only excep­
increases. Table 5-3 shows the percentage tion to this rule occurs with high atomic
of photons transmitted through a 10-cm­ number absorbers, which we will discuss
thick water phantom with various radiation next.
energies. As you can see, the percentage of Atomic Number. Usually, as the radia­
transmitted photons increases as the en­ tion energy increases, x-ray transmission
ergy of the beam increases. With low en­ increases (and attenuation decreases) but,
ergy radiation (20 keV), most of the inter- with high atomic number absorbers, trans­
mission may actually decrease with increas­
ing beam energy. This apparent paradox
Table 5-3. Percent Transmission of
Monochromatic Radiation Through 1 0 occurs because there is an abrupt change
em of Water in the likelihood of a photoelectric reaction
ENERGY TRANSMISSION as the radiation energy reaches the binding
(keV) (%) energy of an inner shell electron. A photon
20 0.04 cannot eject an electron unless it has more
30 2.5
energy than the electron's binding energy.
40 7.0
Thus, a lower energy photon is more likely
50 10.0
60 13.0 to be transmitted than a higher energy
80 16.0 photon, provided one has slightly less and
100 18.0 the other slightly more energy than the
150 22.0
binding energy of the electron.

Table 5-4. Percent Transmission of 60

Monochromatic Radiation Through 1

mm of Lead Cl
(keV) (%) � /29 keV
1- I
50 0.016 I
z •
60 0.40 w I
40 I II
80 6.8 (.)
u.. I \1
88 12.0 u.. I II
-K edge for lead- w
0 I I I
88 0.026 (.) I I I I
30 I I
100 0.14 z I \
\ \
150 0.96 0 I I
I \
1- I \ TIN
< I
20 I
I '\/ \
Table 5-4 shows the percentage trans­ w I \
1- I \
\ \
mission of monochromatic radiation <
through 1 mm of lead. A sudden change 1/) 10 '
' 1ss keV
1/) ''

in transmission occurs at 88 keV, which is < LEAD
:::E ',
the binding energy of the K-shell electron.
This is called the K edge. With a radiation
energy just below the K edge, a fairly large
percentage of the photons is transmitted PHOTON ENERGY (keV)
(12%), while just above the K edge, trans­ Figure 5-5 Comparison of the mass attenua­
mission drops to nearly zero. Of course, the tion coefficients for tin and lead
same thing is true for the low atomic num­
ber elements at the K edge, but the ener­
gies involved are usually less than 1 keV, ium and iodine were originally selected be­
far below those in the useful diagnostic en­ cause of their low toxicities. A better choice,
ergy range. however, could not have been made if they
Figure 5-5 shows a plot of the mass at­ were selected for their physical properties.
tenuation coefficients of tin and lead over Both have ideal K-shell binding energies
the range of energies used in diagnostic (iodine, 32 keV; barium, 34 keV). These
radiology. Remember, the higher the at­ binding energies are approximately the
tenuation coefficient, the lower the num­ same as the mean energy of most diagnostic
ber of transmitted photons. Gram for x-ray beams, so many interactions occur at
gram, tin is a better absorber of x rays than the K-shell level. Attenuation is more in­
lead between 29 and 88 keV. This has a tense than it would be for a higher atomic
practical application. We are not usually number element, such as lead. Most pho­
concerned about attenuation per gram but, tons in a polychromatic beam are less en­
when we are carrying the gram on our back ergetic than the 88-keV K-shell binding
as a protective apron, our perspective energy of lead.
changes. Because tin attenuates more ra­ Table 5-5 shows the atomic numbers
diation per unit weight than lead, it has and K edges of most elements important
recently come into use for this purpose. A in diagnostic radiology. When maximum x­
lighter tin apron gives the same protection ray absorption is desired, the K edge of an
as a standard lead apron, or, if you have a absorber should be closely matched to the
strong back, you can use the same weight energy of the x-ray beam. For example,
of tin for the apron and have more pro­ xeroradiography employs a selenium plate
tection. with a K edge of 12.7 keV as the x-ray
The commonly used contrast agents bar- absorber. This low K edge makes selenium

Table 5-5. Atomic Numbers and K Edges of a difference in tissue densities is one of
Elements Important in Diagnostic Radiology
the primary reasons why we see an x-ray
image. Density determines the number of
ELEMENT SYMBOL (Z) (keV) electrons present in a given thickness, so it
Hydrogen H 1 .013 determines the tissue's stopping power.
Beryllium Be 4 .11 The relationship between density and at­
Carbon c 6 .28 tenuation is linear. If the density of a ma­
Nitrogen N 7 .40
terial is doubled, attenuation doubles. This
Oxygen 0 8 .53
is easiest to understand in terms of a gas
Sodium Na 11 1.1
Aluminum AI 13 1.6 whose density can be changed by com­
Silicon Si 14 1.8 pression. If a quantity of gas in a closed
Sulfur s 16 2.5 container is doubled, the number of elec­
Chlorine Cl 17 2.8
trons doubles. The same thickness of ma­
Potassium K 19 3.6
terial now has twice as much mass, and it
Calcium Ca 20 4.0
Iron Fe 26 7.1 will attenuate twice as many photons.
Copper Cu 29 9.0
Zinc Zn 30 9.7
Effect of Electrons per Gram
Germanium Ge 32 11.1
Selenium Se 34 12.7 The number of Compton reactions de­
Bromine Br 35 13.5 pends on the number of electrons in a
Yttrium y 39 17.0
given thickness. Absorbers with many elec­
Molybdenum Mo 42 20.0
Silver Ag 47 25.5 trons are more impervious to radiation
Cadmium Cd 48 26.7 than absorbers with few electrons. U nfor­
Tin Sn 50 29.1 tunately, the number of electrons is usually
Iodine I 53 33.2 expressed in the unit e/g (a mass unit)
Cesium Cs 55 36.0
rather than e/cm3 (a volume unit). In order
Barium Ba 56 37.4
Lanthanum La 57 39.0 for the number of electrons per gram to
Gadolinium Gd 64 50.2 have meaning, we must know the density
Tantalum Ta 73 67.5 of the material. Density determines the
Tungsten w 74 69.5 number of electrons that will be present in
Gold Au 79 80.7
a given thickness, and this is what deter­
Lead Pb 82 88.0
mines x-ray attenuation as we think of it in
clinical radiology. By multiplying electrons
an excellent absorber for the low energy per gram and density, we get electrons per
radiation (30 to 35 kVp) used for mam­ cubic centimeter:
mography. Selenium would be a poor
choice, however, as an absorber for a high­ e
- X-=-
9 e
energy beam, such as the 350 kVp used for 9 cm3 cm3

chest radiography with a field emission

Now we are talking about a unit we can
unit. For high-energy radiation, tungsten,
understand. A cubic centimeter represents
with a Kedge of 59.5 keV, is a much better
a thickness of 1 em, and this is the unit we
absorber. As we shall see later, the success
use to measure a patient's thickness.
of the rare earth intensifying screens and
Table 5-6 summarizes the relationships
of the cesium iodide image intensifier is in
between density, e/g, and e/cm3 for air, fat,
large part related to excellent matching of
water, and bone.
their K edges to beam energies.
When Compton reactions predominate,
Effect of Density on Attenuation the number of e/cm3 becomes the most im­
Tissue density is one of the most im­ portant factor in attenuation so, even
portant factors in x-ray attenuation, and though bone has fewer e/g than water,

Table 5-6. Comparison of Physical Characteristics of Air, Fat, Water, and Bone

Air 7.64 0.00129 3. 01 X 1023 0.0039 X 1023

Fat 5.92 0.91 3.48 X 1023 3.27 X 1023
Water 7.42 1.00 3.34 X 1023 3.34 X 1023
Bone 13.8 1.85 3. 0 X 1023 5.55 X 1023

bone still attenuates more radiation, be­ electrons per gram for various elements
cause it has more e/cm3. important in diagnostic radiology.
The number of electrons per gram can
be calculated by the equation: POLYCHROMATIC RADIATION

The attenuation of polychromatic radi­

ation is more complex than the attenuation
of monochromatic radiation. Polychro­
N0 = number of electrons per gram matic beams contain a whole spectrum of
N = Avogadro's number (6.02 x 1 023) photons of various energies (see Fig. 2-14),
Z = atomic number the most energetic of which are deter­
A = atomic weight. mined by the peak kilovoltage (kVp) used
to generate the beam. In general, the mean
When comparing one element with an­
energy of polychromatic radiation is be­
other, we can disregard Avogadro's num­
tween one third and one half of its peak
ber, because it is a constant and affects all
energy. A 100-kV p beam has a mean en­
elements equally. The relative number of
ergy of approximately 40 kV. This will vary
electrons per gram then becomes the sim­
somewhat depending on filtration.
ple ratio
As polychromatic radiation passes
Z number of electrons through an absorber, the transmitted pho­
- or
A weight of the atom tons undergo a change in both quantity and
quality. The number of photons decreases
The weight of the atom is the sum of the because some are deflected and absorbed
weights of the protons and neutrons. The out of the beam, just as they are with mon­
number of electrons and protons is equal, ochromatic radiation. In contrast to mon­
so the only factor that varies independently ochromatic radiation, however, the quality
is the number of neutrons, which add mass of the beam also changes because the lower
without affecting the number of electrons. energy photons are more readily attenu­
As the atomic number increases, the num­ ated than the higher energy photons. As
ber of neutrons increases faster than the the low-energy photons are removed from
number of electrons. Hydrogen has no the beam, the mean energy of the remain­
neutrons, and it has twice as many electrons ing photons increases.
per gram as any other element. Oxygen has The attenuation of polychromatic radi­
one neutron per electron, and half as many ation is shown in Figure 5-6. The x-ray
electrons per gram as hydrogen. Lead has beam begins with 1000 photons having a
more neutrons than electrons, so it has mean energy of 40 kV. The first centimeter
even fewer electrons per gram than oxy­ of absorber (water) reduces the number of
gen . In general, the high atomic number photons by 35% and increases their mean
elements have about 20% fewer electrons energy to 4 7 kV. The second centimeter of
per gram than the low atomic number el­ water reduces the number of photons only
ements. Table 5-l shows the number of 27%, because now the remaining photons

kVp= 100 energy of the polychromatic radiation ap­

proaches its peak energy.
1000 365 288
Photons Photons
40 kV 55 kV 57 kV
Mean Mean The photons in an x-ray beam enter a
patient with a uniform distribution and
-27% -21%
emerge in a specific pattern of distribution.
Figure 5-6 Attenuation of polychromatic ra­ The transmitted photons carry the x-ray
image, but their pattern also carries the
memory of the attenuated photons. The
are a little more energetic, and increases transmitted and attenuated photons are
their mean energy to 52 kV. If the process equally important. If all photons were
is continued long enough, eventually the transmitted, the film would be uniformly
mean energy of the beam will approach its black; if all photons were attenuated, the
peak energy. film would be uniformly white. In neither
When the percentage of transmission of case would there be an x-ray image. Image
polychromatic radiation is plotted on semi­ formation depends on a differential atten­
logarithmic graph paper, it results in a uation between tissues. Some tissues atten­
curved line (Fig. 5-7). The initial slope of uate more x rays than other tissues, and
the curve is steep, because many low-en­ the size of this differential determines the
ergy photons are attenuated by the first few amount of contrast in the x-ray image .
centimeters of water. Eventually, the slope To examine differential attenuation we
of the curve becomes similar to the slope will consider a radiograph of a hand so we
for a monochromatic beam as the mean only have to discuss two tissue types, bone
and soft tissue. Figure 5-8 shows the linear
attenuation coefficients of compact bone
and water. Water is used to represent the
soft tissues. Remember, the higher the lin­
(J) ear attenuation coefficient, the greater the
z 200 attenuation. As you can see, x-ray attenu­
� ation is greater in bone than in water, so
I the bones show up as holes in the trans­
a.. mitted photon pattern. We see these holes
0 as white areas on a film. With a 20-keV x­
w ray beam, water has a linear attenuation

� 20 coefficient of 0. 77 cm-1 and that of bone is

(J) 100 kVp
/ 4.8 cm-1• Bone has a greater linear atten­
z uation coefficient by a factor of 6. With
<l such a large differential, the x-ray image
0:: 5
� will display a great deal of contrast. At an
x-ray beam energy of 100 keV, differential
2 attenuation is not nearly as large. Bone still
(Semi- LoQ Graph)
attenuates more x rays than water, but now
4 8 12 16 20 the difference is only by a factor of 0 .6, so
image contrast is considerably reduced.
Figure 5-8 illustrates another point. The
Figure 5-7 Comparison of the attenuation of total linear attenuation coefficient is actu­
polychromatic and monochromatic radiation ally made up of three separate coefficients,

Water Compact Bone


<u 0.6
z 0.3

oL_�aJnnLB���L_L_ __ ��Uil������
20 40 60 80 100 20 40 60 80 100

keY keY

Coherent rzz:t, Compton CJ , Photoelectric CJ

Figure 5-8 Linear attenuation coefficients for water and compact bone

one for each of the basic interactions: co­ and sodium iodide. The number of Comp­
herent scattering, photoelectric effect, and ton reactions decreases only slightly with
Compton scattering. Each contributes its increasing x-ray energies. As the energy of
own component to the total linear attenu­ an x-ray beam is increased from 20 to 100
ation coefficient. At low photon energies keV, the Compton linear attenuation co­
(20 keV) most of the difference in x-ray efficient for bone decreases about 20%
attenuation between bone and soft tisue re­ (while the photoelectric linear attenuation
sults from a difference in the number of coefficient decreases more than 99%).
photoelectric reactions. More photoelectric The shaded portion of Figure 5-9 shows
reactions occur because bone contains cal­ the Compton mass attenuation coefficient.
cium, which has a relatively high atomic Remember, the mass attenuation coeffi­
number. At high photon energies (100 cient indicates the amount of attenuation
keV) the difference in x-ray attenuation be­ per gram, and it is not dependent on den­
tween bone and soft tissues is almost en­ sity. As you can see, a gram of water atten­
tirely the result of the difference in the uates more x rays than a gram of either
number of Compton reactions. Few pho­ bone or sodium iodide by Compton reac­
toelectric reactions occur in either bone or tions because water has more electrons per
soft tissue at 100 keV. When Compton re­ gram (it contains two atoms of hydrogen).
actions predominate; differential attenu­ Sodium iodide has even fewer electrons
ation entirely depends on differences in per gram than bone because of the high
density. The reason is that denser JTlaterials atomic number of iodine, so a gram of so­
have more electrons available to undergo dium iodide attenuates fewer photons by
reactions. Compton reactions than either a gram of
Figure 5-9 shows the Compton linear water or bone. Nevertheless, even when
attenuation coefficients for water, bone, Compton scattering is the only type of basic

Water Compact Bone Sodium Iodide

< u...
� � 0.4

� <


keY keY keY

Compton mass attenuation coefficient 0

Figure 5-9 Compton linear and mass attenuation coefficients for water, compact bone, and
sodium iodide

interaction, there is still a difference in x­ SCATTER RADIATION

ray attenuation between water, bone, and
sodium iodide in favor of the higher atomic Our discussion of attenuation has only
number absorber, but the difference is en­ dealt with primary radiation, which either
tirely dependent on density. passes through the patient unchanged or
Fat and water are always difficult to dif­ is completely removed from the useful
ferentiate radiographically. The effective beam. The primary radiation carries the x­
atomic number of water (7 .4) is slightly ray image. Now enters the villain, scatter
greater than that of fat (5.9), so they can radiation. It has nothing worthwhile to of­
be differentiated by photoelectric attenu­ fer. Scatter radiation detracts from film
ation using low energy techniques. With quality and contributes no useful infor­
higher energy radiation, when Compton mation.
attenuation predominates, differentiation How important is scatter radiation? How
between fat and water depends on density much of the density of an x-ray film is from
and on the number of electrons per gram. scatter radiation? Several factors deter­
Water has a greater density, but fewer elec­ mine its importance. In most of our routine
trons per gram. The net results, the num­ work it is important, frequently making up
ber of electrons per cubic centimeter, are from 50 to 90% of the total number of
nearly the same for both water and fat (Ta­ photons emerging from the patient. The
ble 5-6). Water has approximately 2% reason for the large number of scattered
more e/cmS, which is too small a difference photons will be more apparent if we con­
to demonstrate with only Compton reac­ sider the fate of a diagnostic x-ray beam.
tions. A significant difference in attenua­ With thick parts, such as the abdomen, only
tion between fat and water can only be 1% of the photons in the initial beam reach
demonstrated effectively using low energy the film. The rest are attenuated, the ma­
techniques. jority by Compton scattering. Fortunately,

most of them do not reach the film. Nev­

ertheless, those that reach the film make a
major, and undesirable, contribution to to­
tal film blackening.
It would be more accurate to refer to all
undesirable radiation as secondary radia­
tion, which includes photons and electrons
that might contribute to film fog. In actual
practice, however, none of the electrons
(photoelectric or recoil) have enough en­
ergy to reach the film, and the only char­
acteristic radiation with sufficient energy to
cause a problem occurs during contrast ex­
aminations with barium and iodine. The
only secondary radiation of any signifi­
cance comes from Compton scattering, and
it has become common practice to refer to
all undesirable radiation as scatter radia­
tion, which is fairly accurate.

Factors Affecting Scatter Radiation

Three factors determine the quantity of Figure 5-10 Angle of escap e for scattered
scatter radiation. These are: photons

1. kilovoltage (kVp)
2. part thickness
3. field size of scatter radiation from a narrow beam is
small to begin with, and most of it never
Scatter radiation 1s maxtmum with high­ reaches the plane of the film . For this rea­
kVp techniques, large fields, and thick son, narrow beam attenuation is the
parts-unfortunately, this is what we usu­ method used to measure the attenuation of
ally deal with in diagnostic radiology. We primary radiation, and the terms narrow
rarely have any control over part thickness beam and primary radiation are used syn­
and frequently must use large fields. The onymously.
only variable we can control is kVp, but As the x-ray field is enlarged, the quan­
even here we have less control than we tity of scatter radiation increases rapidly at
would like because patient doses increase first, and then gradually tapers off until
sharply with low-kVp techniques. finally it reaches a plateau, or saturation
Field size is the most important factor in point. A further increase in field size does
the production of scatter radiation. A small not change the quantity of scatter radiation
x-ray field (usually called a "narrow beam") that reaches the film. The total number of
irradiates only a small volume of tissue, so scattered photons in the field increases, but
it generates only a small number of scat­ the number that reaches any particular
tered photons. Most of them miss the film point on the film remains constant. For ex­
because they have a large angle of escape. ample, consider the central point of a cir­
Figure 5-10 shows the angle of escape for cular field 4 em in diameter. Scattered pho­
a scattered photon originating at point P. tons originating from the periphery of the
As you can see, the escape angle is much field can easily penetrate 2 em of tissue to
larger than the narrow angle encompassed reach the field's center. If the field is in­
by the primary beam. Thus, the quantity creased to a 30-cm circle, however, most

scatter photons do not have sufficient production of scatter radiation, we must

range to penetrate the 15 em of tissue be­ find ways of keeping it from x-ray film after
tween the field's margin and center. So, it has been generated. The most important
even though more scattered photons are of these is the x-ray grid, which we will
generated in large fields, they do not in­ discuss in Chapter 8.
crease the quantity reaching any particular
area. The saturation point for scatter ra­ SUMMARY

diation occurs with a field approximately Attenuation is the reduction in the in­
30 X 30 em, only a 12-in. square, which is tensity of an x-ray beam as it traverses mat­
not a large field for diagnostic radiology. ter either by the absorption or deflection
The quantity of scatter radiation reaches of photons from the beam. The attenua­
a saturation point with increasing part tion of monochromatic radiation is expo­
thickness, just as it does with increasing nential; that is, each layer of absorber
field size. The total number of scattered attenuates the same percentage of the pho­
photons keeps increasing as the part be­ tons remaining in the beam. The attenu­
comes thicker, but photons originating in ation of polychromatic radiation is not
the upper layers of the patient do not have exponential. A large percentage of the low­
sufficient energy to reach the film. Unfor­ energy photons are attenuated by the first
tunately we have little control over part few centimeters of absorber, so the quality
thickness and, except for the occasional use (mean energy) of the remaining photons
of a compressed band, we must accept the increases as the beam passes through an
patient as he is. absorber.
The effect of kilovoltage (kVp) on the The amount of attenuation depends on
production of scatter radiation is probably the energy of the radiation and three char­
not as important as part thickness, and cer­ acteristics of the tissue: atomic number,
tainly is not as important as field size. In density, and electrons per gram. Increas­
the low energy range (20 to 30 keV), in ing the radiation energy increases the num­
which the photoelectric effect predomi­ ber of transmitted photons, while increas­
nates, extremely little scatter radiation is ing the atomic number, density, or
produced. As the radiation energy in­ electrons per gram decreases transmission.
creases, the percentage of Compton reac­ Energy and atomic number together de­
tions increases, and so does the production termine the relative percentage of photo­
of scatter radiation. After Compton scat­ electric and Compton reactions. With low­
tering becomes the predominant interac­ energy radiation, and with high atomic
tion, scatter radiation production tends to number absorbers, a large amount of pho­
plateau, just as it does with increasing field toelectric attenuation is superimposed on
size and part thickness. The energy of the a small background of Compton attenua­
point at which the plateau occurs depends tion. As the energy of the radiation is in­
on the atomic number of the tissue, but the creased, photoelectric attenuation dimin­
plateau is not as well defined as it is with ishes, until the background of Compton
field size and part thickness. Even after attenuation is all that remains.
Compton reactions predominate, the Density is one of the most important fac­
quantity of scatter radiation continues to tors affecting attenuation, and radio­
increase with increasing beam energy be­ graphic image contrast is largely depend­
cause more photons scatter in the forward ent on differences in tissue density. The
direction, and their greater energy allows high contrast between air and soft tissues
them to penetrate greater thicknesses of occurs entirely because of density differ­
tissue to reach the film. ences. The number of electrons per gram
Because we have so little control over the plays a lesser role. Generally, as the atomic

number increases, the number of electrons ing an x-ray film increases with increasing
per gram decreases, but the decrease is field size, part thickness, and kilovoltage.
more than compensated by an even greater
increase in density. Thus, high atomic
l. Hubbell, J.H.: Photon Cross Sections, Attenua­
number elements attenuate more radia­ tion Coefficients, and Energy Absorption Coef­
tion, even though they have fewer elec­ ficients from 10 keY to 100 GeV. Washington,
DC, U.S. Government Printing Office, National
trons per gram. Bureau of Standards, Handbook 29, August
The amount of scatter radiation reach- 1969.

6 Filters

Filtration is the process of shaping the 2. Sheets of metal placed in the path of
x-ray beam to increase the ratio of photons the beam (added filtration)
useful for imaging to those photons that 3. The patient
increase patient dose or decrease image
contrast. Diagnostic x-ray beams are com­
posed of photons that have a whole spec­ INHERENT FILTRATION
trum of energies; this is, they are poly­
chromatic. Their mean energy is from one Filtration resulting from the absorption

third to one half of their peak energy, so of x rays as they pass through the x-ray

many photons fall in the lower energy tube and its housing is called inherent fil­

range. As polychromatic radiation passes tration. The materials responsible for in­

through a patient, most of the lower energy herent filtration are the glass envelope en­

photons are absorbed in the first few cen­ closing the anode and cathode, the

timeters of tissue, and only the higher en­ insulating oil surrounding the tube, and
the window in the tube housing. Inherent
ergy photons penetrate through the pa­
filtration is measured in aluminum equiv­
tient to form the radiographic image.
alents, which represent the thickness of
Because the patient's radiation dose de­
aluminum that would produce the same
pends on the number of absorbed photons,
degree of attenuation as the thickness of
the first few centimeters of tissue receive
the material in question. Inherent filtration
much more radiation than the rest of the
usually varies between 0.5 and 1.0 mm alu­
patient. This tissue can be protected by ab­
minum equivalent, and the glass envelope
sorbing the lower energy photons from the
is responsible for most of it.
beam before they reach the patient by in­
In a few special circumstances, unfiltered
terposing a filter material between the pa­
radiation is desirable. Because filtration in­
tient and the x-ray tube. Filters are usually
creases the mean energy of an x-ray beam,
sheets of metal, and their main function in
it decreases tissue contrast. The decrease is
diagnostic radiology is to reduce the pa­
insignificant in the higher energy range but
tient's radiation dose. In the last part of
with lower energy radiation, under 30 kVp,
this chapter we will discuss how heavy
this loss of contrast may be detrimental to
metal filters are also used to improve image image quality. Beryllium window tubes are
contrast. designed to produce an essentially unfil­
In a radiologic examination the x-ray tered beam. The exit portal of the glass
beam is filtered by absorbers at three dif­ envelope is replaced with beryllium (atomic
ferent levels. Beginning at the x-ray source, number 4), which is more transparent to
these are as follows: low energy radiation than glass. The ra­
1. The x-ray tube and its housing (in­ diation from these tubes has a minimum of
herent filtration) inherent filtration.


ADDED FILTRATION teristic radiation. Its own characteristic ra­

diation has so little energy ( 1.5 keV) that it
Added filtration results from absorbers is absorbed in the air gap between the pa­
placed in the path of the x-ray beam. Ide­ tient and filter.
ally, a filter material should � bsorb all low
energy photons and transmit all high en­ Filter Thickness
ergy photons. Unfortunately, no such ma­
After selecting a filter material, usually
terial exists. A material can be selected,
aluminum, the next step is to determine
however, to absorb principally low energy
the appropriate thickness for the filter. The
radiation by utilizing the proclivity of pho­
percentages of x-ray attenuation with 1, 2,
toelectric attenuation for low energy pho­
3, and 10 mm of aluminum for photons ofi
tons . Attenuation is most intense when the
various energies are shown in Table 6-1.
photoelectric effect is the predominant in­
Two millimeters of aluminum absorb
teraction and diminishes when Compton
nearly all photons with energies less than
reactions predominate. The energy of the
20 keV, so most of the advantages of fil­
radiation filtered from the beam can be
tration are achieved by this thickness. An
regulated by selecting a material wi�h an
aluminum filter more than 3 mm thick of­
appropriate atomic number. Alummum
fers no advantage; in fact, excess filtration
and copper are the materials usually se­
. has definite disadvantages. In Table 6-1,
lected for diagnostic radiology. Alummum,
the column on the right shows the per­
with an atomic number of 1 3, is an excel­
centage of photons attenuated by 10 mm
lent filter material for low energy radiation
of aluminum, clearly excessive filtration.
and a good general purpose filter. Copper,
The effect is an overall attenuation of the
with an atomic number of 29, is a better
beam, primarily by the absorption of high
filter for high energy radiation. It is inco�­
energy photons, because all the low energy
venient to change filters between exami­
photons have already been absorb�d by
nations, however, and there is a risk of for­
thinner aluminum filters. The quahty ofi
getting to make the change. For practical
the beam is not significantly altered, but its
reasons, most radiologists prefer to use a
intensity is greatly diminished. This length­
single filter material, usually aluminum.
ens the time required to make an exposure,

Copper is never used by itself as t e filter
. which increases the likelihood of the pa­
material. It is always used in combmauon
tient moving during the examination.
with aluminum as a compound filter. The
The National Council on Radiation Pro­
only reason for going to copper is to cut
tection and Measurements has recom­
down on the thickness of the filter. A com­
mended the following total filtration for
pound filter consists of two or more layers
of different metals. The layers are ar­
Table 6-1. Percent Attenuation of
ranged so that the higher atomic number
Monochromatic Radiation by Various
element, copper, faces the x-ray tube, an ? Thicknesses of Aluminum Filtration
the lower atomic number element, alumi­
num, faces the patient. Most filtration oc­ ENERGY
(keV) 1 mm 2mm 3 mm 10 mm
curs in the copper, and the purpose of the
aluminum is to absorb the characteristic ra­ 10 100 100 100 100
20 58 82 92 100
diation from copper. Photoelectric atten­
30 24 42 56 93
uation in copper produces characteristic
40 12 23 32 73
radiation with an energy of about 8 keV, 50 8 16 22 57
which is energetic enough to reach the pa­ 60 6 12 18 48
tient and significantly increase skin doses. 80 5 10 14 48
100 4 8 12 35
The aluminum layer absorbs this charac-

diagnostic radiology.2 The figu�es include Table 6-2. Exposure Dose to the Skin for
Comparable Density Radiographs of a Pelvic
both inherent and added filtratiOn:
Phantom (18 em thick) with Various
OPERATING kVp TOTAL FILTRATION Thicknesses of Aluminum Filtration
60-kVp BEAM
Below 50 kVp 0.5 mm aluminum
50 to 70 kVp 1.5 mm aluminum ALUMINUM EXPOSURE DOSE DECREASE IN
Above 70 kVp 2.5 mm aluminum
(mm), (mR ) (%)
None 2380
The effect of aluminum filtration on a
0.5 1850 22
90-kV p polychromatic x-ray beam is shown 1.0 1270 47
graphically in Figure 6-1. The unfiltered 3.0 465 80
beam is composed of a spectrum of pho­
tons, many of which have energies in the
radiograph was made without a filter and
10 to 20 keV range. The highest point in
was then repeated with increasing thick­
the curve occurs at 25 keV. Filtration re­
nesses of aluminum filtration. Exposure
duces the total number of photons in the
times were adjusted to produce films of
x-ray beam (area under the curve) but,
equal density, and the radiation dose to the
more importantly, it selectively removes a
skin over the pelvis was measured for each
large number of low energy photons. The
exposure. As you can see in Table 6-2, the
intensity on the low energy side of the
decrease in patient exposure was remark­
curve (left) is reduced considerably more
able, up to 80% with 3 mm of aluminum
than the intensity on the high energy side
filtration. These percentages are for a spe­
of the curve (right), and the highest point
cific examination, but they should give you
in the curve is shifted from 25 to 35 keV.
some insight into the degree of protection
The overall effect is an increase in the
you can expect from adequate filtration.
mean energy of the x-ray beam.

Effect on Exposure Factors

Effect of Filters on Patient Exposure
The major disadvantage of filtration is a
Trout and coworkers demonstrated the
reduction in the intensity of the x-ray
degree of patient protection afford�d y � beam. Filters absorb some photons at all
filters with a series of tests, one of whiCh IS
summarized in Table 6-2.3 In this partic­
energy levels, and we must compensate or �
the loss of higher energy photons by In­
ular experiment, they made multiple ra­
creasing exposure factors (mAs). Even
diographs of an 18-cm-thick pelvic � an­ � when it is necessary to increase exposure
tom using a 60-kV p x-ray beam. The Initial
because of filtration, the patient receives
less radiation than he would from an un­
>- filtered beam. The x-ray tube puts out
(f) more photons but the filter absorbs many
w of them, and the total number reaching the
t- Unfiltered
� patient actually decreases.
/\,.1.. ---.......... I
................ ...
t- /
Wedge Filters
<[ '
...... '
' Filtered ........
0 / ..... Wedge filters are occasionally used in di­
<[ ', ...
a:: agnostic radiology to obtain films of more
// _
uniform density when the part bemg ex­
0 25 50 75 100 .
amined diminishes greatly in thickness
from one side of the field to the other. The
Figure 6-1 Energy and intensity of unfiltered filter is shaped like a wedge. The relation­
and filtered polychromatic radiation ship between the filter and patient is shown

in Figure 6-2. When one side of the patient Table 6-3. Atomic Numbers and K Edges
Important to Heavy Metal Filters
is considerably thicker than the other, the
wedge compensates for the difference.
Less radiation is absorbed by the thinner
Aluminum AI 13 1.6
part of the filter, so more is available to
Molybdenum Mo 42 20
penetrate the thicker part of the patient. Gadolinium Gd 64 50.2
Wedge filters are often used in lower-limb Holmium Ho 67 55.6
angiography when one images from the Erbium Er 68 57.5
Ytterbium Yb 70 61.3
lower abdomen to the ankles with a single
Tungsten w 74 69.5
Iodine I 53 33.17
Barium Ba 56 37.45

(K-Edge Filters) characteristics of iodine and barium. The

The development of high speed inten­ reason we use iodine and barium is to pro­
sifying screens and high capacity x-ray vide contrast, and contrast is greatest when
tubes have made it reasonable to consider the contrast agent absorbs x rays most ef­
the use of heavy metal filters for general ficiently. This maximum contrast is ob­
radiography. These filters make use of the tained when the photon energy of the x-ray
K-absorption edge of elements with atomic beam is close to, but slightly above, the K­
numbers greater than 60, and may offer absorption edge of the absorber in ques­
advantages when imaging barium or io­ tion. For example, just below the K edge
dine. Table 6-3 lists some heavy elements of iodine the mass absorption coefficient of
that have been investigated as filters, and iodine is 6.6 cm2/g. Just above the K edge
compares them to aluminum, iodine, and the coefficient jumps to 36 cm2/g. For io­
barium. dine, the K-absorption edge is at 33.17 keV.
Let us begin by reviewing the absorption The K edge of barium is at 33.45 keV. Ta­
ble 6-4 lists the mass attenuation coeffi­
cient of iodine at several keV levels, illus­
trating that attenuation decreases below
and above the K edge and has a relative
maximum immediately above the K edge.
The purpose of heavy metal filters is to
produce an x-ray beam that has a high
number of photons in the specific energy
range that will be most useful in diagnostic
imaging. An immediate question might be:
what is wrong with our standard aluminum
filter? The energy spectrum transmitted by

Table 6-4. Approximate Mass Attenuation

Coefficient (J.Lp) of Iodine
keV APPROXIMATE ..P (cm2fg)

25 13.7
28 10.2
31 7.27
33.2 36.4
35 31.6
41 21.4
Figure 6-2 Wedge filter

aluminum filters is too wide. The K edge more photoelectric and less Compton at­
of aluminum is 1.6 keV, too low to be useful tenuation). Because of increased beam fil­
as an energy selective filter for diagnostic tration, increased x-ray tube loading (more
radiology. Figure 6-3 illustrates that a mAs) is required when heavy metal filters
2-mm aluminum filter transmits a broad are used. Patient dose reductions up to a
spectrum of bremsstrahlung. The same fig­ factor of two, and mAs increases by a factor
ure illustrates the effect a gadolinium filter of two or more, have been reported. Im­
has on the same x-ray beam. We illustrate proved contrast with a gadolinium filter
the general concept of heavy metal filters has been shown maximal for thin body
with a gadolinium filter 0.25-mm thick be­ parts , while the need for more mAs has
cause that is the most thoroughly studied been minimized by use of low (about 60)
of these filters. Notice that the gadolinium kVp techniques.1 This suggests that heavy
filter transmits increasing numbers of pho­ metal filters may prove to be more useful
tons from the 25 to 50.2 keV range (i.e., for pediatric applications.
the mass attenuation coefficient of Gd is The usefulness of a heavy metal filter in
decreasing from 25 to 50.2 keV). Imme­ imaging iodine is illustrated in Figure 6-4,
diately above 50.2 keV the mass attenua­ in which the mass attenuation coefficient
tion coefficient of gadolinium increases of iodine is compared to that of holmium.4
dramatica1ly and the number of transmit­ The attenuation coefficient of iodine in­
ted photons is correspondingly dimin­ creases dramatically at the 33.17-keV K
ished. The heavy metal filter transmits a edge of iodine, and then decreases with an
significantly narrower spectrum of ener­ increase in photon energy. Conversely , the
gies than aluminum, with decreased num­ attenuation coefficient of holmium de­
bers of both low and high energy photons. creases steadily from 33 to 55.6 keV (the
The reduction in low energy photons will K edge of holmium), at which point atten­
decrease the patient's absorbed dose. uation by holmium increases dramatically.
Image contrast will be improved by the re­
duction in higher energy photons (i.e.,
1.0 ...J
;::- 20
(/) z
z L.U
0 .8
0 10
u.. 8
0 .6
� 6
() \ I
a: \
w z 4 \ I
aJ 0 \I
:2' \

� A � �
w � 2 "Overlap
f= L.U Window"
:5 .2 �
w <l:
20 30 50 80 100
20 40 50.2 60 75 (Log Scale)
of Gd)

PHOTON ENERGY (keV) Figure 6-4 The attenuation coefficients of io­

dine and holmium, showing that the transmis­
Figure 6-3 Approximate bremsstrahlung sion of holmium and attenuation of iodine are
photon transmission of a 75-kVp x-ray beam simultaneously maximized in the region labeled
through an aluminum and a gadolinium filter "overlap window"

The important concept is the "window" in tion remarkably well, frequently reducing
the 33 to 55 keV range. In this window the skin exposures by as much as 80%. Alu­
high transmission by the holmium filter minum is usually selected as the filter ma­
overlaps the region of high attenuation by terial for diagnostic radiology. Most high
iodine (barium would be almost identical energy photons are transmitted through
except its K edge is at 37.45 keV). It is this aluminum, while low energy photons are
"overlap window" of about 20 keV that af­ absorbed by photoelectric interactions.
fords improved contrast when imaging io­ The photoelectric effect selectively absorbs
dine or barium when the x-ray beam fil­ low energy photons because of its depend­
tered by a heavy metal filter is compared ence on energy and atomic number. The
to that produced by an aluminum filter. National Council on Radiation Protection
and Measurement has recommended an
Molybdenum Filters equivalent of 2.5 mm of aluminum per­

A special application of K edge filters is manent filtration for diagnostic x-ray

use of molybdenum filters with molybde­ beams of energies greater than 70 kVp.

num target x-ray tubes used for mammog­ Some higher energy useful photons are ab­

raphy. Recall that a molybdenum target sorbed by filtration, and exposure factors
must be increased to compensate for this
x-ray tube is often used for film-screen
loss. Filters are simple and inexpensive.
mammography to take advantage of the
Nowhere else in all of radiology do we gain
17.5-keV K-alpha and the 19.6-keV K-beta
so much for so little money. Heavy metal,
characteristic radiation of molybdenum.
or K-edge, filters are used to remove
When operated at 30 to 40 kVp, a molyb­
higher energy photons from the x-ray
denum tube will also produce considerable
beam by taking advantage of the increase
bremsstrahlung with energies higher than
in mass attenuation coefficient at the K
20 keV. This higher energy radiation will
edge of certain elements. Except for mo­
reduce contrast in breast soft tissue struc­
lybdenum, these elements have atomic
tures. To reduce the amount of higher en­
numbers greater than 60. Compared to
ergy radiation in the molybdenum tube
aluminum, K-edge filters enhance contrast
spectrum, a molybdenum filter of 0.030
for iodine and barium, reduce patient
mm thickness is commonly used. This filter
dose, and increase tube loading.
will attenuate x rays just above the 20-keV
K-edge of Mo very strongly, but will trans­ REFERENCES
mit 57% of the 17.5 K-alpha and 67% of l. Burgess, A.E.: Contrast effects of a gadolinium
the 19.6 keV K-beta radiation of molyb­ filter. Med. Phys., 8:203, 1981.
2. National Council on Radiation Protection and
denum. Measurements: Medical X-Ray and Gamma-Ray
Protection for Energies up to 10 meV. W ashing­
ton, DC, U.S. Government Printing Office, Re­
port No. 33, 1968.
Filters are sheets of metal placed in the 3. Trout, E.D., Kelley, J.P., and Cathey, G .A.: The
use of filters to control radiation exposure to the
path of the x-ray beam near the x-ray tube
patient in diagnostic roentgenology. Am. ].
housing to absorb low energy radiation be­ Roentgen., 67:942, 1952.
fore it reaches the patient. Their main 4. Villagran, J.E., Hobbs, B.B., and Taylor, K.W.:
Reduction of patient exposure by use of heavy
function is to protect the patient from use­ elements as radiation filters in diagnostic radiol­
less radiation, and they perform their func- ogy. Rad., 127:249, 1978.

7 X-Ray Beam Restrictors

An x-ray beam restrictor is a device that tion for an x-ray beam restrictor but the
is attached to the opening in the x-ray tube flare of the cone is usually greater than the
housing to regulate the size and shape of flare of the x-ray beam, in which case
an x-ray beam. Beam restrictors can be clas­ the base plate that attaches the device to
sified into three categories: the tube housing is the only part that re­
stricts the x-ray beam (Fig. 7-lB). When
1. aperture diaphragms
used in this way, it is little more than an
2. cones and cyclinders
aperture diaphragm. Beam restriction with
3. collimators
a cylinder takes place at the far end of the
barrel, so there is less penumbra (Fig.
7 lC). Cylinders may be equipped with ex­

The simplest type of x-ray beam restric­ tensions to increase their length to give
tor is an aperture diaphragm. It consists of even better beam restriction. A major dis­
a sheet of lead with a hole in the center; advantage of aperture diaphragms, cones,
the size and shape of the hole determine and cylinders is the severe limitation they
the size and shape of the x-ray beam. Its place on the number of available field sizes.
principal advantage is its simplicity. Lead is Even a large assortment cannot approach
soft, so the aperture can be easily altered the infinite variety of field sizes needed in
to any desired size or shape. The principal diagnostic radiology, and changing them is
disadvantage of an aperture diaphragm is inconvenient.
that it produces a fairly large penumbra at
the periphery of the x-ray beam (Fig. COLLIMATORS
7- lA) The center of the x-ray field is ex­
The collimator is the best all around x­
posed by the entire focal spot, but the pe­ ray beam restrictor. It has two advantages
riphery of the field (P in the illustration) over the other types: (1) it provides an in­
"sees" only a portion of the focal spot; this finite variety of rectangular x-ray fields; (2)
partially exposed area is called the penum­ a light beam shows the center and exact
bra. The width of the penumbra can be configuration of the x-ray field. Two sets
reduced by positioning the aperture dia­ of shutters (S1 and S2) control the beam
phragm as far away from the x-ray target dimensions. They move together as a unit
as possible. This is usually accomplished by so that the second shutter aligns with the
attaching the diaphragm to the end of a first to "clean up" its penumbra (Fig. 7-2).
cone. The shutters function as two adjustable ap­
erture diaphragms. Each shutter consists
CONES AND CYLINDERS of four or more lead plates (Fig. 7-3).
The second type of x-ray beam restrictor These plates move in independent pairs.
comes in two basic shapes, conical and cyl­ One pair can be adjusted without moving
indric. The flared shape of a cone would the other, which permits an infinite variety
seem to be the ideal geometric configura- of square or rectangular fields. When the





A B c
Figure 7-1 Aperture diaphragm (A), cone (B), and cylinder (C)

shutters are closed, they meet at the center an angle of 45° (Fig . 7-4). The target of
of the x-ray field. the x-ray tube and light bulb should be ex­
The x-ray field is illuminated by a light actly the same distance from the center of
beam from a light bulb in the collimator. the mirror. As the light beam passes
The light beam is deflected by a mirror through the second shutter opening, it is
mounted in the path of the x-ray beam at collimated to coincide with the x-ray beam.
The distance from the mirror in the colli­
mator to the target of the x-ray tube is crit­
ical . The collimator must be mounted so
that the x-ray target and light bulb are ex­
actly the same distance from the center of
the mirror. If the collimator is mounted too


Figure 7-2 Alignment of collimator shutters Figure 7-3 Collimator shutters (top view)

tors, which position the shutters to exactly

match the size of the film being used. A
1 X-Ray Beam
perfectly aligned collimator will leave an

--+11-+--�---r--- Filter unexposed border on all sides of the de­

#�-----+--Mirror veloped film.
Positive beam limitation (PBL) devices
Light Bulb
(also called automatic light-localized vari­
�--+-- Light Beam
able-aperture collimators) must be accurate
--+-- -Ray 8
1----X to within 2% of the source-to-image dis­
Light Beams
tance (SID). It is required that total mis­
1--- -----t- Lower Shutter
alignment of the edges of the light field
with the respective edges of the x-ray field
along either the length or width shall not
be greater than 2% of the distance from
COLLIMATOR the focal spot of the x-ray tube (the source
of x-rays) to the center of the field. Figure
7-5 illustrates this alignment requirement .
With a standard 40-in. (or 100-cm) SID,
Figure 7-4 Alignment of light and x-ray
beams total misalignment along either the long or
short sides of the true edges of the x-ray
beam must not be greater than 0.8 in . (or
far from the x-ray target, the x-ray source 2 em).
will be further from the second shutter
than the light source, and the x-ray beam Testing X-Ray Beam and Light Beam
will be smaller than the light beam. Alignment
A collimator can also identify the center The alignment of the x-ray beam and
of the x-ray field. This is accomplished by light beam should be checked periodically,
painting a cross line on a thin sheet of Plex­ because the mirror tends to get out of ad­
iglas mounted on the end of the collimator. justment from daily use. The only equip-
The light beam that illuminates the radio­
graphic field also shows its center.
Collimators have a back-up system for X-Ray
identifying field size in case the light bulb Beam
should burn out. The x-ray field size for /
various target-film distances is indicated by
------- 1 ------- �
a calibrated scale on the front of the col­ I
Today, non-mobile radiographic equip­
ment in the United States, with a few spe­
cial exceptions, must be equipped with au­
tomatic collimator s . These are called
positive beam limiting (PBL) devices. l ________ y_______ J

Automatic collimators are the same as
other collimators except that their shutters "'Light
are motor-driven. When a cassette is Beam
loaded into the film holder (Buckey tray),
Figure 7-5 Alignment of the light beam and
sensors in the tray identify the size and x-ray beam for a PBL device must be accurate
alignment of the cassette. These sensors re­ to 2% of the SID. At a 40-in. SID, A + B s0.8
lay this information to the collimator mo- in. and x + y :S 0.8 in.

ment needed to test alignment includes the film exposure, and adjust the mirror in
four L-shaped wires (pieces of paper clips the collimator until the light beam coin­
work nicely), a 14- to 17-in. x-ray film in a cides exactly with the x-ray field (dark area
cassette, and a small lead letter "R." Place on the film). Then repeat the test to make
the film on the top of the x-ray table, open sure that the mirror adjustment has cor­
the collimator shutter to a convenient size rectly aligned the light and x-ray beams.
(10 X 10 in.), carefully position the L­
shaped wires at the corner of the light field, FUNCTIONS OF RESTRICTORS

and place the "R" in the lower right corner. Collimators and other x-ray beam re­
Then make an x-ray exposure (40 in., 3.3 strictors have two basic functions: to pro­
mAs, 40 kVp) to mark the position of the tect the patient and to decrease scatter
x-ray field on the film. Without touching radiation.
the film or wires, enlarge the field size to Although patient protection is the prin­
12 X 12 in. and expose the film for the cipal reason for using collimators today, it
second time (same exposure factors). is interesting to note that the old-time ra­
Figure 7-6 shows a test film taken with diologists restricted x-ray beams before
a collimator whose mirror is out of adjust­ they knew anything about the harmful ef­
ment. The dark center shows the position fects of low doses of radiation. They used
of the x-ray beam, and the wires indicate small fields because they obtained better
the position of the light beam. The second films with small fields. Because they had no
exposure (with an enlarged field) is made grids, their only means of controlling scat­
to ensure visibility of all the wires. The wire ter radiation was to limit field size, so they
in the left lower corner of the illustration collimated their x-ray beams and obtained
would not have been visible on the film excellent films.
without the second exposure.
To adjust a misaligned mirror, return Patient Protection

the processed x-ray film to its original po­ The mechanism by which collimators
sition on the x-ray table. The "R" in the protect the patient is obvious; that is, the
lower right corner assists in orienting the smaller the x-ray field, the smaller the vol­
film properly. Position the light beam to the ume of the patient that is irradiated. If a
images of the wires, as you did earlier for 20- x 20-cm field is collimated to 10 x 10
em, the area of the patient that is irradiated
decreases from 400 to 100 cm2, a fourfold
decrease in area and a corresponding de­
crease in volume. A decrease in field size
is especially significant, because area is a
square function. Trimming only a few cen­
timeters off the edge of the field signifi­
cantly decreases the exposed volume of the
The ideal shape for a radiographic field
for maximal patient protection has never
been established. Initially, all x-ray beam
restrictors were round (cones) and now
they are usually square or rectangular (col­
limators). It was argued that x-ray films are
rectangular. Round fields expose portions
Figure 7-6 Test film of light and x-ray beam of the patient that are not even included
alignment on the film, as shown in Figure 7-7, so x-

.,..,. --�
""' ......
·' I
I .. ,
I \
I \
I \
\ I
',, ..,
...... /
...... __ _

Figure 7-7 Overlap of a round field on a rec­

tangular film

ray beam restrictors were made square. But

the ideal field shapes should be dictated by
the shape of the part being examined, and
not by the shape of either the film or col­
limator. Some parts are better examined
with round fields, such as gallbladders and
paranasal sinuses. Certainly a circular field
that is completely encompassed by the film
is the most suitable shape when a round
part is being examined. a =
It might be interesting to look at a prob­ b B
lem regarding beam restrictor sizes. These
Figure 7-8 The sizes of the aperture and x­
problems can be solved by setting up pro­
ray field are proportional to the target-aperture
portions between the comparable sides of
and target-film distances
similar triangles (Fig. 7-8). For example,
if an aperture diaphragm (A) is 25 em from
is, the larger the field, the more scatter ra­
the x-ray target, and the target film dis­
diation. Figure 7-9 shows the general
tance (B) is 100 em, what size aperture (a)
shape of the curves obtained when the
should be used to produce a 15-cm x-ray
number of transmitted primary and scat­
field (b) for a gallbladder examination? Us­
tered photons are plotted against field size.
ing proportional triangles, the aperture
The exact contribution from scatter de­
hole should be
pends on the thickness of the part being
a 25 25 examined and on the energy (kVp) of the
-=-, or a = 15 x- = 3.75cm
15 100 100 x-ray beam. The amount of primary ra­
diation at the plane of the film is inde­
Decreased Scatter Radiation pendent of field size. The number of trans­
with Collimators mitted primary photons, per unit area, is
The quantity of scatter radiation reach­ the same for a 1- X 1-cm field and a 30-
ing an x-ray film depepds on field size; that x 30-cm field, so the number of the pri-

tors is that they affect exposure factors.

Small fields produce little scatter radiation,
c� so the total quantity of blackening of the
0 rJ>
£. c
:!: x-ray film decreases as the field size de­
c._::> creases. To keep film density constant, as
you decrease the size of the x-ray field you
must increase the exposure factors.

There are three types of x-ray beam re­
strictors: aperture diaphragms, cones (cyl­
inders), and collimators. Their basic func­
0 100 200 300 400 tion is to regulate the size and shape of the
Field Size (cm2) x-ray beam. Closely collimated beams have
two advantages over larger beams. First, a
Figure 7-9 Transmitted radiation for various smaller area of the patient is exposed and,
sizes of x-ray fields
because area is a square function, a de­
crease of one half in x-ray beam diameter
mary photons in Figure 7-9 plots a straight effects a fourfold decrease in patient ex­
line. posure. Second, well-collimated beams
The number of scattered photons reach­ generate less scatter radiation and thus
ing the x-ray film depends on field size. improve film quality. By decreasing the
Small x-ray fields generate little scatter ra­ amount of scatter radiation, collimators
diation. As the field size is enlarged, the also affect exposure factors. As the x-ray
amount of scatter radiation increases rap­ field size is decreased, the exposure factors
idly at first, and then tapers off with larger must be increased to maintain a constant
fields. After the x-ray field reaches a size film density.
of about 30 X 30 em, the total quantity of Collimators are the best general purpose
scatter radiation is near its maximum. Be­ beam restrictors. They offer two advan­
cause collimators are only successful in de­ tages over the other types: the x-ray field
creasing scatter radiation with small fields, is illuminated, which permits accurate lo­
we should restrict the size of our x-ray calization on the patient; and the x-ray field
beams as much as possible. can be adjusted to an infinite variety of rec­
A final fact to remember about collima- tangular shapes and sizes.

8 Grids

The radiographic grid consists of a series pound. The main purpose of the inter­
of lead foil strips separated by x-ray-trans­ space material is to support the thin lead
parent spacers. It was invented by Dr. Gus­ foil strips. Aluminum interspace grids can
tave Bucky in 1 91 3, and it is still the most probably be manufactured more precisely,
effective way of removing scatter radiation and they are structurally stronger than
from large radiographic fields. Figure 8-1 grids with organic interspacers. Patient ex­
shows how a grid functions. Primary ra­ posures are higher with aluminum because
diation is oriented in the same axis as the it absorbs more primary radiation. It also
lead strips and passes between them to absorbs more secondary radiation, how­
reach the film unaffected by the grid. Scat­ ever, so contrast improvement is probably
ter radiation arises from many points better. At the present time both materials
within the patient and is multidirectional, are used, and there is no clear-cut differ­
so most of it is absorbed by the lead strips entiation as to which is better.
and only a small amount passes between

The interspaces of grids are filled either Grid ratio is defined as the ratio between
with aluminum or some organic com- the height of the lead strips and the dis­
tance between them. Figure 8-2 is a cross­

----...J! ---- Tube Anode sectional scale drawing of a grid with a ratio
of 8: l. The lead strips are approximately
0.05 mm thick, so that they may appro­
priately be considered lead foil. The inter­
Primary spaces are much thicker than the lead
Radiation strips. Grid ratios are usually expressed as
two numbers, such as 10:1, with the first
number the actual ratio and the second
number always l. The grid ratio is a pa­
rameter widely used to express a grid's abil­
ity to remove scatter radiation. Ratios usu­
ally range from 4:1 to 16: l. Generally, the
higher the ratio, the better the grid func­
�-<-7---- Scatter tions. Grid ratios are indicated on the top
Radiation of grids by manufacturers.
Grid pattern refers to the orientation of
the lead strips in their longitudinal axis. It
is the pattern of the grid that we see from
a top view. The two basic grid patterns are
linear and crossed.
Figure 8-1 Grid function In a linear grid the lead strips are par-


( "

) ��
( �:

:� 2.0mm

Figure 8-4 Crossed grid

l-.o-1 perimposed linear grids that have the same

focusing distance (Fig. 8-4). T-he grid ratio
h 2.0 of crossed grids is equal to the sum of the
r = = =
D 0.25 ratios of the two linear grids. A crossed grid

made up of two 5:1 linear grids has a ratio
r GRID RATIO of 10:1, and functions about the same as a
linear grid with a 10:1 ratio. Crossed grids
cannot be used with oblique techniques re­
D - DISTANCE BETWEEN quiring angulation of the x-ray tube, and
LEAD STRIPS this is their biggest disadvantage.
A focused grid is a grid made up of lead
Figure 8-2 Cross section of a grid strips that are angled slightly so that they
focus in space (Fig. 8-5). A focused grid
allel to each other in their longitudinal axis may be either linear or crossed, because the
(Fig. 8-3). Most x-ray tables are equipped focusing refers to the cross-sectional plane
with linear grids. Their major advantage is of the lead strips. Most grids are focused.
that they allow us to angle the x-ray tube Linear focused grids converge at a line in
along the length of the grid without loss of space called the convergent line. Crossed
primary radiation from grid "cutoff." grids converge at a point in space called
A crossed grid is made up of two su- the convergent point. The focal distance
is the perpendicular distance between the
grid and the convergent line or point. In
practice, grids have a focusing range that

indicates the distance within which the grid
X- can be used without a significant loss of
Table Top
primary radiation. The focusing range is
fairly wide for a low-ratio grid and narrow

Lin�or Grid for a high-ratio grid. A 5:1 grid focused at

40 in. has a focusing range of approxi­
Figure 8-3 Linear grid mately 28 to 72 in., whereas a 16:1 grid

cused and have a fairly low grid ratio (4: l

to 8: l).


Grids are used to improve contrast by

absorbing secondary radiation before it
reaches the film. The "ideal grid" would
absorb all secondary radiation and no pri­
mary radiation. It would give maximum
film contrast without an unnecessary in­
crease in patient exposure. The ideal grid
does not exist, however, and the design of
all grids is a compromise. The price of bet­
ter film contrast is increased patient ex­
posure. We must always compromise one
Figure 8-5 Focused linear grid
for the other. Grids with high ratios give
maximum contrast, but the improved con­
focused at 40 in. has a range of only 38 to
trast is not always worth the increased pa­
42 in. Focal ranges are indicated on the top
tient exposure. In every clinical situation
of grids by manufacturers.
we must- weigh these two factors. To help
A parallel grid is one in which the lead
in grid selection and design, several tests
strips are parallel when viewed in cross sec­
have been devised to evaluate grid per­
tion. They are focused at infinity, so they
formance. These tests also help us under­
do not have a convergent line. These grids
stand the way a grid functions. We will dis­
can only be used effectively with either very
cuss three methods of evaluating
small x-ray fields or long target-grid dis­
tances. They are frequently used in fluor­
oscopic spot film devices, but otherwise l. primary transmission (Tp)
have little use in modern radiology. 2. Bucky factor (B)
Lines per inch is the number of lead 3. contrast improvement factor (K)
strips per inch of grid. It can be calculated
by adding the thickness of the lead strips Primary Transmission
and interspaces and dividing this sum into Primary transmission is a measurement
l. Because the thickness of the lead strips of the percentage of primary radiation
and interspaces is usually expressed in mil­ transmitted through a grid. Ideally, a grid
limeters, the answer must be multiplied by should transmit l 00% of the primary ra­
25.4, which is the number of mm/in. The diation, because it carries the radiographic
final equation is image. The equipment for measuring pri­

25.4 mary transmission is shown in Figure 8-6.

Lines/in. = --
The x-ray beam is collimated to a narrow
D+ d
pencil of radiation, and the phantom is
D = thickness of interspaces placed a great distance from the grid. With
d thickness of lead strips
this arrangement, no scatter radiation

(both in millimeters)
reaches the grid. A small amount is pro­
A grid (grid-front) cassette is a special x­ duced in the phantom, but it diffuses out
ray cassette, usually used for portable ra­ of the beam in the large air-gap between
diography, with a grid built into the front the phantom and the grid.
of the cassette. Most grid cassettes are fo- Two measurements must be made to de-

Table 8-1 shows the primary transmission

for eight different grids. There is a signif­
icant loss of primary radiation with grids.
For the eight grids tested, the primary
transmission varied between 57 and 72.5%,
and was generally lower for cross grids.
The primary transmission as measured
experimentally is less than would be antic­
ipated. If the geometric relationship be­
Water tween the lead strips and the x-ray target
is accurate, so that there is no grid cutoff,
Phantom and if no primary radiation is absorbed in
the interspaces, then the percentage of the
grid's surface area that is made up of in­
terspaces will be the percentage of antici­
pated primary transmission. This merely
determines the percentage of the grid sur­
Lead face area that is not lead, and is as close as
�--- -
Diaphragm we will ever come to an ideal grid. The
equation for calculating the anticipated
primary transmission is

Calc. Tp = 0 + x 100

Tp = anticipated primary transmission (%)

d = thickness of lead strips
D = thickness of interspaces

The measured primary transmission is al­

Grid ways less than the calculated primary trans­
• J --- Detector The example in Figure 8-7 will help to
Figure 8-6 Apparatus for measuring primary demonstrate the magnitude of this differ­
transmission (Tp) ence. The dimensions are for grid number
8, shown in boldface in Table 8-1, which
termine the percentage of primary trans­
is a 15:1 grid with an experimentally de­
mission. The first measurement is made
termined primary transmission at 64%. Its
with the grid in place to determine the in­
lead strips are 60 1-1 thick and its interspaces
tensity of the radiation transmitted
are 390 1-1 wide. The calculated primary
through the grid, and the second meas­
transmission is 87%, which is 23% more
urement is made after removal of the grid
than the measured primary transmission.
to determine the intensity of the radiation
The difference is largely the result of ab­
directed at the grid. A simple ratio of the
sorption by the interspace material. Also,
intensity with the grid to the intensity with­
there is probably some loss in primary ra­
out the grid gives the fractional transmis­
sion, which is multiplied by 100 to give the diation because of manufacturing imper­

percentage of transmission: fections in the focusing of the lead strips.

TP = _jl_ X 100 Bucky Factor
I' p
The Bucky factor is the ratio of the in­
TP = primary transmission (%)
IP = intensity with grid cident radiation falling on the grid to the
I'P = intensity without grid transmitted radiation passing through the

Table 8-1. Primary Transmission (Tp)

NO. RATIO (IJ.) (IJ.) (IJ.)
1 3.4 50 320 67.5
2 2 X 3.1* 50 350 57
3 11 30 220 69
4 7 60 290 72.5
5 6 80 370 72.5
6 9 50 380 67
7 2 X 7* 50 370 72.5
8 15 60 390 64
(Modified from Hondius Boldingh.')
*Crossed grids

Calc Tp = 100
D + d

II 390
= ___,;.___ X 100
390 + 60

llfr -----­ II J
= 87% 6o .u _.I 1 .....
39o .u _..I I
Figure 8-7 Anti cipated primary transmission

grid. It is a practical determination, be­

cause it indicates how much we must in­
crease exposure factors when we change
from a nongrid to a grid technique. It also
tells us how much the patient's exposure
dose is increased by the use of a grid. The
Bucky factor is similar to primary trans­
mission except for one difference. Primary
transmission indicates only the amount of
primary radiation absorbed by a grid,
whereas the Bucky factor indicates the ab­
sorption of both primary and secondary
radiation. It is determined with a large x­
ray field and a thick phantom (Fig. 8-8).
The transmitted radiation is measured
with the grid in place, and the incident ra­
diation is measured after the grid has been
removed. The equation for calculating the
Bucky factor (B) is

8 = incident radiation
transmitted radiation

The Bucky factor is a measure of the

total quantity of radiation absorbed from Figure 8-8 Apparatus for measuring the
an x-ray beam by a grid so, in part, it is a Bucky factor

measure of the grid's ability to absorb scat­ Table 8-2. Bucky Factor (B)
ter radiation. Table 8-2 shows the Bucky GRID
RATIO 70 kVp 120 kVp
factors for several different grids with two
No grid 1 1
different energies of radiation. Generally,
5:1 3 3
8:1 3.5 4
high-ratio grids absorb more scatter radi­
ation and have larger Bucky factors than 12:1 4 5
low-ratio grids. The size of the Bucky fac­ 16:1 4.5 6
tor also depends on the energy of the x­
ray beam. High-energy beams generate the quantity of scatter radiation, the poorer
more scatter radiation and place a greater the contrast, and the lower the contrast im­
demand on a grid's performance than low­ provement factor. To permit comparison
energy radiation. The Bucky factor for the of different grids, the contrast improve­
16:1 grid in Table 8-2 increases from 4.5 ment factor is usually determined at 100
to 6 as the radiation energy increases from kVp with a large field and a phantom 20
70 to 120 kVp. More scatter radiation is em thick, as recommended by the Inter­
generated by the 120-kVp beam, and the national Commission on Radiologic Units
16:1 grid successfully absorbs it. The 5: 1 and Measurements.3 Table 8-3 shows the
grid in Table 8-2, however, is not as effi­ contrast improvement factor for eight dif­
cient. Its Bucky factor is the same for both ferent grids. It is more closely related to
70- and 120-kVp radiation. Low-ratio grids the lead content of the grid than any other
do not absorb scatter radiation from high­ factor. Generally, the higher the grid ratio,
energy beams as well as high-ratio grids. the higher the contrast improvement fac­
Although a high Bucky factor is desira­ tor, but grid ratio is not as reliable an in­
ble from the point of view of film quality, dicator as lead content.
it is undesirable in two other respects. Although the contrast improvement fac­
higher the Bucky factor, the greater the tor of a particular grid depends on kVp,
exposure factors and radiation dosage to field size, and part thickness, the relative
the patient. If the Bucky factor for a par­ quality of two grids appears to be inde­
ticular grid-energy combination is 5, then pendent of these factors. A grid that has a
exposure factors and patient exposure high contrast improvement factor at 100
both increase 5 times over what they would kVp will have a high factor at 50 or 130
be for the same examination without a grid. kVp. In comparing two grids, the one that
performs better with low energy radiation
Contrast Improvement Factor will also perform better with high energy
The contrast improvement factor (K) is radiation.
the ratio of the contrast with a grid to the
contrast without a grid:
Table 8-3. Contrast Improvement Factor (K)
contrast with a grid LEAD CONTRAST
K contrast without a grid GRID GRID CONTENT IMPROVEMENT
(mg/cm2) (K)


1 3.4 170 1.95

This is the ultimate test of grid perform­
ance because it is a measure of a grid's abil­ 2 2 X 3.1* 310 1.95
ity to improve contrast, which is its primary 3 11 340 2.1
function. 4 7 390 2.1
5 9 460 2.35
Unfortunately, the contrast improve­
6 15 460 2.6
7 2 X 7* 680 2.95
ment factor depends on kVp, field size, and
phantom thickness. This is understanda­ 8 15 900 2.95
ble, because these three factors determine (Modified from Hondius Boldingh.')
the amount of scatter radiation. The larger *Crossed grids.

A B c

h h
3.0mm 3.0mm

o.o3. -o 3__...
mm mm
� 1
Decrease d Decrease D

Grid Ratio 10 : 10 10 :

Lines I inch 71 77 98

Lead Content
600 326 552

Figure 8-9 Relationship between grid ratio, lines per inch, and lead content of a grid

LEAD CONTENT thinner, as in Figure 8-9B, compared to

Figure 8-9A, the number of lines per inch
The lead content of a grid is expressed increases without affecting grid ratio, be­
in g/cm2• An easy way to understand what cause the thickness of lead strips is not con­
this means is to imagine cutting a grid up sidered in determining grid ratio. When
into 1-cm squares and then weighing one the lead strips are made thinner, however,
square. Its weight in grams is the lead con­ there is only a small increase in the number
tent of the grid (ignoring the interspace of lines per inch at a cost of a large decrease
material). The amount of lead in a grid is in lead content. If the number of lines per
a good indicator of its ability to improve inch is increased by decreasing the width
contrast, provided the grid is well de­ of the interspaces, as in Figure 8-9C, then
signed. Poor design would be a solid sheet the lead strips must be made shorter to
of lead. keep the grid ratio constant and, again, the
There is a definite relationship between lead content of the grid decreases. This
the grid ratio, lead content, and number of puts a limitation on the number of lines
lines per inch (Fig. 8-9). If the grid ratio per inch a grid may have and still be ef­
remains constant and the number of lines fective. If a 10:1 grid could be constructed
per inch is increased, the lead content must with 1000 lines per inch, it would be only
decrease. The only ways to increase the 0.2 mm thick, and it would have too little
number of lines per inch is by decreasing lead to be of any real value.
the thickness of either the lead strips or In practice, when grids are constructed
interspaces. If the lead strips are made with many lines per inch, both the thickness

and height of the lead strips are decreased. film, or both edges of the film, depending
These grids are thinner, and improve con­ on how the cutoff is produced. The
trast less than grids of comparable ratios amount of cutoff is alway s greatest with
with fewer lines per inch. A 133-line 10: 1 high-ratio grids and short grid-focus dis­
grid improves contrast about the same as tances. There are four situations that pro­
an 80-line 8: 1 grid. duce grid cutoff:

1. focused grids used upside down

2. lateral decentering (grid angulation)
The primary disadvantage of grids is 3. focus-grid distance decentering
that they increase the amount of radiation 4. combined lateral and focus-grid dis­
that the patient receives. Another disad­ tance decentering
vantage is that they require careful center­
ing of the x-ray tube because of the danger Upside Down Focused Grid
of grid cutoff. Grid cutoff is the loss of All focused grids have a tube side, which
primary radiation that occurs when the is the side of focus of the lead strips. When
images of the lead strips are projected a focused grid is used upside down, there
wider than they would be with ordinary is severe peripheral cutoff with a dark band
magnification (Fig. 8-10). It is the result of exposure in the center of the film and
of a poor geometric relationship between no exposure at the film's periphery. Figure
the primary beam and the lead foil strips 8-11A illustrates how cutoff occurs, and
of the grid. Cutoff is complete and no pri­ Figure 8-11B shows the resultant radio­
mary radiation reaches the film when the graph. The higher the grid ratio, the nar­
projected images of the lead strips are rower the exposed area. When a crossed
thicker than the width of the interspaces. grid is used upside down, only a small
The resultant radiograph will be light in square in the center of the film is exposed.
the area in which the cutoff occurs. With
Lateral Decentering
linear grids there may be uniform light­
ening of the whole film, one edge of the Lateral decentering results from the x­
ray tube being positioned lateral to the con­

---.T..J•i- ARGET
vergent line but at the correct focal distance
(Fig. 8-12A). All the lead strips cut off the
same amount of primary radiation, so
,, there is a uniform loss of radiation over
,, the entire surface of the grid, producing
,, a uniformly light radiograph. This is prob­
,, ably the most common kind of grid cutoff,
,, but it cannot be recognized by inspection
,, of the film. All we see is a light film that is
usually attributed to incorrect exposure
\ \
\ \ factors. Figure 8-12B shows a series of film
\ \ strips that were all taken with the same
\ \

>.------r-\0-GRIO exposure factors, but with increasing

amounts of lateral decentering. The x-ray
tube was centered at the convergent line
\ for the film strip on the left, and then lat­
' \
FILM erally decentered 1, 2, and 3 in. for the
next three strips. The films become pro­
�� L-eur-OFF gressively lighter as the amount of lateral
Figure 8-10 Grid cutoff decentering increases, but the exposure is

Figure 8-11 Cutoff from an upside down fo­

cused grid (A) and radiograph resultmg from Figure 8-12 Cutoff from lateral decentering,
an upside down focused grid (B) (A) and series of radiographs resultmg from m­
creasing amounts of lateral decentering (B)
still uniform. The center and both edges
of the film are equally exposed, and it's
grid ratio and decentering distance in­
impossible to recognize the cutoff from in­
crease, and cutoff decreases as the focal
spection of the film.
distance increases (Tables 8-4 and 8-5).
Three factors affect the magnitude of
The loss of primary radiation for any given
cutoff from lateral decentering: grid ratio,
amount of lateral decentering can be min­
focal distance, and the amount of decen­
imized with low-ratio grids and a long focal
tering. The equation for calculating the
distance. With a 5:1 grid focused at 72 in.,
loss of primary radiation with lateral de-
there is a 14% loss of primary radiation
centering IS
with 2 in. of lateral decentering. If the grid
rb ratio is increased to 16: 1, the loss of pri­
L =- X 100
fo mary radiation increases to 45%, and if the

L = loss of primary radiation (%) focal distance is then decreased to 40 in.,

r = grid ratio the loss of primary radiation goes up to
b = lateral decentering distance (inches) 80%. When exact centering is not possible,
f0 = focal distance of grid (inches)
as in portable radiography, low-ratio grids
The amount of cutoff increases as the and long focal distances should be used

Table 8-4. Loss of Primary Radiation from Target

Lateral Decentering for Grids Focused at 40

- X-Ray
RATIO 2 4 5 6

5:1 13 25 38 50 63 75
6:1 15 30 45 60 75 90
8:1 20 40 60 80 100
10:1 25 50 75 100
12:1 30 60 90 100
16:1 40 80 100

whenever possible. It takes 8 in. of lateral

decentering for 100% cutoff of primary ra­
Figure 8-13 Cutoff from an off-level grid
diation with a 5:1 grid focused at 72 in.,
whereas it takes only 2.5 inches of lateral
decentering for complete cutoff with a nitude. The cutoff is greater with near than
16:1 grid focused at 40 in. with far focus-grid distance decentering.
Off-Level Grids. When a linear grid is Figures 8-14 and 8-15 show the cutoff be­
tilted, as it frequently is in portable radi­ coming progressively greater with increas­
ography, there is a uniform loss of primary ing distance from the film center. The cen­
radiation across the entire surface of the tral portion of the film is not affected, but
grid (Fig. 8-13). The effect on the film is the periphery is light. Because the frac­
the same as the effect of lateral decenter­ tional loss of radiation is not uniform, it
mg. must be calculated for a point, which ac­
tually represents two parallel lines running
Focus-Grid Distance Decentering the length of the grid on either side of the
In focus-grid distance decentering, the center line. The loss of primary radiation
target of the x-ray tube is correctly cen­ is directly proportional to the grid ratio and
tered to the grid, but it is positioned above distance from the center line.

or below the convergent line. If the target

is above the convergent line, it is called far
focus-grid distance decentering; if the tar­
get is below the convergent line, it is called
near focus-grid distance decentering. The
results are the same, but they differ in mag-

Table 8-5. Loss of Primary Radiation from

Lateral Decentering for Grids Focused at 72


RATIO 2 3 4 5 6

5:1 7 14 21 28 35 42
6:1 8 17 25 33 42 50 -""'-'"'-'r..L.l..LJ....l.J-U...U'--L..ll...l..i..!...lo.o�..U.*-;:-_-_:Primary
8:1 11 22 33 45 56 67
10:1 14 28 42 56 70 83
12:1 17 33 50 67 83 100
Figure 8-14 Cutoff from near focus-grid dis­
16:1 22 45 67 90 100
tance decentering

with far focus-grid distance decentering.

At the margin of a 14-in. wide film (7 in.
from grid center), the loss of primary ra­
diation is 35% for a 6: l grid and 94% for
a 16: l grid with near focus-grid distance
decentering. With far focus-grid distance
decentering, the loss at the film margin de­
creases to 21% for a 6: l grid and 64% for
a 16: l grid.
Parallel grids are focused at infinity so,
of course, they are always used with near
focus-grid distance decentering (Fig.
8-16). They usually have a low grid ratio
to minimize cutoff. The only time there is
no significant cutoff is with long target-grid
distances or small fields. A film taken with
a parallel grid has a dark center and light
Transmitted edges because of near focus-grid distance
""""'"""""'-'""'""-� decentering.

Figure 8-15 Cutoff from far focus-grid dis­ Combined Lateral and Focus-Grid
tance decentering Distance Decentering
The most commonly recognized kind of
The equations for calculating the loss of
grid cutoff is from combined lateral and
primary radiation for near and far focus­
focus-grid distance decentering. It is prob­
grid distance decentering are as follows:
ably not as common as lateral decentering
Near Focus-Grid Far Focus-Grid alone, but lateral decentering cannot be
Distance Decentering Distance Decentering recognized as such on the resultant radio­

L = rc (.! - .!)
f, f0
x 1 00 L = rc (.! - .!,)
f0 f
x 100
graph. Combined decentering is easy to
recognize. It causes an uneven exposure,
resulting in a film that is light on one side
L = loss of primary radiation
at point c (%) and dark on the other side. There are two
r = grid ratio kinds of combined decentering, depending
f0 = grid focusing distance on whether the tube target is above or be­
f, = target-grid distance (below low the convergent line. The amount of
convergent line; inches), cutoff is directly proportional to the grid
f2 the target-grid distance (above
ratio and decentering distance, and in­

convergent line; inches)

versely proportional to the focal distance
c = distance from center of grid (inches)
of the grid. With high-ratio grids and large
Table 8-6 shows the extent of cutoff at decentering errors, there is a large loss of
various distances from the center of grids primary radiation. With long focus-grid
focusing at 40 in. with a 10-in. focus-grid distances, there is less loss of primary ra­
distance decentering error. The left half of diation.
the table indicates near (30 in.) and the Combined lateral and focus-grid dis­
right half far (50 in.) focus-grid distance tance decentering above the convergent
decentering. The loss of primary radiation line is illustrated in Figure 8-17. The pro­
increases as the grid ratio increases and as jected images of the lead strips directly be­
the distance from the center of the grid low the tube target are broader than those
increases; the loss is greater with near than on the opposite side, and the film is light

Table 8-6. Loss of Primary Radiation from Near and Far Focus-Grid Distance Decentering for a
Grid Focused at 40 Inches

TARGET AT 30 in. TARGET AT 50 in.

RATIO 3 5 7 3 5 7

6:1 5 15 25 35 3 9 15 21
8:1 7 20 33 47 4 12 20 28
10:1 8 25 42 58 5 15 25 35
12:1 10 30 50 70 6 18 30 42
16:1 13 40 66 94 8 32 48 64

on the near side. Cutoff is greatest on the MOVING GRIDS

side directly under the x-ray tube. Com­
The moving grid was invented by Dr.
bined lateral and focus-grid distance de­
Hollis E. Potter in 1920 and, for many
centering below the convergent line is il­
years, a moving grid was called a Potter­
lustrated in Figure 8-18. The projected
Bucky grid. In recent years the name has
images of the lead strips are broader on
been shortened to Bucky grid, which is un­
the side opposite the tube target than on
fortunate, because the name of the inven­
the same side, and the film is light on the
tor is omitted. Grids are moved to blur out
far side. Cutoff is least on the side under
the shadows cast by the lead strips. Most
the x-ray tube. With equal decentering
moving grids are reciprocating, which
errors the amount of cutoff is greater with
means they continuously move 1 to 3 em
combined decentering below the conver­
back and forth throughout the exposure.
gent line than with combined decentering
They start moving when the x-ray tube an-
above the convergent line.

X- Ray


Figure 8-17 Cutoff from combined lateral

Figure 8-16 Cutoff from a parallel grid and far focus-grid distance decentenng

strip on the film. Between pulses, the grid

moves one interspace distance so that the
next lead strip is over the position of the
preceding strip and, with a new pulse, the
\ two images are superimposed. Thus, even
\ though the grid moves, the grid lines will
\ be distinct.
\ Moving grids have several disadvan­
tages. They are costly, subject to failure,
\ may vibrate the x-ray table, and put a limit
\ on the minimum exposure time because
\ they move slowly. An even more serious
disadvantage is that they increase the pa­
tient's radiation dose. There are two rea­
sons for this, the first of which is lateral
decentering. Because the grid moves 1 to
3 em during the exposure, the tube is not
Figure 8-18 Cutoff from combined lateral centered directly over the center of the grid
and near focus-grid distance decentering
during most of the exposure. This lateral
decentering may result in a loss of as much
ode begins to rotate. Older grids move only as 20% of the primary radiation with a
in one direction and must be cocked for high-ratio grid and a short focusing dis­
each exposure. They also have timers that tance.
are set for a little longer than the exposure The second cause of increased patient
time to avoid grid lines. These single-stroke exposure is more difficult to understand.
grids are inconvenient, and are seldom The photons are spread out uniformly on
used in modern equipment. the film by a moving grid. With the iden­
Moving grids are advantageous because tical number of photons per unit area, a
they eliminate grid lines from the film. film is 15% darker with a stationary grid
Many years ago this was important, because because the photons are concentrated be­
the lead strips were thick and unevenly tween the lead strips. Figure 8-19 attempts
spaced. With improved manufacturing to illustrate how this happens. The total
methods, however, grids have improved so quantity of density (blackness) on the film
that grid lines are not nearly as distracting, is represented by the water level in Figure
and many radiologists now prefer station­ 8-19A. This is the amount of density that
ary grids. is spread out uniformly over the film by a
If you are using a moving grid and want moving grid. In Figure 8-l9B a grid is im­
to avoid grid lines, you must take two pre­ mersed in the water to represent the grid
cautions. First, the grid must move fast lines on a film. Now all the water is con-
enough to blur its lead strips. If it moves
too slowly, you will see either the grid lines
themselves or random density variations in
the film that are just as distracting as the

lines. Second, the transverse motion of the
grid should be synchronous with the pulses
of the x-ray generator. When this happens,
the shadow of each lead strip is superim­
posed on the shadow of its neighbor. Each Figure 8-19 Analogy between water level and
pulse produces a faint image of the lead film density for moving and stationary grids

centrated between the grid lines, and the there is a great deal of scatter radiation,
water level rises just as the amount of black­ such as in biplane cerebral angiography.
ness on the film rises. The total quantity of The efficiency of scatter radiation ab­
density on the film is the same with both sorption by various grids is shown in Figure
moving and stationary grids, but with sta­ 8-20, which plots the fractional transmis­
tionary grids the film is made up of many sion of scatter radiation against grid ratio.
narrow voids (grid lines) containing no There is little decrease in transmitted scat­
density, interspersed with black lines con­ ter beyond an 8:1 ratio grid, and almost no
taining all the density. change between 12:1 and 16:1. For this
reason 12:1 grids are preferable to 16:1
GRID SELECTION grids for routine radiography. The im­
There is no simple rule to guide the cli­ provement in film quality in going from a
nician in choosing a grid for any particular 12:1 to a 16:1 grid is not worth the greater
situation. A compromise is always involved. patient exposure.
The price of increased "cleanup" with
high-ratio grids is that patient exposure is
considerably increased and that x-ray tube The x-ray grid is the most important
centering becomes more critical. You must means of reducing scatter radiation with
weigh the value of a better-quality radio­ large radiographic fields but, under certain
graph against your moral obligation to circumstances, an alternate method, an air
keep the patient's exposure at a minimum. gap, produces comparable results with less
There is little to be gained using high ratio patient exposure. The basic principle of
grids in the low kVp range. Usually, 8:1 the air gap technique is quite simple. Scat­
grids will give adequate results below 90 ter radiation arising in the patient from

kVp. Above 90 kVp, 12:1 grids are pre­ Compton reactions disperses in all direc­
ferred. Crossed grids are only used when tions, so the patient acts like a large light
bulb (Fig. 8-21A). In the illustration, scat­
tered photons are shown radiating out
from a point source, but the point actually
� 1.0 represents a tiny block of tissue. Each ray
Q represents a separate scattering event, and
� numerous scatterings within the small
Q block produce the array shown. The closer
Gl: the patient is to the film, the greater the
..... concentration of scatter per unit area. With

u an air gap, the concentration decreases be­

0.5 cause more scattered photons miss the film.
0 The name "air filtration" has been applied
z to air gap techniques, but this is a misno­
II) mer. Scatter radiation decreases not from
filtration but from scattered photons miss­
ing the film. Negligible quantities of radi­
� 0.1
ation are absorbed in the gap, and the
beam is not appreciably hardened, so the
0 2 4 6 8 10 12 14 16 18
name "air filtration" should be discarded.
GRID RATIO There is no strong bias for forward scat­
tering in the diagnostic energy range, so
Figure 8-20 Fractional transmission of scatter
radiation through grids (Courtesy of Sven the distribution of Compton scattering is
Ledin and the Elema Shoenander Company) nearly random. A photon is almost as likely

Figure 8-21 Air gap techniques

to scatter at 90° as at 45°, or 180°. With cause a natural "gap" separates the film
increasing energy, more photons scatter in from the scattering site. The difference in
the forward direction, but the increase is the contribution from the two sides is even
negligible in the diagnostic range. When greater than the illustration implies. Many
the energy of the incident photon is in­ scattered photons from the input surface
creased from 10 to 140 keV, a much greater (Fig. 8-21C) are absorbed during their
change than is likely in clinical radiology, long journey through the patient, whereas

the most probable angle of scatter only those originating near the exit surface (Fig.

changes from 53o to 45°.4 With a strong

8-2lB) have only a short escape distance.
As a result of these two factors, a greater
forward bias, as in megavoltage therapy,
angle of capture and less tissue attenuation,
air gaps are ineffective as a means of con­
most of the scattered photons reaching the
trolling scatter radiation.
film arise in the immediately adjacent su­
Figure 8-21B and C compares the dis­
perficial tissue layers, an ideal arrangement
tribution of scattered radiation arising
for an air gap technique. You can see from
from superficial tissue blocks on opposite
the illustration that the air gap is most ef­
sides of the patient. More radiation reaches fective in removing scatter radiation when
the film from scattering near the exit sur­ the scatter originates close to the film (Fig.
face (Fig. 8-21B), because the film inter­ 8-2lB).
cepts a larger angle of the scattered
"beam." The angle of interception is Optimum Gap Width
smaller for scattering occurring on the in­ Air gap techniques are used in two clin­
put side of the patient (Fig. 8-21C), be- ical situations, magnification radiography

and chest radiography. With magnification change is gradual without an abrupt tran­
techniques the object-film distance is opti­ sition point. The optimum gap width is a
mized for the screen-focal spot combina­ matter of judgment. The following four
tion and the air gap reduces the scatter guidelines should be used to select a gap
radiation to acceptable levels as an inciden­ width:
tal bonus. In chest radiography the air gap 1. The thicker the part, the more ad­
is used instead of a grid and techniques are vantageous a larger gap. Figure 8-22
designed around the air gap. For example, shows that a change from a 1- to a 15-in.
image sharpness deteriorates as the gap width produces a more marked decrease
widens, so the focal-film distance is usually in scatter radiation for a 22-cm phantom
lengthened from 6 to 10 ft to restore sharp­ than for a 10-cm phantom.
ness. Air gaps and grids have the same 2. The first inch of air gap improves
function, to remove as much scatter radi­ contrast more than any subsequent inch.
ation and as little primary radiation as pos­ The increase from a 1- to a 2-in. gap im­
sible. As a measure of performance, gap proves contrast more than an increase
width is analogous to grid ratio. A large air from a 14- to a 15-in. gap.
gap, like a higher ratio grid, removes a 3. Image sharpness deteriorates with in­
larger percentage of scatter radiation. creasing gap width unless the focal-film dis­
Figure 8-22 shows the ratio of scatter to tance is increased to compensate for the
primary radiation with air gap widths from greater magnification. An increase in focal­
% to 15 in. for three different absorber film distance from 6 to 10 ft is customary
thicknesses. A ratio of 5 means that 5 scat­ in chest radiography to compensate for this
tered photons reach the film for each pri­ loss of sharpness. A further increase in fo­
mary photon. At the ideal ratio of zero, all cal-film distance is usually not phy sically
the transmitted photons are primary. As possible because of space limitations, or
the gap increases, the ratio of secondary to technically feasible at an acceptable expo­
primary radiation decreases, but the sure time because of the heat tolerance of
the x-ray tube. Remember, a 50% increase
z in distance doubles the exposure factors.
0 8
;:::: 4. If the gap is widened by moving the
0 ' patient away from the film with a fixed fo­
<( '
a: ' cal-film distance, the patient is closer to the
>- 6 ' x-ray tube and his exposure increases. The
a: '
<( ' increase can be corrected by a comparable
::::!; ,
a: ' increase in the focal-film distance, but this
0.. ...,
, , may not be a possible alternative for the
0 " , ...,
I- 4 ...,
... , Absorber reasons stated above.
a: ...,
w ..., ... , Thickness Table 8-7 shows that a 5-in. air gap ap­
...... ', , (em)
...... ... proaches the performance of a light (7: 1)
u ..... ...... ''·22
(f) 2 ...... ......
......... ......
l4. ...... ......
0 ... ...... ... Table 8-7. Comparative Contrast
..----- ..... 15
0 Improvement Factors for Grid and Air Gap
....... _·10
I- Techniques (100 kVp, 30- x 30-cm Field)
0 5 10 15 THICKNESS
(em OF
AIR GAP (inches) WATER) 5 in. 10 in. 7:1 15:1

Ratio of scatter to primary ra- 10 1.39 1 . 81 1.50 1.85

Figure 8-22
20 1.37 2.21 2.10 2.85
diation for various air gaps and absorber thick-
nesses (Data from Gould and Hale.•)

grid for thin patients. A 10-in. gap is the must be increased for the air gap technique
same as a heavy (15: 1) grid for thin pa­ because of the larger focal-film distance.
tients, but not as good for thicker patients. Apply ing the inverse square law (constant
The final decision on air gap width is usu­ KVp), exposures would have to be in­
ally made empirically. Changing widths creased 2.8 times (1202 ..;- 722 = 2.8) but
from patient to patient is inconvenient. Ex­ the actual increase is somewhat less, be­
perience has shown that a 10-in. gap and cause only 65% of the primary photons are
a 10-ft focal-film distance with an energy transmitted through the grid. The increase
from 110 to 150 kVp produces satisfactory is 1.8 times (2.8 x 0.65 = 1.8), a substantial
results. difference that taxes the heat capacity of
the x-ray tube.
Exposure Factors with Air Gaps Patient exposures are usually less with air
Figure 8-23 compares a 6-ft focal-film gaps than grids, because grids absorb pri­

distance ( FFD) with a grid to a 10-ft focal mary photons. With the grid technique

film distance with a 10-in. air gap. The ex­ shown in Figure 8-23, 1.54 primary pho­

posure factor, patient exposure, and trans­ tons must pass through the patient for each
mitted primary radiation are all assigned a one reaching the film (1.0 ..;- 0.65 = 1.54);

value of one for the grid technique, which the difference (35%) is absorbed in the

is used as the standard for comparison. In grid. Only 1.19 primary photons need pass
this example, the primary transmission of through the patient with the air gap tech­

the grid is 65%, and the experimental de­ nique to produce a comparable concentra­
sign was chosen to produce films of equal tion per unit area of film; the difference

density (same total number of photons) (1.19 - 1.0 = 0.19 or 19%) is lost as a

and contrast (same ratio of primary to sec­ consequence of the inverse square law. The

ondary photons). The x-ray tube exposure air gap loses less primary radiation, so the
patient's exposure is less. If the patient ex­
posure is unity with the grid, it will be 0.77
with the air gap (1.19 + 1.54 = 0.77), a
significant but not dramatic savings.
1- -EXPO SURE-� Magnification with Air Gaps

Two factors determine the amount of

magnification, the object-film distance and
the focal-film distance. Magnification is
greatest with a short focal-film distance and
721n--FFD - • 120 in
a long object-film distance. As a general

I rule, image sharpness deteriorates with

magnification. The objective with an air
PATIENT gap technique is to preserve image sharp­
1•- �EXPO SURE-·· .77
ness by lengthening the focal-film distance

I until magnification returns to pre-air-gap
levels. Some deterioration is inevitable,
1 19- l- however, for parts close to the film. Figure
8-24 shows that the objective is not accom­
.1 plished with a 120-in. focal-film distance,
at least not for a 10-in. air gap and a rea­
sonably sized (10-in. thick) patient. Mag­
Figure 8-23 Comparison of grid and air gap nification is less with the 72-in. focal-film
techniques for chest radiography distance for all object-film distances up to

plest way of characterizing a grid. Grids

come in two patterns, linear and crossed.
AP CHEST Crossed grids consist of two superimposed
linear grids; their great disadvantage is that
they cannot be used with oblique tech­
1.30 niques. Most grids are focused to a line in
space called a convergent line, but they can
z be used over a variable range of distances
Q 1.25
,__ called the "focal range."
0! Three tests have been devised to evaluate
z 1.20 grid performance. ( 1) Primary transmis­
::::. sion is a measurement of the percentage of

1.15 primary radiation transmitted through a

grid (usually between 60 and 70%). (2) The
Bucky factor is a measurement of the total
radiation (primary and scatter) absorbed
by a grid. It indicates how much exposure
factors must be increased because of grids,
and also reveals how much more radiation
the patient receives. Bucky factors range
0 3 6 9 12 15 18
from 3 to 7, depending on the grid ratio.
(3) The contrast improvement factor is a
(Plus 10 '" for the A1r Gop Techn1quel measurement of a grid's ability to improve
contrast, and it is the best test of grid per­
Figure 8-24 Magnification with 72-inch (no formance. Generally, high-ratio grids with
air gap) and 120-inch (air gap) focal-film dis­
a high lead content have the highest con­
tances (FFD)
trast improvement factors.
The grid ratio, lead content, and number
15 in. The longer focal-film distance pro­ of lines per inch are all interrelated. A grid
duces a more uniform magnification from with many lines per inch (over 100) is gen­
the front to the back of the patient, a de­ erally thinner and has a lower lead content
sirable characteristic. This uniformity is than a grid of comparable ratio with fewer
preserved with lesser air gaps. With a 5-in. lines per inch.
gap, the curve would maintain its slope but Grid cutoff is the loss of primary radia­
would slide down the magnification scale tion that occurs when the images of the
to cross the 72-in. curve at a shorter object­ lead strips are produced wider than they
film distance. Experienced chest radiog­ would be with ordinary magnification. The
raphers do not feel that image sharpness amount of cutoff with a particular decen­
deteriorates noticeably with air gap tech­ tering error is greatest with high-ratio grids
mques. that have a short focusing distance. Lateral
decentering is the most common error, and
SUMMARY it causes a uniform loss of primary radia­
Radiographic grids consist of lead foil tion, so it is impossible to recognize on the
strips separated by x-ray transparent spac­ film. With focus-grid distance decentering,
ers. They are used to absorb scatter radi­ the film is light on both sides, whereas with
ation and to improve radiographic image combined lateral and focus-grid distance
contrast. The grid ratio is defined as the decentering, the film is light only on one
ratio of the height of the lead strips to the side.
distance between: them, and it is the sim- Moving grids eliminate the image of the

lead strips from the film, and this is their face and diminishes rapidly at increasing
chief advantage. Generally, exposure fac­ distance from the surface. If the film is
tors are a little greater with moving grids placed at a distance, the scatter simply
than with stationary grids because of a lat­ misses the film. Focal-film distances are in­
eral decentering error during a portion of creased with air gap techniques in an at­
the exposure, and because the exposure is tempt to maintain image sharpness and, as
spread out over the entire surface of the a consequence, x-ray exposure factors are
film. With properly positioned stationary greater with the air gaps than grids. Patient
grids there is no decentering error, and the exposures are generally less with air gaps,
exposure is concentrated between the lead however, because the grid absorbs some
strips. primary radiation.
Grid selection involves a compromise be­ REFERENCES
tween film quality and patient exposure. l. Hondius Boldingh, W.: Grids to reduce scattered
High-ratio grids produce films with better x-ray in medical radiography. Eindhoven, Neth­
erlands, Philips Research Laboratories, 1964, No.
contrast at a cost of increased patient ex­ 1, Philips Research Reports Supplement.
posure. Generally, low-ratio grids are ad­ 2. Characteristics and Applications of X-Ray Grids,
rev. ed. Liebel Flarsheim Company , 1968.
equate for low energy radiation; 8:1 grids
3. International Commission on Radiologic Units
should be used with energies less than 90 and Measurements: Methods of Evaluating
kVp, and 12:1 grids for higher energy ra­ Radiologic Equipment and Materials. W ashing­
ton, DC, U.S. Government Printing Office, Na­
tional Bureau of Standards, Handbook 89, Au­
Air gaps are an alternative method of gust, 1963.
eliminating scatter radiation with large ra­ 4. Gould, R.G., and Hale, J.: Control of scattered
radiation by air gap techniques: Applications to
diographic fields. The intensity of scatter
chest radiography. Am. J. Roentgenol., 122:109,
radiation is maximum at the patient's sur- 1974.

9 Luminescent Screens

The x-ray photons making up the radi­ cent and photofluorographic screens). Re­
ographic image cannot be seen by the hu­ cent advances in technology have resulted
man eye. This information is converted in the introduction of a number of new
into a visual image by one of two methods. phosphors, such as cesium iodide in image
A photographic emulsion can be exposed intensifier tubes and several new intensi­
to the x rays directly. More commonly, the fying screen phosphors, including barium
energy of the x rays is converted into ra­ strontium sulfate, yttrium, and the rare
diation in the visible light spectrum, and earths gadolinium and lanthanum, and the
this light may be used to expose x-ray film rare earth tantalates.
(radiography or photofluorography), or
the light may be viewed directly (fluoros­
copy). The sensitivity of film to direct x-ray
exposure is low. Direct exposure of film Intensifying screens are used because
would require prohibitively large patient they decrease the x-ray dose to the patient,
x-ray doses for most examinations. There­ yet still afford a properly exposed x-ray
fore, almost all radiographic film exami­ film. Also, the reduction in exposure allows
nations require that the radiographic use of short exposure times, which be­
image be converted into light at some stage comes important when it is necessary to
by a luminescent screen. minimize patient motion.
The x-ray film used with intensifying
FLUORESCENCE screens has photosensitive emulsion on
Luminescence refers to the emission of both sides. The film is sandwiched between
light by a substance. It can be caused by two intensifying screens in a cassette, so
varying kinds of stimuli (e.g., light, chem­ that the emulsion on each side is exposed
ical reactions, or ionizing radiation). The to the light from its contiguous screen. Re­
term fluorescence is applied to that form member, the screen functions to absorb the
of luminescence produced when light is energy (and information) in the x-ray beam
emitted instantaneously (within 1 o-s sec of that has penetrated the patient, and to con­
the stimulation). If the emission of light is vert this energy into a light pattern that has
delayed beyond 1 G-8 sec, the term phos­ (as nearly as possible) the same information
phorescence is used. as the original x-ray beam. The light, then,
Fluorescence, as used in radiology, is the forms a latent image on x-ray film. The
ability of crystals of certain inorganic transfer of information from x-ray beam
salts (called phosphors) to emit light when to screen light to film results in some loss
excited by x rays. For many years the crys­ of information. In this and subsequent
tals, or phosphors, of importance in radi­ chapters we will discuss some factors in­
ology were calcium tungstate (the phos­ volved in the degradation of information
phor in intensifying screens) and zinc and describe what can be done to keep this
cadmium sulfide (the phosphor in fluores- to a minimum.


Construction Protective Layer. The protective layer

An intensifying screen has four layers: applied over the phosphor is made of a
plastic, largely composed of a cellulose
1. a base, or support, made of plastic or compound that is mixed with other poly­
cardboard mers. It forms a layer about 0.7 to 0.8 mils
2. a reflecting layer (Ti02) thick. This layer serves three functions: it
3. a phosphor layer helps to prevent static electricity; it gives
4. a plastic protective coat physical protection to the delicate phos­

The total thickness of a typical intensifying phor layer; and it provides a surface that

screen is about 15 or 16 mils (1 mil= 0.001 can be cleaned without damaging the phos­

in. = 0.0254 mm). phor layer.

Base. The screen support, or base, may Figure 9-1 shows a cross section of a typ­

be made of high-grade cardboard or of a ical par speed intensifying screen with a

polyester plastic. The base of Du Pont in­ plastic base. Unlike the highly mechanized

tensifying screens is a polyester plastic (My­ operation of x-ray film production, screens

lar*) that is 10 mils thick (10 mils = 254 are largely a product of hand labor.

f.Lm). Kodak X-Omatic and Lanex screens

have a similar (Estart) base 7 mils thick.
Reflecting Coat. The light produced by The original phosphor used in x-ray in­
the interaction of x-ray photons and phos­ tensifying screens was crystalline calcium
phor crystals is emitted in all directions. tungstate (CaW04). New (since about
Much of the light is emitted from the 1973) screen phosphor technology is being
screen in the direction of the film. Many developed to increase screen speed over
light photons, however, are also directed that available with calcium tungstate, and
toward the back of the screen (i.e., toward has resulted in the introduction of a be­
the base layer) and would be lost as far as wildering variety of screens and corre­
photographic activity is concerned. The re­ sponding films. Let us first consider cal­
flecting layer acts to reflect light back to­ cium tungstate and then review some of the
ward the front of the screen. The reflecting new phosphors.
coat is made of a white substance, such as Natural calcium tungstate (scheelite) is
titanium dioxide (Ti02), and is spread over no longer used because synthetic calcium
the base in a thin layer, about 1 mil thick. tungstate of better quality is produced by
Some screens do not have a reflecting layer fusing sodium tungstate and calcium chlo­
(e.g., Kodak X-Omatic fine and regular). ride under carefully controlled conditions.
Phosphor Layer. The phosphor layer, The first commercial calcium tungstate
containing phosphor crystals, is applied screens were made in England and Ger­
over the reflecting coat or base. The crys­ many in 1896; they were first made in the
tals are suspended in a plastic (polymer) United States in 1912. The calcium tung­
containing a substance to keep the plastic state crystal must be absolutely free of any
flexible. We will discuss the phosphor in contaminant if it is to fluoresce properly. It
more detail later. The thickness of the
phosphor layer IS
· about 4 m1'ls 10r
c par LAYER 0.7-o.s l


{ �mz�:z::!I
speed screens. The thickness of the phos-
phor layer is increased 1 or 2 mils in high- 4-6
� ·

speed screens, and is decreased slightly in

10 mil{
detail screens.

*Mylar is a trademark of E. I. du Pont de Nemours

and Company, Inc. Figure 9-1 Par speed x-ray intensifying
tEstar is a trademark of Eastman Kodak Company. screen (cross section)

produces light primarily in the blue re­ the conversion efficiency of the phosphor
gion of the visible spectrum, with a wave­ are known, the number of light photons
length range of 3500 to 5800 A (1 ang­ generated is easily calculated. For example,
strom = 0.0001 f.L = 0.00000001 em) and assume a 50-keV x-ray photon (50,000 eV)
a peak wavelength of about 4300 A (430 is absorbed in a calcium tungstate screen
nm), which is seen by the eye as a violet that emits most of its light at a wavelength
color. Although the eye is not sensitive to of 4300 A. The energy in electron volts
light of this wavelength, x-ray film emul­ (eV) of this light photon is calculated by
sion exhibits maximum sensitivity to light
from calcium tungstate screens. Figure 9-2 A. (wavelength)
= keV
diagrams the spectrum of calcium tung­ 12.4
4300 =
state fluorescence, the response of the eye keV
to light of different wavelengths, and the 12.4
keV = -- = 0.003
sensitivity of x-ray film. Film sensitivity is 4300
eV= 3
seen to be high throughout most of the
range of light emitted by the screen, a fact The energy of a 4300 A (430 nm) blue light
that ensures maximum photographic ef­ photon is about 3 eV. At 100% efficiency,
fect. Note that the film does not exhibit a 50,000 eV x-ray photon would produce
photosensitivity to red light, so red light about 17,000 light photons of 3 eV energy:
can be used in the darkroom without pro­
ducing any photographic effect on the film. -- = 17,000
Intensifying Action of Screens. An in­
tensifying screen is used because it can con­ Because the conversion efficiency of cal­
vert a few absorbed x-ray photons into cium tungstate is only 5%, the actual num­
many light photons. The efficiency with ber of light photons emitted by this phos­
which the phosphor converts x rays to phor is about 850 (17,000 X 0.05 = 850).
light is termed the intrinsic conversion One reason some new intensifying screen
efficiency of the phosphor; this is more phosphors are faster than calcium tung­
accurately defined as the ratio of the light state is that the new phosphor has a higher
energy liberated by the crystal to the x-ray conversion efficiency (up to 20%).
energy absorbed. The intrinsic conversion The ability of light emitted by the phos­
efficiency of calcium tungstate is about 5%. phor to escape from the screen and expose
If the energy of the absorbed x-ray photon, the film is termed the "screen efficiency."
the wavelength of the emitted light, and For typical screens about half the gener­
ated light reaches the film; the rest is ab­
sorbed in the screen and is wasted. In the
previous example, only half the 850 light
photons generated would be able to escape
from the screen and expose the film.
To examine the manner in which the
screen amplifies the photographic effect of
x rays, let us consider the case of a film­
screen combination compared to a film
3000 4000 3000 6000 7000 (ANGSTROMS I alone, each irradiated by one thousand 50-
VIOLET keV x-ray photons. Of the 1000 photons,
a high-speed calcium tungstate screen will
Figure 9-2 The spectral emission of a calcium
absorb 40% (we will discuss this aspect in
tungstate x-ray intensifying screen compared to
the spectral sensitivity of x-ray film and of the more detail later), or 400. Each of the 400
eye x-ray photons will produce 850 light pho-

tons, for a total of 340,000 light photons. sification factor of a screen is the ratio of
Half of these light photons ( 170,000) will the x-ray exposure needed to produce the
escape the screen and expose the film. same density on a film with and without
Let us assume that 100 light photons are the screen (intensification factor is com­
needed to form one latent image center. monly determined at a film density of 1.0).
The screen-film system in our example has exposure required when
caused the formation of 1700 latent image Intensification screens are not used

centers. (We will discuss the concept of the factor exposure required with
latent image in Chapter 10.) At this point,
let us simply define a latent image center The intensification factor of CaW04
as the end product of the photographic ef­ screens will increase with kVp of the x-ray
fect of light or x rays on the film emulsion. beam (Fig. 9-3). The K-absorption edges
It is possible to measure the magnitude of of silver and bromine in the film are 26
the photographic effect of an exposure by keV and 13 keV, respectively, while the K­
counting the number of latent image cen­ absorption edge of tungsten is 69.5 keV.
ters formed as a result of that exposure. In Photoelectric absorption of low-kVp x rays
the case of x rays exposing film directly, will be relatively greater in film because of
only about 5% of the x-ray photons are the low-keV K-absorption edge of silver
absorbed by the film, so our example will and bromine. High-kVp x rays will be more
cause 50 x-ray photons to react with the abundantly absorbed by the photoelectric
film emulsion. It is usually considered that process in calcium tungstate screens. For
one latent image center is formed for each this same reason, a heavily filtered x-ray
x-ray photon absorbed by a film emulsion. beam will increase the screen intensifica­
The result of the same 1000 x-ray photons tion factor by removing the low-energy
acting directly on film is to produce only components of the beam. Thus, CaW04 in­
50 latent image centers, versus 1700 latent tensifying screens are relatively faster in
image centers resulting from the use of in­ radiography of a thick body part, such as
tensifying screens. In our example, the ra­ the lumbar spine, than when an extremity
tio of the photographic effect of screen ver­ is examined. That is, only higher energy x
sus nonscreen, or direct, exposure is 34: 1. rays can penetrate the thick part and reach
If kVp remains constant, direct film ex­ the screen.
posure will require 34 times as many mAs There will actually be a small number (3
as a film-screen exposure. or 4%) of x-ray photons directly absorbed
A review of the example of the intensi­ by the film when a screen-film combination
fying effect of screens is presented in Table is exposed. This direct action of x rays
9-l. Stated another way, patient exposure would create so few latent image centers
is decreased greatly when intensifying compared to those caused by the fluores­
screens are used. A measure of this de­ cence of the screen that it may be ignored
crease in exposure is termed the "intensi­ as having no detectable influence on total
fication factor" of the screen. The inten- film response to the exposure.

Table 9-1. Comparison of the Photographic Effect of 1000 X-Ray Photons Used With and
Without Intensifying Screens

Intensifying screen 400 340,000 170,000 1700

Direct exposure 50 50


KILOVOLTS PEAK Figure 9-4 Screen unsharpness is caused by

light diffusion in the intensifying screen
Figure 9-3 Variation of the intensification
factor with kV p
photons generated by this absorbed x-ray
photon are emitted in all directions. Not
Speed of Calcium Tungstate Intensify­ all the light will reach the film, but that
ing Screens. Several factors determine how which does will expose an area of the film
"fast" or "slow" a calcium tungstate screen that is much larger than the size of the
will be. These include thickness of the calcium tungstate crystal that emitted the
phosphor layer, size of the phosphor crys­ light (Fig. 9-4). In addition, some light
tals, presence or absence of light-absorb­ scattering takes place in the screen, and
ing dye in the phosphor layer, and phos­ further increases the area of illumination.
phor conversion efficiency. Of course, the The resultant light diffusion obviously
faster screen will allow a lower x-ray ex­ causes images to have less sharp borders.
posure to the patient, but a price for this Consider Figure 9-5. If an x-ray beam is
speed must be paid. The speed of a cal­ directed onto a film-screen combination
cium tungstate screen and its ability to that has a thick lead block covering half the
record detail are in reciprocal relation­ film, one would expect half the film to be
ship; that is, high speed means less detail. exposed and half to be entirely unexposed,
This statement is also true for new screen with the line between the two areas being
phosphors, but the subject is more complex very sharp. If the x-ray beam were to ex-
because higher speed does not always re­
quire a thicker screen. These screens are
classified as fast, medium (par speed), and X-RAY
slow (detail), with intensification factors in BEAM

the range of 100, 50, and 25, respectively.
A thicker phosphor layer will result in a FILM -SCREE�>-=LE= D B ":'O C
= ::::':':": K
L ':"=I=======
faster screen because the thick layer will COMBINATION

absorb more x-ray photons than a thin

layer. Thick screens will be faster but will
cause a decrease in the clarity of the image
recorded on the film. This decrease in
image clarity is primarily caused by diffu­ SLOW SCREEN 4 ) BLACKENING
sion of light in the phosphor layer. If a
thick phosphor layer is employed, an x-ray I
photon may be absorbed in the phosphor Figure 9-5 Unsharp image borders produced
at some distance from the film. The light by light diffusion in intensifying screens

pose film alone (without screens), the line lines and two spaces per mm. Each line is
between exposed and unexposed areas � mm wide, and each space is � mm wide,
would be very sharp. If the film is sand­ thus making each line pair � mm wide. (A
wiched between intensifying screens, how­ more detailed discussion of resolving
ever, the border is less sharp because some power is presented in Chapter 14). X-ray
light will diffuse into the area under the film is able to record up to 100 line pairs
edge of the lead block and cause exposure per mm, but the slowest screens can record
of film in an area that actually receives no only a little over 10.
x-ray exposure. This light, which diffused
Screen-Film Contact
under the edge of the lead block, is not
The cassette in which the intensifying
available to expose the portion of the film
screens are mounted provides a light-tight
that is not covered by the block. The final
container for the film. It also serves to hold
result is the production of a more gradual
the film in tight contact with the screens
change in density, which is unlike the ideal
over its entire surface. With good film­
abrupt transition from black to unexposed
screen contact a dot of light produced in
(clear) film. {Attempts to measure the mag­
the screen will be recorded as a comparable
nitude of this problem of light diffusion
dot on the film. If contact is poor, this dot
and scattering will lead us to consideration
of light will diffuse before it reaches the
of the concepts of line spread function and
film, so that its radiographic image is un­
modulation transfer function in Chapter
There is a simple method for testing
A thin screen causes less light diffusion
film-screen contact. A piece of wire screen
than a thick one because light photons are
is placed on the cassette, and a radiograph
produced closer to the film. Another way
of the wire screen is made. The sharpness
to decrease light diffusion is to incorporate
of the image of the wire in all regions of
a substance that absorbs light in the screen.
the film is compared, and any areas of poor
The substance is commonly a yellow dye,
film-screen contact become obvious be­
but the question of which color is most ef­
cause the image of the wire appears fuzzy.
fective has not been resolved. The light
The wire screen should be made of iron,
photons that emerge from the crystal im­
brass, or copper (aluminum and plastic
mediately adjacent to the film will obviously
screens fail to absorb enough x rays), and
travel the shortest distance before leaving
should have a heavier wire than that used
the phosphor layer. Scattered photons
in ordinary window screen. It is best ·to
must travel longer distances in the phos­
mount the test wire screen between two
phor layer, and thus they have a better
pieces of plastic, composition board, or
chance of being absorbed by the dye. The
cardboard. Poor film-screen contact will
dye will decrease the speed of the screen
occur quickly if the cassette is handled
because it decreases the amount of light
roughly, but eventually any cassette will
emitted. Light-absorbing dyes are included
wear out with routine use.
in screens designed to produce greater
detail. Cleaning
Intensifying screens must be kept clean.
Resolving Power Any foreign material on the screen, such
The maximum number of line pairs pet as paper or blood, will block light photons
millimeter that can be resolved by the and produce an area of underexposure on
screen-film system is called the resolving the film corresponding to the size and
power. A line pair means a line and a shape of the soiled area. Cleaning will elim­
space. For example, two lines (or two line inate the "high spots" on a screen; these
pairs) per mm means that there are two high points are the major source of exces-

sive wear. The cause of screen failure is beam (and is twice as fast as par speed).
mechanical attrition. Under normal con­ The Hi Plus screens are not twice as thick
ditions of use, x-ray photons will not dam­ as par screens; they are about 2.3 times as
age screens. Screens are best cleaned with thick as the par screens. Why is this? If 100
a solution containing an antistatic com­ x-ray photons struck a pair of par screens,
pound and a detergent; the solution should 80 photons would be transmitted through
be applied gently (never rub vigorously) the screens (20 are absorbed). A second set
with a soft lint-free cloth. The cassette of par screens placed in the path of the 80
should never be closed after cleaning until transmitted photons would absorb 20%, or
it is absolutely dry. 16 photons, allowing 64 to pass through.
Thus, a double thickness of par speed
screens absorbs 36% rather than 40% of
It is usually agreed that the maximum the x-ray beam. It is possible to continue
useful speed of calcium tungstate screens
to increase the speed of CaW04 screens by
has been achieved. New screen phosphors
increasing screen thickness, but unsharp­
have permitted greater speed, but this
ness caused by light diffusion in the thick
changing technology is complex and per­
phosphor layer becomes unacceptable. In
calcium tungstate screens a compromise is
made between speed and sharpness.
Increasing Screen Speed
Conversion Efficiency. In 1972, reports
We must first consider how intensifying
describing rare earth oxysulfide phosphors
screen speed can be increased. There are
for use in intensifying screens were pub­
three possible ways. The phosphor layer
lished.10 This work was supported by a con­
can be made thicker to absorb more of the
tract from the Atomic Energy Commission
x-ray beam. The phosphor may have a
to the Lockheed Corporation, and this class
higher conversion efficiency of x rays to
of phosphors is sometimes referred to as
light. The phosphor may be able to absorb
the Lockheed rare earth phosphors. Rare
the x-ray beam more effectively. To review,
earth intensifying screens have been com­
there are three possible ways to increase
mercially available since about 1973.
screen speed:
Chemists divide the periodic table of the
1. thicker phosphor layer elements into four basic groups: alkaline
2. higher conversion efficiency phos­ earths, rare earths, transition elements,
phor and nonmetals. The term "rare earth" de­
3. higher absorption phosphor veloped because these elements are diffi­

Before entering this long discussion of new cult and expensive to separate from the

phosphor technology, please review Table earth and each other, not because the el­
9-2, which catalogs the noncalcium tung­ ements are scarce. The rare earth group

state screens available from Du Pont and consists of elements of atomic numbers 57
Kodak, listing phosphor composition and (lanthanum) through 71 (lutetium), and in­

color of light emitted by the phosphor. cludes thulium (atomic number 69), ter­
Thicker Phosphor Layer. Thick phos­ bium (atomic number 65), gadolinium
phor layers absorb more of the x-ray beam (atomic number 64), and europium (atomic
than thin layers . This is the major way in number 63). Because lanthanum is the first
which the speed of calcium tungstate element, the rare earth group is also known
screens is increased. A pair of Du Pont Par as the lanthanide series . Lanthanum (La)
Speed CaW04 screens absorbs about 20% and gadolinium (Gd) are used in the rare
of the x-ray beam. A pair of Du Pont Hi earth phosphors. A related phosphor, yt­
Plus screens absorbs about 40% of the x-ray trium (Y), with atomic number 39, is not a

Table 9-2. Non-Calcium Tungstate Intensifying Screens Manufactured by Du Pont and Kodak

Du Pont Cronex Quanta Ill Lanthanum oxybromide:thulium acti- Blue

vated (LaOBr:Tm)
Quanta Detail Yttrium tantalate:thulium activated Ultraviolet/blue
Quanta Fast Detail Yttrium tantalate:niobium activated Ultraviolet blue
Cronex Quanta V Lanthanum oxybromide:thulium acti- Blue
vated plus gadolinium oxysul- and
fide:terbium activated (LaOBr:Tm green
plus Gd202S:Tb)
Kodak X-Omatic Fine Barium lead sulfate, yellow dye Blue
X-Omatic Regular Barium strontium sulfate:europium Blue
activated, neutral dye
Lanex Fine Gadolinium oxysulfide:terbium acti- Green
vated, neutral dye (Gd202S:Tb}
Lanex Medium Gadolinium oxysulfide:terbium acti- Green
vated, yellow dye (Gd202S:Tb)
Lanex Regular Gadolinium oxysulfide:terbium acti- Green
vated (Gd202S:Tb)

rare earth but has some properties similar quoted are CaW04 5%, LaOBr:Th 18%,
to those of the rare earths. Gd202S:Tb 18%, and Y202S:Tb 18%.
The rare earth phosphors are produced A fairly new addition to the family of
as crystalline powders of terbium-activated rare earth screens are the rare earth tan­
gadolinium oxysulfide (Gd202S:Tb) and talates, designated by the generic formula
thulium-activated lanthanum oxybro­ LnTa04 where Ln is any of the lanthanide
mide (LaOBr:Th). Unlike CaW04, the rare series. Tantalate phosphors can be made
earth phosphors do not fluoresce properly with most of the lanthanides, but the com­
in the pure state. For example, maximum mercial screens that have been made use
fluorescence of Gd202S occurs when about yttrium tantalate (YTa04}.
0.3% of the gadolinium atoms are replaced
Table 9-2 lists two tantalate screens of
Du Pont manufacture. In the YTa04:Tm
by terbium.
(Quanta Detail) some of the Y ions have
The x ray to light conversion efficiency
been replaced by Tm ions. In the
of rare earth phosphors is significantly
YTa04:Nb (Quanta Fast Detail) some of
greater than that of CaW04• It is generally
the Ta04 ions have been replaced by Nb04
correct to state that the x ray to light con­
ions. The brightest phosphor involves Nb
version efficiency of CaW04 is about 5%,
substitution (YTa04:Nb), which has an es­
whereas that of the rare earth phosphors
timated x ray to light conversion efficiency
is about 20%. Stated another way, one ab­ of 11%.
sorbed x-ray photon in a CaW04 intensi­ Higher Absorption. The fraction of the
fying screen will produce about 1000 light x-ray beam absorbed by a pair of par speed
photons, and the same photon absorbed in CaW04 screens is about 20%, high-speed
a rare earth screen will produce about 4000 CaW04 screens about 40%, and rare earth
light photons. The conversion efficiency of screens about 60%. The increased absorp­
yttrium oxysulfide screens is about the tion in the faster CaW04 screens occurs
same as that of gadolinium. In the litera­ because a thicker screen is used; the in­
ture, actual conversion efficiency figures creased absorption in rare earth screens is

z Thanks, Jack!
a result of improvement in the absorption 0
characteristics of the phosphor. �
At diagnostic x-ray energies, absorption 0
is almost entirely caused by the photoelec­ <( k-edge Tungsten
tric effect in the high atomic number ele­
ments of the phosphor. As discussed in

Chapter 4, a photoelectric reaction is most I

likely to occur in elements with high atomic z
numbers, and when the x-ray photon en­ w
ergy and the binding energy of the ejected (/)
K-shell electron are almost the same. >
A phosphor with a higher atomic num­ <(
ber would have higher absorption. Because n::

the atomic numbers of tungsten (74) and 25 50 75 100

lead (82) are almost at the end of the pe­ X-RAY ENERGY
riodic table, the potential for this type of
improvement is limited. It is interesting to Figure 9-6 Approximate x-ray absorption as
a function of x-ray photon energy in a calcium
note that the improvement in the input
tungstate intensifying screen
phosphor of image intensifier tubes in­
volved a change from zinc sulfide (atomic tage over CaW04, which persists until the
numbers 30 and 16) to cesium iodide 70-keV K edge of W is reached. It is ap­
(atomic numbers 55 and 53). Actually, each parent that Gd, which has a lower atomic
of the phosphors used in the new fast in­ number than W, enjoys an absorption ad­
tensifying screens has an atomic number vantage over W in the x-ray energy range
lower than that of tungsten. Table 9-3 lists of 50 to 70 keV because of the differences
some of the elements found in intensifying in photoelectric absorption related to the
screens, together with their atomic number K-shell electron binding energies of these
(Z) and K-shell binding energy. two elements. In diagnostic radiology,
Figure 9-6 diagrams the approximate x­ however, a large percentage of x rays in
ray absorption curve of a calcium tungstate the beam will have an energy less than 50
screen. Notice that the screen is less ab­
sorbing as radiation energy increases until
z k-edge Gadolinium
the 70 keV K edge of tungsten is reached. 0

In Figure 9-7 the approximate absorp­

�" Gd202S
tion curve of Gd202S:Tb is added to the (!) :�\/
CaW04 curve. Up to 50 keV, both phos­
\ k-edge Tungsten

phors absorb about the same. But, at the

50.2-keV K edge of Gd, the rare earth 0 ' '
I ' '
screen develops a four- or five-fold advan- z
' '
Table 9-3. Atomic Number and K-Shell (fJ
Binding Energy of Some Screen Phosphors >


25 50 75 100
Yttrium (Y) 39 17.05 X-RAY ENERGY
Barium (Ba) 56 37.4 keV

Lanthanum (La) 57 38.9

Figure 9-7 Comparison of the approximate
Gadolinium (Gd) 64 50.2
x-ray absorption as a function of x-ray photon
Tungsten (W; wolfram) 74 69.5
energy in CaWO. and Gd202S

keV, making the advantage of gadolinium

much less. This is why lanthanum is added
to the phosphor of some rare earth screens
(such as the Du Pont Cronex Quanta V ).
Lanthanum, with its Kedge at 39 keV, ef­
fectively "closes the window" between 39
( -edge Tungsten

and 50 keV.
In Figure 9-8 the approximate x-ray ab­
sorption curves of LaOBr, Gd202S, and
CaW04 are compared. By combining the
two rare earth phosphors, the rare earth
. screen has a significant absorption advan­
tage over CaW04 in the 39- to 70-keV
Figure 9-9 shows the approximate ab­ 25 50 75 100
sorption of yttrium oxysulfide (Y202S:Tb
is the phosphor in the GAF Rarex B Mid­
speed screens) compared to that of CaW04• Figure 9-9 Comparison of the approximate
x-ray absorption curves of CaW04 and Y202S
Notice that CaW04 and Y202S match from
17 to 70 keV. Although these two phos­
phors absorb about the same x-ray energy, gion of increased absorption for these

the Y202S screen produces more light pho­ phosphors. This is the principal reason that
tons because of its higher x ray to light these phosphors have higher absorption of
conversion efficiency (18 versus 5%). x rays used in diagnostic radiology as com­

Let us summarize this section. The K pared with CaW04, even though the
edges of Ba, La, and Gd correspond closely atomic number of tungsten (74) is higher
to the maximum intensity of x rays in the than that of barium (56), lanthanum (57),
primary beam, causing much of the energy or gadolinium (64). The absorption for yt­
in the primary beam to decrease in the re- trium (39) is the same as that of tungsten
from 17 to 70 keV, but yttrium is "faster"
because of a higher conversion efficiency.
k-edge Gadolinium

� Emission Spectrum
This section will deal with the wave­
length (color) of light emitted by the vari­
ous phosphors, an important consideration
because the light output of the screen and
the maximum sensitivity of the film used
must be matched.
The spectrum of light emitted by some
of the new screen phosphors is quite dif­
k-edge _j
Lanthanum ferent from that of CaW04, which pro­
duces a continuous spectrum of light in the
blue region with a peak wavelength of
25 50 75 100
X-RAY ENERGY about 4300 A (430 nm) shown in Figure
keV 9-2. Some phosphors other than calcium
Comparison of the approximate
tungstate also produce light in the blue­
Figure 9-8
x-ray absorption curves of CaW04, LaOBr, and violet range. Table 9-4 lists some of these
Gd202S phosphors. Natural silver halide films are

Table 9-4. Spectral Emission of Several Screen Phosphors


Tungsten 430 Du Pont Cronex family

Barium lead sulfate 360 Kodak X-Omatic Fine
Du Pont Hi Speed
Barium strontium sulfate 380 Kodak X-Omatic Regular
Yttrium tantalate:niobium 410 Du Pont Quanta Fast Detail

maximally sensitive to light of this wave­ peaks in the blue, blue-green, yellow, and
length. Figure 9-10 shows in graphic form red areas. Because standard x-ray silver
the approximate spectral emission of the halide film will not absorb (i.e., is not sen­
screen phosphors listed in Table 9-5, com­ sitive to) light in the green area, a special
pared to the sensitivity of natural silver hal­ film must be used with these screens. Such
ide film (such as Du Pont Cronex films and films are Kodak Ortho G and Du Pont Cro­
Kodak X-Omatic films). Keep in mind that nex 8. (This subject will be explored in
the speed of these screens covers a wide more detail in Chapter 11.) Figure 9-11 is
range. a graph of the spectral emission of the Ko­
We must now consider the light emission dak Lanex regular screen (which contains
of the rare earth phosphor Gd202S:Tb. Gd202S:Tb in its phosphor) and the sen­
The spectral emission of this earth phos­ sitivity of Kodak Ortho G film. Because the
phor is produced by the terbium ion. The sensitivity of natural silver halide film stops
terbium emission is not a continuous spec­ at about 500 nm, use of this film with some
trum (as is CaW04) but is concentrated in rare earth screens would result in loss of
narrow lines with a very strong peak at 544 much of the speed of these screens because
nm, which is in the green light part of the the film would be insensitive to much of
spectrum. There are less intense emission the light produced.


(/) t-
w z
z "(/)
w � ...if;\
a:: _.J
(/) LL.

w w
> >
t- t-
<l: <l: .....

_.J _.J
w w
a:: a::
250 300 350 400 450 500

Figure 9-10 Approximate spectral emission of several different intensifying screen phosphors
compared to the natural sensitivity of silver halide

Table 9-5. Speed Class of Various Intensifying Screens


Du Pont Cronex Par Speed cawo. Blue Cronex 4 100

Cronex Hi Plus cawo. Blue Cronex 4 250
Cronex Quanta Ill LaOBr:Tm Blue Cronex 4 800
Cronex Quanta V LaOBr:Tm Blue Cronex 4 320
Gd202S:Tb Green Cronex 8 400
Quanta Detail YTaO.:Tm Ultraviolet/ Cronex 4 100
Quanta Fast Detail YTaO.:Nb Ultraviolet/ Cronex 4 400

Kodak X-Omatic Fine BaPbSO., Blue XRP 32

yellow dye
X-Omatic Regular BaSrSO. :Eu, Blue XRP 200
neutral dye
Lanex Fine Gd202S:Tb, Green Ortho G 100
neutral dye
Lanex Medium Gd202S:Tb, Green Ortho G 250
yellow dye
Lanex Regular Gd202S:Tb Green Ortho G 400
Note: Dupont Cronex 4 is now largely replaced by Cronex 7 and Cronex 10, and Kodak Ortho G
film by T-Mat G. The older films still accurately reflect relative intensifying screen speeds, which is
the purpose of this table.

The spectral emtsston of yttrium oxy­ nm, though, is still present. Yttrium oxy­
sulfide screens is similar to that of the rare sulfide screens may be used with natural
earth screens in that it shows line emission, silver halide films, but will exhibit even
but a large fraction of the emission from a more speed when used with green-sensitive
Y202S:Tb screen falls within the sensitivity films.
range of natural silver halide films. The The Du Pont Cronex Quanta V, contains
principal emission of the terbium at 544 two phosphors, terbium-activated gadolin­
ium oxysulfide (Gd202S:Tb) and thulium­
activated lanthanum oxybromide
(LaOBr:Tm). It is interesting to note that
the color of light emitted by a rare earth
phosphor depends on the activator used;
terbium (Tb) produces green light and thu­
lium (Tm) blue. For example, Gd202S:Tb
is a green-emitting intensifying screen
phosphor, and LaOBr:Tm is a blue-emit­
ting intensifying screen phosphor. The
300 400 500 600
spectral emission of LaOBr:Tm is illus­
trated in Figure 9-12. The Du Pont
WAVELENGTH in nm Quanta V intensifying screen is interesting
in that it contains both a blue (LaOBr:Tm)
Figure 9-11 Approximate spectral emission
and a green (Gd202S:Tb) emitting phos­
of the Kodak Lanex Regular screen (rare earth)
compared to the sensitivity of Kodak Ortho G phor. This "double" spectral emission al­
film. (This is similar to charts supplied through lows a wider choice of suitable films, be­
the courtesy of the Eastman Kodak Company) cause both blue- and green-sensitive films

are matched to the spectral emission of the 100. Remember that a film that matches
screens. the spectral emission of the screen must be
An example of a rare earth screen with used. In Table 9-5 Du Pont Cronex 4 and
a blue-emitting phosphor is the Du Pont Kodak XRP are films sensitive to blue light.
Cronex Quanta III, which has LaOBr:Tm Green-light-sensitive films are represented
at its phosphor (see Fig. 9-12). Recent ad­ by Du Pont Cronex 8 and Kodak Ortho G
vances in separating rare earths have made films. When comparing the relative speed
suitably pure supplies of lanthanum avail­ of various intensifying screens, the films
able for use in intensifying screens . that were used must be specified. (In Chap­
The yttrium tantalate screens emit light ter 11 the characteristics of these films will
in the ultraviolet and blue wavelengths. Yt­ be presented in more detail, and in Chap­
trium tantalate (YTa04) has an inherent ter 14 we will examine potential problems
broadband emission with its peak at 337 associated with very fast screens.)
nm. Activation with thulium (YTa04:Tm)
adds another peak at 463 nm. Activation Response to Kilovoltage
with niobium (YTa04:Nb) produces a
In the past we compared the speeds of
broadband emission with its peak at 410
other screens to that of CaW04 as if the
nm. The emission spectrum of YTa04:Nb
CaW04 screen had a constant response to
is illustrated in Figure 9-10.
kilovoltage. The response of CaW04 to
Table 9-5 is an attempt to compare
kVp is not flat, but drops down at low kVp.
roughly the relative speed (speed class) of
This is illustrated in Figure 9-13, which
the newer screens we have been discussing.
shows the kVp response of CaW04,
(The term "speed class" will be discussed
LaOBr:Tm, and a combined Gd202S:Tb
in more detail in Chapter 11.) It is common
and LaOBr:Tm phosphor. Figure 9-13
to express the speed of an intensifying
shows that the speed of the rare earth
screen relative to that of a par speed cal­
screens varies much more with kVp than
cium tungstate screen (specifically, the Du
Pont Cronex Par Speed screen). In our
speed class scheme the speed of the par
2� . . . . . . .... . . .
speed CaW04 screen is assigned a value of .. . . ..... . .. ... .. ...

t ........... ----- --
.··,.--- ......
� .....
2.0 /.
Cl 1.'
w� ,_.. LaOBr: Tm /Cronex 7
r463nm wl ,_..
a. a:: I·
' (Quanta ill)
(f) E 1.6 ti\
0 374nm z !/) f.: Gd202S:Tb �
Ui + ww 0 t: plus LaOBr:Tm Cranex 8
"0 ,_..
� a:: Q) 1.2 Ii (Quanta 12:)
w (.) !/) I:
(f) !/)
z t:

w :::E a.
CaW04 I Cronex 4
5 ...J
� Q8 li
(/) G: , . (Hi Plus)
w 303 nm
> +

60 80 100 120 140

300 400 500 600 kVp
WAVELENGTH (nm) Figure 9-13 Response to kVp of a Du Pont
CaWO. (Hi Plus), LaOBr:Tm (Quanta III), and
Figure 9-12 Approximate spectral emissiOn dual phosphor Gd202S:Tb plus LaOBr:Tm
of LaOBr:Tm. (Data courtesy of E.I. du Pont (Quanta V) screen. (Data courtesy of E.I. du
de Nemours & Company, Inc.) Pont de Nemours & Company, Inc.)

does the CaW04 screen. ote that the rare timers with very fast screens will usually
earth screens show maximum speed at require special consideration, and one
about 80 kVp, with less speed at both low should plan for appropriate changes in ad­
and high kilovoltages. The need for ex­ vance .
posure technique adjustment because of
screen speed variation with kVp is not Quantum Mottle
great, and is usually of little importance Quantum mottle, commonly called
unless kilovoltage lower than 70 kVp is "noise;' results from statistical fluctua­
used. It is generally advisable to establish tion in the number of x-ray photons ab­
constant kVp (with variable mAs) exposure sorbed by the intensifying screens to form
technique charts when using noncalcium the light image recorded on film. With the
tungstate screens. fast screens and films now available, it is
Figure 9-13 introduces two areas we will possible to use a film-screen system so fast
not explain until Chapter 11. Film-screen (i.e., the screen does not .have to absorb
speed is often expressed as the reciprocal many x rays) that noise makes the resulting
of the exposure in milliroentgens required image unsatisfactory. (This subject will be
to produce a net film density of 1.0. This reviewed in painful detail in Chapter 14).
explains the numbers used to express film­ At this time, we only wish to emphasize that
screen speed. The speed of a film-screen any discussion of fast screens demands
system depends on the speed of the screen careful evaluation of system noise. One
and on the speed of the film. The films pays a price for speed.
used to determine the curves in Figure
9-13 were a medium-speed and half-speed PHOTOSTIMULABLE PHOSPHOR
blue-sensitive film (Cronex 4 and Cronex Conventional radiography uses an x-ray
7) and a medium-speed green-sensitive intensifying screen, x-ray film, or both as
film (Cronex 8). We will repeat Figure 9-13 the image receptor. The exposed film is
in Chapter 11; don't worry if these areas then developed and viewed. Now, a process
are not entirely clear now. sometimes called computed radiography
(CR) uses a photostimulable phosphor as
Phototimers the image receptor. A phosphor that is cur­
Some phototimers being used today will rently used is composed of europium-ac­
not work well with rare earth screens. Many tivated barium fluorohalide. The phosphor
of these phototimers have a response time crystals are coated on a screen that looks
of about 30 milliseconds (i.e., it takes 30 like an intensifying screen. The phosphor
milliseconds for the phototimer to measure coated screen is contained in a cassette sim­
the radiation and terminate the exposure). ilar to standard film-screen cassettes. A ra­
With fast screens a response time of 3 mil­ diographic exposure is made using con­
liseconds is recommended.3 In addition, ventional x-ray equipment, but here the
the phototimer must be able to recognize similarity to conventional radiography
differences in screen speed caused by var­ ends . The photostimulable phosphor ab­
iation in kVp. If a photomultiplier-type sorbs some of the energy in the x-ray beam
phototimer is used, the phosphor in the and stores a portion of this energy as va­
fluorescent screen used to convert x rays lence electrons stored in high energy traps
to light should be matched to the phosphor to create a latent image. When this latent
in the intensifying screen. The photomul­ image is scanned by a laser beam, the
tiplier tube selected should be sensitive to trapped electrons return to the valence
the light output of the phosphor. Ioniza­ band with the emission of light. This light
tion chamber-type phototimers can be is viewed by a photomultiplier tube, and
made to work satisfactorily. Use of photo- the output of the photomultiplier tube con-

stitutes the signal. The signal is fed to a diography (photostimulable suggests that
computer where it is processed and stored, luminescence is going to be induced by ab­
and may be displayed in any of the normal sorption of light photons).
formats. This technology is sometimes The material used in one commercially
called "filmless," or "electronic," or "com­ available photostimulable phosphor is eu­
puted" radiography. ropium-activated barium fluoride bromide
(BaFBr:Eu). Europium is an activator
Luminescence added to BaFBr. (In the discussion of semi­
Luminescence describes the process of conductors, added materials were called
emission of light (optical radiation) from a "impurity atoms," but they perform the
material caused by some process other than same function as activators.)
heating to incandescence. Luminescent The reason for using an activator is to
materials can absorb energy, store a frac­ make traps at the activator site in the crys­
tion of it, and convert the energy into light, talline structure. Other types are produced
which is emitted. when atoms are missing from their normal
There are two forms of luminescence position in the lattice structure. These traps
that are determined by the time it takes for are localized energy levels at the activator
light to be emitted after an exciting event or vacancy site that do not normally exist
takes place. In fluorescence, light is emitted in the crystalline structure. If by chance an
within 1 o-s sec of the stimulation. This oc­ electron is captured in the trap, it behaves
curs when an electron is excited from one very much like an electron trapped
energy level, and then returns to that en­ (bound) in an atom. We remember that to
ergy level without anything additional hap­ remove an electron bound to an atom, we
pening to it. must supply it with the ionization energy
Phosphorescence is the emission of light of that particular atom. In doing so, we
that is delayed beyond 1 O-s sec. This delay produce a positive ion (which is the atom
requires some additional interaction of the without an electron) and an electron that
electron; normally this is described as a is free to move away from the atom. The
trapping of the electron in some localized same is true for an electron that is trapped
energy state. Intensifying screens are made at an activator or vacancy site. If we want
from fluorescent material that emits its to free the electron from the trap, we must
light immediately after stimulation. This is supply it with sufficient energy to free itself
just fine, because film is there to capture from the trapping mechanism. If we sup­
the light. However, if we would like the ply the energy, the electron is free to move
screen to retain the x-ray image, we would through the crystalline structure, and it
use a phosphorescent material. In radiol­ leaves behind an empty trap (that looks like
ogy, there are two such applications. One a positive ion) ready to capture another
is thermoluminescent dosimetry (TLD), in electron. A photostimulable phosphor is a
which the phosphor is encouraged to emit material that has activator and vacancy
its light by heating. A thermoluscent do­ traps from which electrons may be re­
simeter is a small chip of activated lithium moved by absorbing energy from light
fluoride that is exposed to radiation. The photons. Without looking further into the
chip is then heated, and emits light in pro­ physics of trapping mechanisms, we will
portion to the amount of radiation it re­ describe how a photostimulable phosphor
ceived (the term "thermoluminescence" in­ plate is used in clinical radiography.
dicates heating to liberate luminescence). Let us acknowledge that we have not pre­
The second application of phosphores­ sented a complete and totally accurate pic­
cence is our current topic of photostimu­ ture of the physics of BaFBr:Eu. We would
lable phosphors as used in computed ra- like to quote Hartwig Blume from the Phil-

ips PCR Technical Review, Issue No. 1, p. high intensity sodium discharge lamps.
3, to give a more accurate picture of the This erases any image remaining from a
trapping of electrons and holes in the pho­ prior exposure.
tostimulable image plates. We have no de­ The plate is now exposed to x rays, using
sire to get into the physics of such things standard radiographic equipment. During
as hole traps and vacancy traps. the exposure, electrons in the phosphor
are excited into a higher energy state.
Let me first describe a bit of the physics of
the stimulable phosphor x-ray detector: The About half of these electrons return to the
majority of x-ray photons which interact with normal energy state instantly and are no
the phosphor material interact via the pho­ longer available for image formation. The
toelectric effect and create a high-energetic
rest of the excited electrons are trapped
photoelectron. The photoelectron in turn,
through a cascade process, generates a large
and form the latent image. It seems self
number of low energetic electrons and holes. evident that the number of trapped elec­
(A hole is a missing electron in the crystalline trons per increment of area is directly pro­
structure of the phosphor.) In an ordinary portional to the amount of x rays absorbed
intensifying screen, the electrons and holes
(otherwise, this would not be an imaging
recombine with each other and create "spon­
method). The trapped electrons may be
taneous" luminescence which exposes the x­
ray film. thermally removed from the trap on a sta­
In our stimulable phosphor, about half of tistical basis. The thermal agitation of elec­
the electron-hole-pairs recombine at euro­ trons out of the traps will degrade the la­
pium centers during the x-ray exposure and
tent image. In BaFBr:Eu, the image will be
create spontaneous, purple or near-ultravi­
olet luminescence characteristic for euro­ readable for up to about 8 hours at room
pium in BaFBr. The spontaneous lumines­ temperature. We should note that the
cence is not utilized. The other half of the image should be retrieved as soon as pos­
electrons and holes are trapped separately, sible to prevent degradation due to thermal
holes at divalent europium sites, electrons at
halogen vacancies in the crystalline lattice of
the phosphor. The local concentration of
The image is retrieved by putting the
trapped holes and electrons is proportional imaging plate in a reader. The reader is
to the local x-ray exposure and represents the made so that the plate is scanned by a small
latent image of the x-rayed object. dot ( a bo ut 100 microns) of light from a
By stimulation with red light, electrons can
helium-neon laser (633-mm wavelength).
be liberated from their traps. The electrons
The trapped electrons exposed to this
may migrate to trapped holes and recombine
with them to generate europium lumines­ small dot of laser light will absorb the laser
cence. To enhance the efficiency of the stim­ light energy. This addition of energy will
ulating light, the phosphor screen is actually free the electrons from their traps, and
prepared with a very high, uniform and sta­
they will become free to return to the lower
ble concentration of trapped electrons at hal­
energy state associated with the crystalline
ogen vacancies. Obviously, by stimulation
with (red) light, the latent image is destroyed . structure. Note that the crystalline state of
the lattice structure has a lower energy level
Photostimulable Phosphor Imaging than the electron trap. When the electrons
Let us now look at the process of pho­ return to this lower energy level they will
tostimulable phosphor (or computed ra­ emit a light photon. Since the energy level
diographic) imaging by steps as follows: of the normal crystalline structure is lower
than the activator energy level, the light
The phosphor prior to exposure
photon emitted on return to the lower en­
Formation of the latent image
ergy level will have more energy (shorter
Detecting the image
wavelength) than the energy of the laser
To prepare a BaFBr:Eu plate for imag­ light used to excite them out of their traps.
ing, the plate is flooded with light from That is to say, the stimulated emission has

a wavelength that is shorter than the ex­ about 10,000:1 (104:1). Photostimulable
citing laser beam. In BaFBr:Eu, this wave­ phosphors are said to have a very large
length is a narrow band in the 300 nm (ul­ dynamic range compared to film. In ad­
traviolet) to 500 nm (greenish) range. dition, the dynamic range is linear. This
That the two lights (the laser light and means that the amount of photostimulable
stimulated emission) have different wave­ luminescence created by an exposure in­
lengths is critical for image retrieval. After creases, or decreases, in direct proportion
all, we have an intense red laser light and to the amount of x-ray exposure the phos­
a somewhat feeble greenish light in the phor receives. The actual exposure range
same vicinity, and we would like to detect over which this linear response is seen cov­
the greenish light as the image signal. One ers a range of about 5 microroentgens (5
way to do this is to use a filter that will x 10-3 mR) to about 50 milliroentgens (5
absorb the red light but be transparent to X 101 mR), yielding an exposure range of
the greenish light. If, after the filter, we about 104:1. Figure 9-14 is a rough
use an optical fiber, we can conduct the graphic representation of this concept. No­
signal light to a photomultiplier tube at a tice that as the exposure increases from 5
remote position so that the photomultiplier X 10-3 mR to 50 mR (horizontal axis of
tube and the laser light do not interact. In Fig. 9-14), the intensity of the resulting
reality, we would expect to have a long fil­ photostimulable luminescence increases in
ter and a fiber bundle so that we would a linear fashion from a relative value of 1
scan the laser across the image plate to pro­ to a value of 104• At this time you may want
duce a scan line. The entire plate can be to take a glance ahead to Figure 11-5 and
read, a scan at a time, by moving the image notice that film has a more narrow dynamic
plate perpendicular to the scan line of the range, and a non-linear response to ex­
laser beam. posure at both the low and high ends of
In the output of the photomultiplier the exposure range. This concept of dy­
tube, we have a continuous point-by-point namic range (we will often call this "ex­
scan of the image. The output is an elec­ posure latitude") will be re-examined in
trical analog signal corresponding to ab­ much more detail in Chapter 11.
sorbed x rays in the image plate. This signal What is the advantage of a system with
must be amplified, converted to a digital a large, linear dynamic range? The most
signal, and stored in a computer. From practical advantage deals with the techno!-
here on, the imaging is just like any other
imaging method whose information is
stored in a computer before being read by ()
z 104
a radiologist. w
u. ()
0 (J)
Dynamic Range of Photostimulable � z 10 3
(ii �
Phosphors z ::::l
w 2
When we study x-ray film in Chapter 11, I- w 10
� ...J
we will find that high-contrast x-ray film w <(
can record exposure differences of about > ...J
i= � 10
100:1 (102:1). That is, the exposure that ::s i=
w (J)
causes the film to be developed as very a: 0
black is about 100 times greater than the
I- 100
� 10-3 10-2 10-1 10° 101 102
exposure that causes such a faint gray that c..
it can barely be detected by the eye. It has
been found that the exposure range that a Figure 9-14 The dynamic range of a photo­
photostimulable phosphor can record is stimulable phosphor imaging plate

agist's correct choice of exposure factors 2. higher conversion efficiency phos­

(kVp and mAs) for a radiographic exami­ phor
nation. The wider the range of exposures 3. higher absorption phosphor
the imaging system will accept, the more
CaW04 screen speed is determined largely
latitude there is in choice of kVp and mAs.
by the thickness of the phosphor layer. The
A major problem in radiology departments
newer fast phosphors exhibit a higher con­
results from the need for repeat exami­
version efficiency. Higher absorption of x
nations because of incorrect exposures. A
rays in the diagnostic range is a function
wide latitude system will decrease the need
of the matching of the K-absorption edge
for repeat exposures.
of the screen phosphor to the energy spec­
Because of this wide dynamic range, it is .
t:um m the x-ray beam. The spectral emis­
necessary to process photostimulable phos­
siOn of some phosphors requires that an
phor �mages in two steps. First, the imaging
appropriately sensitized film be used, or
plate IS scanned by a weak laser beam about
much of the light will be wasted. Intensi­
1.8 mm wide in an operation called "pre­
fying screens are usually used in pairs, with
reading. " Only a small portion of the image
a double-emulsion x-ray film sandwiched
stored in the plate is read out at this time.
befween the two screens in a light-tight cas­
The data in this pre-reading image is an­
sette. System noise becomes important
alyzed to compute the exposure level and
when the fastest film-screen systems are
exposure range of the stored image. In this
way, such things as the sensitivity of the
Computed radiography uses a photo­
photomultiplier tube can be adjusted to the .
sumulable phosphor as the image receptor.
�xp.o�ure level a�d exposure range for any Th phospho is composed of europium
mdividual exammation. This pre-reading � :
activated banum ftuorohalide, which is
thus allows examinations made with a wide
coated on an imaging plate. The phosphor
range of mAs settings to be converted into
diagnostically useful images. As a result,
�emporanly . stores a latent image on the
Imagmg plate. The latent image is con­
underexposed and overexposed as well as
verted to a light image using laser stimu­
correctly exposed examinations will be
lated luminescence. A photomultiplier
processed as satisfactory images. Digital
tube conve:ts the light image into an analog
processing of these images can produce a .
. electncal signal. This analog signal is am­
vanety of other image changes which we
plified, converted to a digital sig�al, and
will consider in another chapter.
stored in a computer.

l. Alves, R.V., and B'uchanan, R.A.: Properties of
Intensifying screens are used in diag­ Y,O,S:Tb x-ray intensifying screens. Presented at
the IEEE meeting in Miami, December, 1972.
nostic radiology because they reduce x-ray
2. Balter, S., and Laughlin, J.S.: A photographic I
dose to the patient. Also, the lower expo­ method for the evaluatwn of the conversion ef­
sures required (less mAs) allow the shorter ficiency of fluoroscopic screens. Radiology,
exposure times needed to reduce motion 86:145, 1966.
3. Bates,LM.: Properties of radiographic films and
unsharpness. Until about 1971, calcium mtens1fymg screens which influence the efficacy
tungstate was the phosphor used in most of a film-screen combmatwn. Contained in the
screens. New technology has resulted in syllabus of the joint BRH-ACR Conference: First
Image Receptor Conference: Film/Screen Com­
very fast rare earth and other phosphors. binations, Arlington, V irginia, November 13-15
There are three ways in which screens can 1975. Washington, DC, U.S. Government Print�
be made "faster": ing Office, HEW Pub!. No. (FDA) 77-8003, 1975,
p. 123.
4. Buchanan, R.A., Finkelstein, S.l., and Wicker­
1. thicker phosphor layer sheim, K.A.: X-ray exposure reduction using

rare-earth oxysulfide intensifying screens. Radi­ scopic and intensifying screens. In Medical Phys­
ology, 105:85, 1972. ics. Vol. 2. Edited by 0. Glasser. Chicago, Year
5. Castle, J.W.: Sensitivity of radiographic screens Book Medical Publishers, 1950.
to scattered radiation and its relationship to 15. Patterson, C.V.S.: Roentgenography: Fluoro­
image contrast. Radiology, 122:805, 1977. scopic and intensifying screens. In Medical Phys­
6. Coltman,J.W., Ebbighausen,E.G., and Alter, W.: ics. Vol. 3. Edited by 0. Glasser. Chicago, Year
Physical properties of calcium tungstate x-ray Book Medical Publishers, 1960.
screens.]. Appl. Physics, 18:530, 1947. 16. Patterson X-Ray Screens: Sales Manual for Du
7. Holland, R.S.: Image analysis. Contained in the Pont Patterson X-Ray Screen Dealers. W ilming­
syllabus of the AAPM 1975 Summer School, The ton, Photo Products Department, E.I. du Pont de
Expanding Role of the Diagnostic Radiologic Nemours and Company, Inc., 1952.
Physicist, Rice University, Houston, July 17. Properzio, W.S., and Trout, E.D.: The deterio­
27-August 1, 1975, p. 103. ration of x-ray fluoroscopic screens. Radiology,
8. Ishida, M., Kato, H., Doi, K., Frank, P.: Devel­ 91 :439, 1968.
opment of a new digital radiographic image proc­ 18. Roth, B.: X-ray intensifying screens. Contained
essing system. SPIE Vol. 347: Application of Op­ in the syllabus of the AAPM 1975 Summer
tical Instrumentation in Medicine X, 1982, p. 42. School, The Expanding Role of the Diagnostic
9. Lawrence, D.J.: Kodak X-Omatic and Lanex Radiologic Physicist, Rice University, Houston,
screens and Kodak films for medical radiography. July 27-August 1, 1975, p. 120.
Rochester, N.Y., Radiography Markets Division, 19. Sonoda, M., Takano, M., Miyahara,]., Kato, H.:
Eastman Kodak Company File No. 5.03, June Computed radiography utilizing scanning laser
1976. stimulated luminescence. Radiology, I 48:833,
10. Ludwig, G.W., and Prener, J.S.: Evaluation of 1983.
20. Philips Medical Systems, Inc.,710 Bridgeport Av­
Gd,O,S:Tb as a phosphor for the input screen of
enue, Shelton, CT, 06484. PCR Technical Review,
x-ray image intensifier. IEEE Trans. Biomed.
Issue No. I.
Eng., 19:3, 1972.
21. Stevels, A.L.N.: New phosphors for x-ray screens.
11. Meredith, W.J., and Massey, J.B.: Fundamental
Medicamundi 20:12, 1975.
Physics of Radiology. Baltimore, W illiams & Wil­
22. Ter-Pogossian, M.: The efficiency of radio­
graphic intensifying screens. In Technological
12. Meritt, C.: Computed radiography: a new ap­ Needs for Reduction of Patient Dosage from Di­
proach to plain film imaging. Diagnostic Imaging. agnostic Radiology. Edited by M.L. Janower.
January, 1985, p. 58. Springfield, IL, Charles C Thomas, 1963.
13. Morgan, R.H.: Characteristics of x-ray films and 23. Thompson, T.T., Radford, E.L., and Kirby, C.C.:
screens. Radiology, 49:90, 1947. A look at rare-earth and high-speed intensifying
14. Patterson, C.V.S.: Roentgenography: Fluoro- screens. Appl. Radio!., 6:71, 1977.

10 Physical Characteristics
of X-Ray Film and Film

When an x-ray beam reaches the patient, tion-sensttive, emulsion that is usually
it contains no useful medical information. coated on both sides of a transparent sheet
After the beam passes through and inter­ of plastic, called the base. Firm attachment
acts with the tissues in the part examined, between the emulsion layer and the film
it contains all the information that can be base i'S achieved by use of a thin layer of
revealed by that particular radiographic adhesive. The delicate emulsion is pro­
examination. This is represented by vari­ tected from mechanical damage by layers
ation in the number of x-ray photons in known as the supercoating (Fig. 10-l).
different areas of the emergent beam. We
are unable to make direct use of the infor­ Film Base
mation in this form, however, and must
The only function of the film base is to
transfer it to a medium suitable for viewing
provide a support for the fragile photo­
by the eye. The method of transfer might
graphic emulsion. Three characteristics of
involve a magnetic tape or disc, a fluoro­
the base must be considered. First, it must
scopic screen, or xerography. The most im­
not produce a visible pattern or absorb too
portant material used to "decode" the in­
much light when the radiograph is viewed.
formation carried by the attenuated x-ray
Second, the flexibility, thickness, and
beam is photographic film. The film may
strength of the base must allow for ease of
be exposed by the direct action of x rays.
More commonly, the energy of the x-ray processing (developing) and produce a ra­
beam is converted into light by intensifying diograph that "feels right" when handled
screens, and this light is used to expose the (a film too "floppy" to "snap" under the
film. hangers of a viewbox gets a cool reception).

. .
It is unfortunate that the transfer of in­ Third, the base must have dimensional sta-
formation from the x-ray beam to the

- �j
SUPERCOAT 0005 in.
screen-film combination always results in a

loss of information. In reviewing x-ray "---- .· . . · . · .· . · · · · . · .

.. .· > .
··. . ·.· · · ·. ·
· .. J
film, we must examine both the film and

those factors that influence the amount of O'i "
information lost in the transfer process. BASE-

EMULSION - · · · ... . .. .

. ..
. ,.
·· . .

X-ray film is photographic film consist­ Figure 10-1 Cross section of a double emul­
ing of a photographically active, or radia- sion x-ray film


bility; that is, the shape and size of the base cellulose triacetate base was developed. In
must not change during the developing 1960 the first medical radiographic film us­
process or during the stored life of the film. ing a polyester base was introduced. Poly­
Figure 10-2 illustrates a radiograph in ester as a film base offers the advantage of
which the base has slowly diminished in size improved dimensional stability, even when
over a period of 22 years. Notice how stored under conditions of varying humid­
shrinking of the base has thrown the un­ ity, and it is much stronger than acetate.
shrinking emulsion into folds, producing a What is polyester? It is similar to Dacron*
wrinkled appearance. polyester fiber used in clothing. The raw
Original x-ray "plates" consisted of a materials, dimethyl terephthalate (DMT)
glass plate with the emulsion coated on one and ethylene glycol, are brought together
side. It is no longer possible to make a "flat under conditions of low pressure and high
plate of the abdomen," because x-ray plates temperature to form a molten polymer that
are not available. The onset of World War is then literally stretched into sheets of ap­
I cut off the supply of photographic glass propriate size and thickness to form film
from Belgium and created a demand for a base. Cronex* film is an example of a poly­
less fragile x-ray film for use by the Army. ester-base film, and the characteristics of
Cellulose nitrate, previously used as a base polyester may be appreciated by examining
for photographic film, was adapted for use a piece of "cleared" Cronex film.
in x-ray film in 1914. Cellulose nitrate is Triacetate and polyester bases are clear
quite flammable, however, and several fires and colorless. In 1933 the first commer­
were caused by improper handling and cialized blue tint was added to x-ray film in
storage of the film. Because of this fire haz­ an effort to produce a film that was "easier"
ard a new "safety base" film was urgently to look at, causing less eyestrain. Present
sought and, in 1924, a "safety" film with a x-ray film is tinted blue. The blue dye may
be added to the film base or to the emul­
Triacetate base was about 0.008 in. (8
mils) thick, and polyester base is 7 mils
thick. The slightly thinner polyester base
has handling properties approximately
equal to those of the thicker acetate.
Photographic emulsion would not ad­
here to the finished base if applied directly.
Therefore, a thin layer of adhesive sub­
stance is applied to the base to ensure per­
fect union between base and emulsion.

The two most important ingredients of
a photographic emulsion are gelatin and
silver halide. The exact composition of the
various emulsions is a closely guarded in­
dustrial secret. Most x-ray film is made for
use with intensifying screens, and has
emulsion coated on both sides of the base.
Emulsion thickness varies with film type,
Figure 10-2 Wrinkled emulsion resulting *Dacron and Cronex are trademarks of E. I. du Pont
from shrinking of the film base de Nemours and Company, Inc.

but is usually no thicker than 0.5 mil. A

thicker emulsion would not be useful be­
cause of the inability of light to penetrate
to the deeper layers.
Gelatin. Photographic gelatin for x-ray
film is made from bone, mostly cattle bone
from India and Argentina. (Useless infor­
mation: India and Argentina do not have
better cows, but they have a lot of them.
They also have inexpensive labor to proc­
ess the bones.) Gelatin satisfies several ex­
acting requirements better than any other
suspension medium. It keeps the silver hal­
ide grains well dispersed and prevents the
clumping of grains. Processing (developing
and fixing) solutions can penetrate gelatin
rapidly without destroying its strength or
permanence, and gelatin is available in a
reasonably large quantity and uniform
Silver Halide. Silver halide is the light­
sensitive material in the emulsion. The hal­
ide in medical x-ray film is about 90 to 99%
silver bromide and about 1 to 10% silver
iodide (the presence of Agi produces an
emulsion of much higher sensitivity than a
pure AgBr emulsion). The silver iodo­ Figure 10-3 The silver iodobromide crystal
bromide crystals are precipitated and
emulsified in the gelatin under exacting
conditions of concentration and tempera­
ture, as well as the sequence and the rate emulsions. Crystal size might average 1.0
at which these chemicals are added. The to 1.5 microns (1 micron = 0.001 milli­
method of precipitation determines crystal meter) in diameter with about 6.3 x 109
size, structural perfection, and concentra­ grains per cubic centimeter of emulsion,
tion of iodine. In general, the precipitation and each grain contains an average of
reaction involves the addition of silver ni­ 1,000,000 to 10,000,000 silver ions.
trate to a soluble halide to form the slightly The silver iodobromide grain is not a
soluble silver halide: perfect crystal because a perfect crystal has
almost no photographic sensitivity. There
AgN03 + KBr ----> AgBr + KN03
are several types of crystal defects. A point
The silver halide in a photographic defect consists of a silver ion that has
emulsion is in the form of small crystals moved out of its normal position in the
suspended in the gelatin. The crystal is crystal lattice; these interstitial silver ions
formed from ions of silver (Ag+), ions of may move in the crystal (Fig. 10-4). A dis­
bromine (Br-), and ions of iodine (I-) ar­ location is a line imperfection in the crystal,
ranged in a cubic lattice (Fig. 10-3). These and may be thought of as a brick wall that
grains, or crystals, in a medical x-ray film contains one row in which the bricks are
emulsion are small but still relatively large not the same size as all the other bricks,
compared to fine-grain photographic thus causing a strain in the wall structure.

+ - + -
Ag -Br-Ag-Br
A site of crystal imperfection, such as a dis­
location defect, or an AgS sensitivity speck,
may act as an electron trap where the elec­
tron is captured and temporarily fixed.
The electron gives the sensitivity speck a
negative charge, and this attracts the mo­
bile interstitial Ag+ ions in the crystal. At
the speck, the silver ion is neutralized by
the electron to form a single silver atom:
+ _, +
Ag - I _,_Ag -Br Ag + + electron� Ag

+ This single atom of silver then acts as an

' .

- +
.... l+ ,_ electron trap for a second electron. The
negative charge causes a second silver ion
to migrate to the trap to form a two-atom
silver nucleus. Growth of silver atoms at the
Figure 10-4 A point defect site of the original sensitivity speck con­
tinues by repeated trapping of electrons,
This may be the way in which the iodine followed by their neutralization with inter­
ion strains the crystal. stitial silver ions. The negative bromine
Chemical sensitization of a crystal takes ions that have lost electrons are converted
several forms. It is commonly produced by into neutral bromine atoms, which leave
adding a sulfur-containing compound, the crystal and are taken up by the gelatin

such as allylthiourea, to the emulsion, of the emulsion. Figure 10-5 diagrams the

which reacts with silver halide to form sil­ development of a two-atom latent image

ver sulfide. The silver sulfide is usually lo­ according to the Gurney-Mott hypothesis.6
cated on the surface of the crystal and is A single silver halide crystal may have
referred to as the sensitivity speck. It is one or many of these centers in which

the sensitivity speck that traps electrons to atomic silver atoms are concentrated. The
begin formation of the latent image cen­ presence of atomic silver is a direct result
ters. of the response of the grain to light ex­
posure, but no visible change has occurred
Latent Image in the grain. These small clumps of silver
can, however, be seen with electron mi­
Metallic silver is black. It is silver that
croscopy. These clumps of silver atoms are
produces the dark areas seen on a devel­
termed latent image centers, and are the
oped radiograph. We must explain how ex­
sites at which the developing process will
posure of the sensitized silver iodobromide
cause visible amounts of metallic silver to
grains in the film emulsion to light (from
be deposited. The difference between an
x-ray intensifying screens), or to the direct
emulsion grain that will react with the de­
action of x rays, initiates the formation of
veloping solution and thus become a visible
atomic silver to form a pattern. The energy
silver deposit and a grain that will not be
absorbed from a light photon gives an elec­
"developed" is the presence of one or more
tron in the bromine ion enough energy to
latent image centers in the exposed grain.
escape. The electron can move in the crys­
At least two atoms of silver must be present
tal for relatively large distances as long as
at a latent image center to make a grain
it does not encounter a region of impurity
developable (i.e., to become a visible de­
or fault in the crystal.
posit of silver). In practical terms the min­
Br- + light photon � Br + electron imum number to produce developability is



@ �


f of the


� /



Figure 10-5 Formation of the latent image

probably between three and six. The more photoelectric absorption or Compton scat­
silver atoms that exist at a latent image cen­ tering, and have rather long ranges in the
ter, the greater the probability that the emulsion. In fact, an electron produced in
grain will be developed. Some centers will this way may react with many grains in an
contain several hundred silver atoms. Un­ emulsion. The manner in which the energy
der the usual conditions, the absorption of of the electrons is imparted to the photo­
one quantum of light by a silver halide graphic emulsion is complex and will not
grain will produce one atom of silver and be considered in detail. The final result is
one of bromine. freeing of electrons from the bromide ion,
producing bromine atoms and an electron
Direct X-Ray Exposure that can move to a trapping site and begin
The photographic effect of direct ab­ the process of latent image formation. The
sorption of x rays by the emulsion is not energy of one absorbed x-ray photon can
caused by electromagnetic radiation itself produce thousands of silver atoms at latent
but by electrons emitted when the x-ray image sites in one or several grains. Even
photon interacts with the silver halide in this large number of silver atoms is low,
the emulsion. These electrons come from considering the energy of the absorbed

photon. Most of the energy of the absorbed veloped film may be used as an indication
photon is lost in processes that do not pro­ of how much x-ray exposure (i.e., how
duce any photographic effect (such as many milliroentgens) the film has received.
losses to gelatin). Only 3 to 10% of the pho­ Because the sensitivity of the film varies
ton energy is used to produce photolytic greatly with the energy (kVp) of the x rays,
silver. The photographic effect of direct however, blackening of a piece of film does
x-ray exposure on an emulsion can be in­ not give an accurate estimation of the ex­
creased by a factor of almost 100 by proper posure to which the film has been sub­
chemical sensitization of the emulsion. jected. For example, a film subjected to an
The sensitivity of film to direct x-ray ex­ exposure of 50 mR at an x-ray energy of
posure varies significantly (by a factor of 50 kVp will, after development, exhibit a
20 to 50) with the energy (kV p) of the x-ray much higher density (amount of blacken­
beam. This x-ray spectral sensitivity is most ing) than an identical film subjected to an
important when considering use of film to exposure of 50 mR by 200-kVp x rays. The
measure x-ray exposure dose (i.e., film problem of variation of film sensitivity with
badge monitoring). Above 50 kVp, the ef­ radiation energy is partially solved by plac­
ficiency with which absorbed x-ray photons ing various metal filters in front of the film
are utilized to produce a photographic ef­ in an attempt to control the energy (kVp)
fect decreases significantly with increasing of the x rays that reach different areas of
photon energy. At about 50 kVp the aver­ the film. The accuracy of film badge mon­
age keV of the x rays produced will be close itoring of x-ray exposure is about ± 20%.
to the K-shell binding energy of silver (25.5 Film badge monitoring of personnel ex­
keV) and bromine (13.5 keV). This will posure offers several ad vantages over
cause the film to exhibit maximum pho­ other methods, such as ionization cham­
toelectric absorption of 50-kVp x rays. Fig­ bers. The film badge provides a permanent
ure 10-6 shows, in a rough graphic form, record, and is small in size and weight, rug­
the way in which the x-ray sensitivity of film ged, and inexpensive.
varies with kV p.
The sensitivity also varies greatly with
the way in which the film is developed. The Covering the emulsion is a thin layer,
amount of blackening (density) on the de- commonly gelatin, that serves to protect
the emulsion from mechanical damage. In
special types of film this supercoat, or anti­

1- abrasive coating, may contain substances
> that make the film surface smooth and
1- slick. This is a desirable quality in film that
z must be transported through a cut film
CJ) 60 rapid film changer.

1L.. Development
w 30
> Development is a chemical process that

1- amplifies the latent image by a factor of
....J millions (about 100,000,000) to form a vis­
a:: ible silver pattern. The basic reaction is re­
20 60 100 140 180 200 duction (addition of an electron) of the sil­
kVp ver ion, which changes it into black metallic
Figure 10-6 Film sensitivity varies with radi­
ation quality Ag+ + electron__,. Ag

The developer is the reducing agent. De­ and inactivation of the developing agent
velopment is generally an all-or-none phe­ and the liberation of hydrogen ions. Note
nomenon, because an entire grain is de­ that the reaction must proceed in an al­
veloped (reduced) once the process begins. kaline solution. When hydroquinone is ox­
The process is usually initiated at the site idized to quinone, two electrons are liber­
of a latent image speck (commonly on the ated to combine with the two silver ions to
surface of the grain). It is believed that the form metallic silver (Fig. 10-7A). The re­
action of the silver atoms in the latent action of phenidone is similar (Fig. 10-7B).
image is to accelerate (catalyze) the reduc­ The silver thus formed is deposited at
tion of the silver ions in the grain by the the latent image site, gradually enlarging
developing chemicals. The silver in a grain this initially microscopic black spot into a
that does not contain a latent image can be single visible black speck of silver in the
reduced by the developer, but at a much emulsion.
slower rate. Thus, time is a fundamental In addition to developing agents, the de­
factor in the developing process. Devel­ veloping solution contains ( 1) an alkali to
opment should be discontinued when the adjust the pH, (2) a preservative (sodium
differential between exposed developed sulfite), and (3) restrainers, or antifog­
grains and unexposed undeveloped grains gants. The alkali adjusts the hydrogen ion
is at a maximum. concentration (pH), which greatly affects
Modern developing solutions contain the developing power of the developing
two developing agents, hydroquinone plus agents, especially hydroquinone. In addi­
phenidone or metol. H ydroquinone was tion, the alkali serves as a buffer to control
discovered to be a developing agent in the hydrogen ions liberated during the de­
1880. Hydroquinone requires a strong al­ velopment reaction. Most radiographic de­
kali to activate it. Developers made of hy­ velopers function at a pH range of 10 to
droquinone are characterized by high con­ 11.5. Typical alkalies include sodium hy­
trast. Metol developers became available in droxide, sodium carbonate, and borates
1891, and are characterized by high speed, (sodium metaborate and sodium tetrabor­
low contrast, and fine grain. Phenidone was ate).
discovered in 1940, and is similar to metol. Sodium sulfite is added for two reasons.
Both metol and phenidone are used mainly The oxidation products of the developing
in combination with hydroquinone. This agents decompose in alkaline solution and
statement is usually expressed the other form colored materials that can stain the
way around by stating that hydroquinone emulsion. These products react rapidly
is used mainly in combination with metol with sodium sulfite to form colorless solu­
or phenidone. Two agents are used be­ ble sulfonates. In addition, sodium sulfite
cause of the phenomenon of synergism, or acts as a preservative. In alkaline solution
superadditivity. The mixture results in a the developing agent will react with oxygen
development rate greater than the sum of from the air. The sulfite acts as a preser­
the developing rate of each developing vative by decreasing the rate of oxidation,
agent. The reasons for development syn­ especially that of hydroquinone. Sulfite re­
ergism are complex and not fully under­ moves oxygen from the air dissolved in the
stood, so we will not explore the details. solution, or at the surface of the solution,.
The chemistry of developing is not our before it has time to oxidize the developing
chief interest, but the formulas for the basic agent.
reactions help in gaining a good under­ Fog is the development of unexposed
standing of the process . As shown in Figure silver halide grains that do not contain a
10-7, the developing agent reduces silver latent image. In a complex manner, dilute
ions to metallic silver, causing oxidation concentrations of soluble bromide (potas-

OH 0

¢1 Alkaline
+2Ag- ----+ ¢ +2Ag+2H
:-... Solution

OH 0

Developer Silver Oxidized Silver Hydrogen

(Hydroquinone) Ions ----+ Developer + Atoms + Ions

+2 Ag- --=-s--
o lu-,t i-'
+ 2 Ag + 2H.


Developer Silver Oxidized + Silver + Hydrogen

(Phenidone) + Ions -------+ Developer Atoms Ions

Figure 10-7 Basic chemical reactions involved in the development process. Hydroquinone (A)
or phenidone (B) may be used as the developer

sium bromide) decrease the rate of fog for­ ference in commercial x-ray developing so­
mation. To a lesser degree the bromide also lutions is in the antifoggants present.
decreases the rate of development of the Developer formulas also contain other
latent image. Soluble bromide produced as ingredients designed to influence swelling
a byproduct of the development process of the x-ray film emulsion, development
also affects the activity of the developer. rate, and physical properties. All devel­
The development reaction in a 90-sec x-ray opers contain the same basic functional
processor must be completed in about 20 components: developer (reducing agent),
sec. This rapid rate of development re­ alkali, preservatives, and bromide. Differ­
quires that the temperature of the devel­ ences in antifoggants and other ingredients
oping solution be quite high, usually be­ are often proprietary, so we cannot give
tween 90 and 95° F. This rapid, high­ specific examples.
temperature rate of developing requires The bromide ions released by the re­
that modern x-ray developers contain ad­ duction of silver ions to silver atoms pass
ditional antifoggants to permit rapid de­ into the developing solution. It is mainly
velopment of exposed grains but minimize this increase in bromide concentration that
fog development. The most significant dif- limits the life of developing solutions.

Replenishment cessor is about 2 to 3 months. A typical

We have seen that, during use, devel­ replenishment rate is to replace 60 ml of

oping solutions consume developing developer with replenisher for each 14- X

agents and preservatives, but acquire hy­ 17-in. film processed.

drogen ions and bromide. Each time a film Now, let us consider a tank of developer

is processed in an automatic processor a that sits for long periods of time with few

small portion (about 60 ml in a 10-L tank) films being processed. In this situation, ox­

of the developing solution is removed and idation of the developer becomes more im­

replaced with a replenishment solution. portant than the development reaction.

The purpose of this replenishment is to The oxidation reaction may be written:

maintain developing agent concentration,

H20 + Na2S03 + 02
preservative concentration, bromide con­ (hydroquinone + sodium sulfite + oxygen)
centration, and pH at a constant level for I
the lifetime of the developer solution. Al­ Oxidation

most all radiographs are now processed in I

automatic film processors with automatic HQS03Na + NaOH + Na2S04
developer replenishment. Traditional re­ (hydroquinonemonosulfonate +

plenishment was developed for high vol­ sodium hydroxide + sodium sulfate)

ume operations where many films are proc­

essed each day. However, many automatic Notice that the oxidation reaction raises

processors operate in small installations the pH of the developer by forming so­

where few films are developed each day. dium hydroxide. This is just the opposite

Under these circumstances, oxidation of of the development reaction in which the

developer is more important than the con­ acid formed lowers developer pH. Also,

sequences of the development process. We the oxidation reaction produces no bro­

will discuss replenishment requirements mide. Since few films are processed, re­
for both high volume (developing reac­ plenishment is infrequent. Also, since

tion dependent) and low volume (oxida­ standard replenisher has a higher pH than

tion reaction dependent) situations. developer and no bromide, routine replen­

The development reaction may be writ­ ishment will maintain the high pH and di­

ten: lute bromide. Bromide concentration

drops rapidly, and this has an adverse ef­
2 AgBr + H20 + Na2SOa
fect on film sensitometry.
(silver bromide + hydroquinone + sodium sulfite)
Developer and replenisher formulas are
Development modified for low volume applications. De­
� velopers have a lower pH and higher sulfite
2 Ag + HBr + HQS03Na + NaBr concentration to retard oxidation, and a
(Silver + hydrobromic acid +
high buffering capacity to minimize the pH
hydroquinonemonosulfonate + sodium bromide)
effects of oxidation. Replenishers compen­
Notice that each time a film is processed, sate mostly for oxidation rather than de­
bromide and acid are formed and some velopment. The replenisher has a lower
developer is consumed. Replenishment of pH than developer and contains bromide.
developer must compensate for these Replenishment rate is usually higher
changes by being free of bromide, by con­ (about 90 ml per 14- X 17-in. film) to in­
taining alkaline agents and buffers and, to crease developer turnover rate.
a lesser extent, restoring depleted preser­ The main consequence of an abnormal
vative and developing agents. With normal composition of developer ingredients is to
use in a busy department, the lifetime of a produce films with a sharp decrease in toe
tank of developer in an automatic film pro- gradient (we will discuss the terms "toe"

and "gradient" in Chapter 14). Such a common fixing agent, or "hypo." Why is it
change makes it hard to detect low contrast called hypo? In earlier chemical nomencla-·
images in the lighter (lower density) areas ture, the compound we call sodium thio­
of a radiograph. sulfate (Na2S203) was given the name hy­
In summary, the chemical composition posulfite of soda, and "hypo" it remains to
of the developer and replenisher, and the photographers. At least three silver thio­
replenishment rate, must provide a con­ sulfate complexes are formed in the fixing
stant composition of solution in the devel­ solution; their identities need not concern
oper tank of an automatic processor. The us. A typical reaction might be
most important parameters are pH and Silver bromide + sodium thiosulfate ---+

bromide concentration. Restoration of con­

sumed preservatives and developing silver thiosulfate complex + sodium bromide

agents is also required. The ammonium thiosulfate salt is more ac­

tive, and is used in fixer supplied in the
Fixing form of a liquid concentrate.
Only part of the silver halide in the emul­ In addition to thiosulfate, the fixing so­
sion is reduced to silver during developing. lution contains a substance to harden the
The remaining silver halide impairs both gelatin. Hardening results in a decrease in
the immediate usefulness and permanence the swelling of gelatin, making it tougher
of the developed radiograph. Therefore it and more resistant to abrasion. The hard­
must be removed, but the fixing solution ener is usually a chromium or aluminum
must remove silver halide without dam­ compound. The fixing bath also contains
aging the image formed by metallic silver. an acid, stabilizers, and a buffer to maintain
The solubility of silver halide (we will use the acidic pH level.
silver bromide as an example) in a water An incompletely fixed film is easily rec­
solution is controlled by the concentration ognized because it has a "milky" or cloudy
of silver and halide ions. Silver bromide is appearance. This is a result of the disper­
only slightly soluble in water. The product sion of transmitted light by the very small
of the silver and bromide ions in solution silver iodobromide crystals that have not
is always constant for any given tempera­ been dissolved from the emulsion.
ture, and may be expressed by the equation
Silver ion x bromide ion constant
After developing and fixing, the film

If the concentration of silver ions could be must be well washed with water. Washing
reduced, the concentration of bromide serves primarily to remove the fixing-bath
ions would have to increase, which means chemicals. Everyone has seen an x-ray film
that more silver bromide would have to dis­ that has turned brown with age. This is the
solve from the emulsion. Thus, the solu­ result of incomplete washing. Retained
bility of silver halide would increase. The hypo will react with the silver image to
function of the fixing agent is to form wa­ form brown silver sulfide,just as silverware
ter-soluble complexes in which silver ions acquires a brown tarnish when exposed to
are tightly bound. The soluble complex the hydrogen sulfide produced by cooking
thus formed effectively removes silver ions gas. The general reaction is
from the solution. Hypo + silver---+ Silver sulfide (brown)
Two agents form satisfactory stable com­ + Sodium sulfite

plexes with silver ions, cyanides and thio­

sulfates. Cyanides are poisonous and are SUMMARY
not generally used. Thiosulfate in the form X-ray film is a photographic film coated
of the sodium or ammonium salt is the with emulsion on both sides of the film

base. The light-sensitive material in the 2. Fuchs, A.W.: Evolution of roentgen film. Am. J.
Roentgenol., 75:30, 1956.
emulsion is a silver iodobromide crystal. X­ 3. James, T.H., and Higgins, G.C.: Fundamentals
ray film is only slightly sensitive to direct of Photographic Theory. 2nd Ed. New York,
Morgan and Morgan, 1968.
x-ray exposure. Impurities in the silver
4. Martin, F.C., and Fuchs, A.W.: The historical ev­
halide crystal structure increase the light olution of roentgen-ray plates and films. Am. J.
sensitivity of the film emulsion. Light, or Roentgenol., 26:540, 1931.
5. Mees, C.E.K., and James, T.H.: The Theory of
x-ray, exposure causes the grains in the
the Photographic Process. 3rd Ed. New York,
emulsion to develop an invisible latent Macmillan, 1969.
image. The developing process magnifies 6. Neblette, C.B.: Photography , Its Materials and
Processes. 6th Ed. New York, Van Nostrand,
the latent image to produce a visible pat­ 1962.
tern of black metallic silver. 7. Wayry nen, R.E., Holland, R.S., and Trinkle, R.J.:
Chemical Manufacturing Considerations and
The sensitivity of x-ray film to direct
Constraints in Manufacturing Film, Chemicals
x-ray exposure varies with the kVp of the and Processing. Proceedings of the Second Image
x-ray beam. Receptor Conference: Radiographic Film Proc­
essing, Washington, DC, March 31-April 2, 1977,
pp. 89-96. Washington, DC, U.S. Government
REFERENCES Printing Office, 1977, Stock No. 017-015-00134-
l. Baines, H., and Bomback, E.S.: The Science of 8. Wuelfing, P.: High stability developer for medical
Photography. 2nd Ed. London, Fountain Press, x-ray processing. SPIE Vol. 555 Medical Imaging
1967. and Instrumentation, 1985, p. 91.

11 Photographic
of X-Ray Film

The diagnostic accuracy of a radio­ This chapter will discuss the response of
graphic film examination depends, in part, the x-ray film to exposure.
on the visibility of diagnostically important
information on the film. Understanding
the relationship between the exposure a PHOTOGRAPHIC DENSITY
film receives and the way the film responds
to the exposure is essential to intelligent When the x-ray beam passes through
selection of proper exposure factors and body tissues, variable fractions of the beam
type of film to provide maximum infor­ will be absorbed, depending on the com­
mation content of the radiograph. position and thickness of the tissues and on
What is meant by the term "exposure" the quality (kVp) of the beam. The mag­
of an x-ray film or film-screen combina­ nitude of this variation in beam intensity is
tion? Exposure is proportional to the prod­ the mechanism by which the x-ray beam
uct of the milliamperes of x-ray tube cur­ acquires the information transmitted to the
rent and the exposure time. Thus, an film. This pattern of varying x-ray intensity
exposure of 100 milliamperes for 1 second has been called the x-ray image. Webster's
is expressed as 100 milliampere-seconds, Collegiate Dictionary defines an image as
usually written 100 mAs. An exposure of "a mental representation of anything not
100 mAs could also be produced by using actually present to the senses." This defi­
50-mA tube current for 2 sec, 200 rnA for nition is particularly applicable to the idea
0.5 sec, 500 rnA for 0.2 sec, and so forth. of the x-ray image. The x-ray image ac­
In this chapter, film exposure is assumed tually exists in space, but it cannot be seen
to mean exposure of the x-ray film by light or otherwise detected by the natural senses.
from x-ray intensifying screens, unless oth­
The x-ray image is the pattern of infor­
erwise stated.
mation that the x-ray beam acquires as it
Exposure (mAs) of the x-ray film pro­
passes through and interacts with the pa­
duces film blackening, or density. The qual­
tient; that is, the beam is attenuated by the
ity of the x-ray beam (kVp) has more effect
patient. The x-ray image is present in the
on image contrast. Two general but not
space between the patient and the x-ray
completely accurate statements should be
film (or x-ray intensifying screen). The in­
kept in mind:
formation content of the x-ray image must
1. mAs controls film density be transformed into a visible image on the
2. kVp controls image contrast x-ray film with as little loss of information


as possible. We have seen how the energy called transmittance. Useful densities in di­
of the x-ray beam may be used to produce agnostic radiology range from about 0.3
a visible pattern of black metallic silver on (50% of light transmitted) to about 2 (1%
the x-ray film; the degree of film black­ of light transmitted). A density of 2 means
ening is directly related to the intensity of Io
that log- = 2. Because the log of 100 =
radiation reaching the film or intensifying I,
screen. The measurement of film blackness
is called "photographic density"; usually, 2, !c! = 100. Thus, for every 100 light pho-
only the word density is used. Density is tons incident on the film, only 1 photon,
expressed as a number that is actually a
or 1%, will be transmitted. Table11-1
logarithm, using the common base 10.
Photographic density is defined by
li ��s some (
commo � values for opacity

D =
(I,) .
, dens1ty log I, ) , and percentage of

light transmission. Note that an increase in
D = density film density of 0.3 decreases transmitted
10 = light incident on a film light to 50% of its previous value. For ex­
I, = light transmitted by the film ample, an increase in density from 0.6 to

Refer to Figure 11-1. If ten arrows (pho­ 0.9 decreases the amount of transmitted
tons) of light strike the back of the film, light from 25 to 12.5%. This emphasizes

and only one photon passes through the the fact that the number used to signify a

film, then I0 = 10 and I, = 1: certain density has no units, but is a loga­

rithm. The number 0.3 is the logarithm of
Dens1ty . =
log 1 = Iog 10 = 1 2. Thus, an increase in density of 0.3

( f) log means an increase in opacity

Note that !c! measures the opacity of the
film (the ability of film to stop light). The
(f) of 2; opacity is doubled by a density

reciprocal of density, ..!:, measures the frac­ increase of 0.3.

Io Higher density means a blacker film (less
tion of light transmitted by the film, and is light transmission). In routine x-ray work,
a density of 2 (1% of light transmitted) is

llll ll
black when viewed on a standard viewbox,
and a density of 0.25 to 0.3 (50% of light
transmitted) is very light.

Table 11-1. Percentage of Light Transmitted

by X-Ray Films of Various Densities


1 0 100
2 0. 3 50
4 0.6 25
8 0.9 12.5
10 10 1.0 10
LOG101 30 1. 5 3.2
100 2 1
t 1,000 3 0.1
10,000 4 0.01
Figure 11-1 Photographic density

If an unexposed x-ray film is processed, transmitted light photons between density

it will demonstrate a density of about 0.12. 0.3 and density 0.6 is 250 (500 to 250), but
This density consists of base density and between density 0.6 and density 0.9 it is
fog. The plastic material used to make the only 125 photons (250 to 125). The eye will
film base absorbs a small amount of light. interpret density 0.3 as being exactly as
Also, the blue dye used to color some film much brighter than density 0.6 as density
bases adds slightly to base density. Total 0.6 is brighter than density 0.9. The eye
base density will average about 0.07. A few has "seen" the equal differences in density
of the silver halide grains in an x-ray film rather than the unequal differences in the
emulsion develop without exposure. These number of light photons transmitted.
unexposed but developed grains compose The third reason for expressing density
the density known as fog. Fog density of a as a logarithm deals with the addition or
fresh x-ray film averages about 0.05. We superimposition of densities. If films are
will refer to the subject of base and fog superimposed, the resulting density is
density several times in this chapter and in equal to the sum of the density of each film.
Chapter 14. Consider two films, one of density 2 and
Why is density expressed as a logarithm? one of density 1, which are superimposed
There are three primary reasons. First, log­ and put in the path of a light source with
arithms conveniently express large differ­ an intensity of 1000 units (Fig. 11-3). The
ences in numbers on a small scale. For film of density 1 absorbs 90% of the light
example, the difference in the light trans­ (100 units are transmitted) and the film of
mission represented by going from a den­ density 2 absorbs 99% of these 100 units,
sity of 1 (10% of light transmitted) to a to allow final transmission of 1 unit of light.
density of 2 (1% of light transmitted) is a We started with 1000 units (10 = 1000) and
factor of 10. ended with 1 unit of light (I, = 1), so we
Second, the physiologic response of the may calculate density:
eye to differences in light intensities is log­
10 1000
arithmic (Fig. 11-2). Assume that a film D = log - = log -- = 3
I, 1
having regions of density that equal 0.3,
0.6, and 0.9 is transilluminated by a light
source of 1000 photons. The number of
light photons transmitted will be 500, 250,
and 125, respectively. The difference in




log _Q_ = DENSITY log 2 = 0.3 log 4 = 0.6 log 8 = 0.9 DENSITY 3 = +
Figure 11-2 The reduction in intensity caused Figure 11-3 The density of superimposed
by three films having densities of 0.3, 0.6, and films is the sum of the density of the individual
0.9 films

The effect of superimposing films with a

density of 2 and 1 is the same as using a
single film of density 3. Almost all the film
used in radiology has two emulsions, one
on each side of the base. The total density
exhibited by the radiograph is the sum of
the density of each emulsion.

It is necessary to understand the rela­ C/)
tionship between the exposure a film re­ lLI
ceives and the density produced by the
exposure. The relationship between ex­
posure and density is plotted as a curve,
known as the "characteristic curve" or "H
and D curve" (named after F. Hurter and
V.C. Driffield, who first published such a
curve in England in 1890). The concept of
the characteristic curve of an x-ray film ex­
posed by light from x-ray intensifying
screens is illustrated in Figure 11-4. Film
density is plotted on the vertical axis and Figure 11-4 Typical characteristic curve of a
screen-type x-ray film, exposed with x-ray in­
film exposure on the horizontal axis. The
tensifying screens
shape and location of this curve on the
graph are important and will, we hope,
take on some meaning as this chapter pro­ kVp and rnA, doubling the time of expo­
gresses. sure will double the mAs). The exposure
Characteristic curves are derived by giv­ is recorded as the relative exposure.
ing a film a series of exposures, developing The term relative exposure tends to cre­
the film, and plotting the resulting density ate confusion. Actually, the radiologist and
against the known exposure. The actual ex­ technologist think in terms of relative ex­
posure the film received may be measured posure when evaluating radiographs. For
in the laboratory, but such measurements example, if a radiograph of the abdomen
are not important to use and understand exposed with factors of70 kVp and75 mAs
the characteristic curve. (Using medium­ is judged to be underexposed, the correc­
speed intensifying screens and 80-kVp x tion might involve increasing the mAs to
rays, a density of 1.0 on the x-ray film re­ 150. In other words, the correction in­
quires that about 3 x 104 x-ray photons volves doubling the exposure. The actual
hit each square mm of the screen.) By film exposure (such as the number of milli­
exposure we refer to the product of the roentgens or the number of x-ray photons
intensity of the exposure (milliamperes of per square mm) is not known, and does not
x-ray tube current) and time of exposure have to be. The relationship between the
(expressed in seconds). Exposure is ex­ two exposures, however, is important. One
pressed in terms of milliampere-seconds, function of the characteristic curve is to
usually abbreviated mAs. One way to pro­ allow the amount of change necessary to
duce a characteristic curve is to expose dif­ correct an exposure error to be predicted.
ferent areas of a film with constant kilo­ For example, if a film with the character­
voltage and milliamperage while varying istic curve shown in Figure 11-5 is under­
the time of exposure (e.g., with constant exposed so that its average density is 0.35,

the corresponding relative exposure is 4. 3.5

If the exposure (mAs) is doubled (relative
exposure is increased to 8), it can be pre­
dicted that the average density will increase
to about 0.8. The exposure is also recorded
as the logarithm of the relative exposure, >-
mainly for two reasons. First, use of a log­
arithmic scale allows a very wide range of z
exposures to be expressed in a compact 0
graph. Second, use of log relative exposure ...1
makes analysis of the curve easier. Two ex­ �
posures whose ratio is constant (e.g., one 0
is twice the other) will always be separated
by the same distance on the exposure scale,
regardless of their absolute value. Refer to 0.5
Figure 11-5, in which both the relative ex­
posure and the log relative exposure are
indicated, and note that an increase in the 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7
log relative exposure of 0.3 always rep­ LOG RELATIVE EXPOSURE
resents a doubling of the relative expo­
sure. Figure 11-6 The regions of the characteristic
Analysis of the characteristic curve of a
particular x-ray film provides information
about the contrast, speed (sensitivity), and 11-6. Even at 0 exposure the film density
latitude of the film. Please refer to Figure is not 0 but will usually be 0.2, or less. This
density is made up of fog (development of
unexposed grains of silver halide in the
emulsion) and base densities (opacity of the
film base), which have been previously dis­
3.0 cussed. Therefore, total density on an ex­
posed and developed film will include base
2.5 and fog densities . The minimum density
caused by base and fog in a "fresh" film is
about 0.12. To evaluate density produced
>- 2.0
1- by the exposure alone, base and fog den­
C/) sities must be subtracted from the total
UJ 1.5 density. Second, note that at low density
(toe) and high density (shoulder), the film
shows little change in density despite a rel­
atively large change in log relative expo­
sure (Fig. 11-6). The important part of the
0.5 characteristic curve is between the toe and
shoulder, and in this region the curve is
2 4 8 16 32 64 128 256 512
almost a straight line. In this "straight line"
0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7
portion the density is approximately pro­
LOG RELATIVE EXPOSURE portional to the log relative exposure. For
exampie, if log relative exposure 1.1 pro­
Figure 11-5 The relationship between rela­
tive exposure and the corresponding log rela­ duces a density of 1.0, and log relative ex­
tive exposure posure 1.3 produces a density of 2.0, we

can predict that a density of about 1.5 will number of x rays in each part of the atten­
be produced by a log relative exposure of uated x-ray beam to be sufficient to pro­
1.2 (these figures correspond roughly to duce correct overall density in the proc­
the characteristic curve of Figure 11-6). essed film. Because exposure (mAs)
determines the total number of x rays in
Film Contrast the beam, mAs may be considered analo­
The information content of the invisible gous to light in producing an ordinary pho­
x-ray image is "decoded" by the x-ray film tograph. Too little, or too much, mAs re­
into a pattern of variations in optical den­ sults in an underexposed or overexposed
sity, known as "radiographic contrast." radiograph.
Radiographic contrast is the density dif­ Subject contrast may be thought of as
ference between image areas in the radi­ one factor that controls the log relative ex­
ograph. There are many definitions of con­ posure that reaches the film. That is, film
trast, but we will use the simple definition directly under a bone receives a low ex­
that contrast is the difference in density posure, whereas a high exposure reaches
existing between various regions on the the film under soft tissue areas of the sub­
film. Radiographic contrast depends on ject. A consideration of film contrast must
subject contrast and on film contrast. Sub­ examine how the film responds to the dif­
ject contrast depends on the differential ference in exposure produced by subject
attenuation of the x-ray beam as it passes contrast. Film contrast depends on four
through the patient. The discussion of at­ factors:
tenuation in Chapter 5 has already intro­
l. characteristic curve of the film
duced the important aspects of subject con­
2. film density
trast. Subject contrast was seen to be
3. screen or direct x-ray exposure
affected by the thickness, density, and
4. film processing
atomic differences of the subject, the ra­
diation energy (kVp), contrast material, Shape of the Characteristic Curve. The
and scatter radiation. The major theme of shape of the characteristic curve tells us
the remainder of this chapter will be film how much change in film density will occur
contrast. as film exposure changes. The slope, or
The information content of the x-ray gradient, of the curve may be measured
image is the pattern of varying intensity of and expressed numerically. One such
the x-ray beam caused by differential at­ measurement is called film gamma. The
tenuation of x rays by the subject. Few x gamma of a film is defined as the maximum
rays reach the film through areas of bone slope of the characteristic curve, and is de­
or opaque contrast material, while many scribed by the formula
photons are transmitted through soft tis­
02- D,
sue, and the air around the patient stops Gamma=_---=. ..:..._
_ .:..
log E2- log E,
almost no x-ray photons. The kVp must be
selected with care so that the numbers of where D2 and D1 are the densities on the
photons attenuated by bone and soft tissue steepest part of the curve resulting from
are in the proper proportion to produce log relative exposures E2 and E1• Figure
an x-ray image of high information content 11-7 shows an example of how film gamma
for the film intensifying screen to "de­ is calculated. The gamma of x-ray films ex­
code." The correct kVp is tremendously posed with intensifying screens ranges
important in producing proper subject from 2.0 to 3.5.
contrast. This relationship (kVp and con­ In radiology the concept of film gamma
trast) will be examined in detail in Chapter is of little value because, as illustrated in
14. Using the correct mAs causes the total Figure 11-7, the maximum slope (steepest)

3. 5

3. 0


o2- D 1 ?"?------------------
Gamma -,--
�-:-'--=­ >- Average Gradient •

log E2 -log E

t 2.0 !1.��-�:�-----------­
2.0 - 1.5
1 !-
2.0 2.2- 0.45
1.4 - 0.85
• .!..:I1
1.35 -1.2
z �
� 1 .5 !>_,_·-��--------- 0.5 Q
1.5 Average Gradient • 3.2

• o:i5
Gamma 3.33


0.5 Q �;i
_ � _______
Loa E,:
0.2 ----- 1.4

0.3 2..1 2.4 2.7 3.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7


Figure 11-7 The gamma of an x-ray film Figure 11-8 The average gradient of an x-ray

portion of the characteristic curve is usually

very short. We are interested in the slope the hand (exposed with a screen-film com­
of the curve over the entire range of useful bination) made with an x-ray beam of
radiographic densities (0.25 to 2.0). The proper energy (kVp) to cause the bones to
slope (gradient) of a straight line joining absorb four times as many photons as the
two points of specified density on the char­ soft tissue. This means that four times as
acteristic curve is called the average gra­ many x-ray photons will reach the film un­
dient. The average gradient is usually cal­ der soft tissues as will reach the film under
culated between density 0.25 and 2.0 above bone. The difference in log relative ex­
base and fog for radiographic films. Such posure between bone and soft tissue reach­
a calculation is shown in Figure 11-8, ing the film under the hand is 0.6 (log 4
which is the same curve from which gamma = 0.6). If we assign a value of 1.5 as the
was calculated in Figure 11-7. In calculat­ log relative exposure in the soft tissue area
ing average gradient, D1 is always 0.25 and (E5), then the log relative exposure corre­
D2 is 2.0; therefore, D2 - D1 is always 1.75. sponding to the bones (EB) is 0.9 (Fig.
In Figure 11-8, log E2 and E1 are 1.4 and 11-9). These two exposures will produce
0.85, respectively. film densities of 0.8 (bone density) and 2.8
If the average gradient of the film used (soft tissue density) on the hypothetical
is greater than 1, the film will exaggerate film's characteristic curve depicted in Fig­
subject contrast and, the higher the aver­ ure 11-9. This is a density range of 2.0,
age gradient, the greater this exaggeration corresponding to an overall brightness
will be. A film with average gradient of 1 range of 100:1 when the film is transillu­
will not change subject contrast; a film with minated and viewed (100 is the antilog of
an average gradient of less than 1 will de­ 2.0). Thus, subject contrast resulting in an
crease subject contrast. Because contrast is exposure range to the x-ray film of 4: l has
very important in radiology, x-ray films all been exaggerated, or amplified, in the
have an average gradient of greater than viewed radiograph into a brightness range
1. For example, consider a radiograph of (radiographic contrast) of 100: 1.

depicted in Figure 11-1 0. If the factors of

time, milliamperes, and focus-film distance
(the inverse square law) are correct, the log
3. 0 Ds relative exposures will produce film density

falling along the steep portion of the char­

2.5 acteristic curve. This will produce a density
difference (radiographic contrast) of 0.6,
f- 2.0 or a difference in light transmission of 4:1
[/) (antilog 0.6 = 4). If the exposure puts the
UJ developed densities on the toe of the curve,
Cl however, the film is underexposed (not
enough mAs), and the density difference
will fall to 0.13, or a difference in light
transmission of 1 .35: 1 (antilog 0.13

1 .35). Note that the exposure ratio has re­

mained the same (i.e., log relative exposure
difference of 0.2) because the kVp has not
0.3 0.9 1.5 2. 1 2.7
been changed. Similarly, overexposure, or
LOG RELATIVE EXPOSURE too many mAs, will result in densities in
the shoulder region of the characteristic
Figure 11-9 Film contrast amplifies subject curve o£ our hypothetic film-screen com­
contrast if the average gradient is greater than
bination. As shown in Figure ll-10, this
will result in a density difference (contrast)
of 0.2, corresponding to a difference in
Film Density. The slope of the charac­
teristic curve (i.e., film contrast) changes
with density. This is especially true in the
toe and shoulder regions (see Fig. ll-5). 3.0
Let us emphasize that the ratio of the dif­
ference in log relative exposure is deter­ 2.5
mined by the kVp selected (such as the 4:1
ratio between soft tissue and bone in the >-
example of the hand). If the kVp remains f- 2.0

constant, this ratio will remain constant for
any one examination despite change in ex­ UJ 1.5
posure time, milliamperes, or focus-film
distance. All these last mentioned factors,
however, will determine the actual value of
the exposure, thereby determining the lo­
cation of the exposure on the log relative 0.5
exposure axis of the characteristic curve of
the film.
Let us consider an x-ray study of the ab­
domen in which the kVp chosen results in LOG RELATIVE EXPOSURE
one area transmitting about 1 .6 times more
radiation than another. This means a log � UNDER 111111 PROPERLY OVER
relative exposure difference of 0.2 (log 1.6
= 0.2). We will assume that the film used Figure 11-10 Incorrect exposures result in
for this study has the characteristic curve loss of contrast

light transmission of 1.59:1 (antilog 0.2 = tween 50 and 100 kVp is not very great,
1.59). Exposures producing density at the however, and can usually be ignored in
level of 3 (0.1% of light transmitted) also clinical radiology.
produce less visible contrast under ordi­ Stated another way, the average gradient
nary viewing conditions, because the hu­ of a double-emulsion x-ray film will be
man eye has low sensitivity to contrast at greatest when the film is exposed with in­
low brightness levels. This is why a spot­ tensifying screens. Direct x-ray exposure
light must be used to aid in viewing regions will produce a lower average gradient. The
of high density. photochemical reasons for this phenome­
S creen or Direct X-Ray Exposure. If a non are unknown.
film designed for exposure by light from Film Processing (Develop ment). In­
intensifying screens is exposed to x rays di­ creasing the time or temperature of de­
rectly, its characteristic curve has a consid­ velopment (or both) will, up to a point, in­
erably different shape than the curve ob­ crease the average gradient of a film (film
tained from exposure with screens. speed is also increased). If development
Remember, considerably more exposure time is only 40% of normal, the gradient
(mAs) is required if no screens are used, will be reduced to about 60% of maximum.
because the intensification factor of screens Fog will also be increased with increased
may range from about 15 to 50 or more. development time or temperature, though,
Films exposed with par speed intensifying and fog decreases contrast. Therefore, it is
screens will require an x-ray exposure of important to adhere to the manufacturer's
approximately 1 mR to produce a density standards in processing film. Automatic
of 1; this value will rise to 30 mR or more film processing equipment has eliminated
with direct x-ray exposure. some problems associated with tempera­
At the same density, contrast is always ture of solutions and development time. To
lower for a film exposed to x rays only than summarize, increasing the time or temper­
for the same film exposed by light from ature of development will
intensifying screens. The reason for this
1. increase average gradient (increase
difference in contrast is not precisely
film contrast)
known. It probably is related to the com­
2. increase film speed (increase density
plex manner in which the film emulsion
for a given exposure)
responds to the energy of absorbed x-ray
3. increase fog (decrease film contrast)
In addition, intensifying screens are rel­ Figure 11-11 shows the effect of devel­
atively more sensitive than film to higher opment time (or temperature) on average
energy x rays. The primary x-ray beam gradient, film speed, and fog.
transmitted through the patient is of
higher energy than the secondary, or scat­ Speed
ter, radiation. Scatter radiation decreases The speed of a film-screen system is de­
contrast, as will be discussed in detail in fined as the reciprocal of the exposure in
Chapter 14. Because intensifying screens roentgens required to produce a density of
are relatively less sensitive to the lower en­ 1.0 above base plus fog density:
ergy of the scatter radiation, decrease in
contrast will be minimized by screens. X­ Speed= ---­

ray film, because it is more sensitive to
lower kVp x rays, may record the scatter As an example we will consider the speed
radiation better than the higher energy, in­ of a commonly used film-screen system, Du
formation-containing primary beam. The Pont Cronex 4 film combined with Du Pont
variation in film sensitivity to x rays be- Cronex Hi Plus (CaW04) screens. Analysis


FOG (/)
---!--- w



0 1.5

Figure 11-11 Development time influences 1.0

average gradient, speed, and fog

of the curve in Figure 11-12 shows that an
exposure of 0.57 mR (0.00057 R) was re­
quired to produce a net density of 1.0 QL--L--�--��--J----L�
.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
above base and fog. Therefore, the speed
(S) of the film-screen system for this kVp
lS Figure 11-12 A Cronex 4 sensitometric curve
(70-kVp exposure, Hi Plus screens) shows that
1 0.57 mR is used to reach a 1.0 net density. The
S = = 1750 R-1
0.00057 R system speed for 70 kVp is

The shape of the characteristic curve is = 1750 R-'
controlled by film contrast; the film speed 0.00057

determines the location of the curve on the (Modified from the Cronex 4 medical x-ray film
log exposure scale. data sheet available from E.I. du Pont de
Figure 11-13 shows the curves of two Nemours Company, Inc. Modified with the help
of Russell S. Holland, Ph.D.)
films that are identical except that film B
is 0.3 log relative exposure units to the
right of film A. Both films will show iden­ density of 1.0 than film A). At a density of
tical film contrast, but film B will require 1.5 both films have the same speed, and
twice (antilog 0.3 =2) as much exposure above density 1.5 the higher contrast film
(mAs) as film A. Because the gradient of a A is faster than film B. For medical x-ray
characteristic curve varies with density, the films, relative film speed is usually com­
relative speed between two films will be pared at a density of 1.0 above base and
found to vary with the density at which fog.
speed is measured. For example, if a low­
contrast and a high-contrast film are com­ Speed Class System
pared for speed, as in Figure 11-14, the The measured speed of a film-screen sys­
relative speed between the two films might tem depends on a number of variables,
actually reverse with change in density. In such as the kVp, amount of scatter radia­
our example, speed calculated at a density tion, x-ray absorption by the cassette or x­
of 1.0 will show the low-contrast film (film ray table top, and the way in which the film
B) to be the fastest (i.e., film B requires a is processed. The American National
lower log relative exposure to produce a Standards Institute (ANSI) has attempted

3.5 125, 80, 70, and 60 kVp. Film processing

is strictly controlled. If all manufacturers
3.0 followed such a rigid standard to deter­
mine speed, direct comparison of products
2.5 would be easy and meaningful. As things
exist now, however, published measure­
ments of speed must be used as approxi­
t: 2.0
t.n mations only. For this reason, we suggest
IJJ that the "speed class system" concept be
0 1.5
used to assign relative speeds to various
film-screen systems.
1.0 Refer to Table 11-2, which is the same
as Table 9-5. Notice that film-screen speed
0.5 is listed as speed class.* The numbers in
the speed class system make up a sequence
in which the logarithm of each number dif­
0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 fers from the next number by 0.1 (i.e., this
is a 0.1 log system). Table 11-3 lists the
sequence from 100 to 1000 and gives the
Figure 11-13 Film speed logarithm of each number. Note that the
log of each number increases or decreases
to standardize the way in which speed (and by 0.1 (this makes each number approxi­
contrast) of medical x-ray film-screen sys­ mately 25% greater or less than its neigh­
tems is measured.1 This standard defines bor). Thus, the ratio between numbers in
test objects that are used to simulate radi­ the sequence is a constant. Because log 2
ography of the chest, the skull or pelvis, = 0.3, addition of 0.3 to the log of a num­
and the extremities. Various exposures are ber multiplies the number by 2 (e.g., log
made at four kilovoltages, approximately 160 = 2.2, log 320 = 2.5, etc.). All of us
who use a camera have used this 0.1 log
system, usually without realizing it. Con­
sider the speed of the film you use in your
camera. Don't the numbers ASA 25, 32, 64,
125, 200, and 400 sound familiar? Table
11-3 does not list the sequence below 100,
but simply divide by 10 to go on down (i.e.,
100, 80, 64, 50, 40, 32, 25, 20, 16, and so
Why propose this speed class system?
First, it is a handy number sequence. It is
easy to multiply or divide by 10, the num­
bers increase or decrease by an easily de­
fined ratio, and a change in log of 0.3 dou­
bles or halves a number. The other reason
is that the currently available measure­
ments of film-screen system speed are not

LOG RELATIVE EXPOSURE *This idea was proposed to us by Robert Wayrynen,

Ph.D.: Reid Kellogg, Ph.D.; and Russell Holland,
Figure 11-14 Relative film speed varies with Ph.D., of the photo products department of E. I. du
the density at which speed is measured Pont de Nemours & Company, Inc.

Table 11-2. Speed Class of Various Intensifying Screens


Du Pont Cronex Par Speed cawo. Blue Cronex 4 100

Cronex Hi Plus cawo. Blue Cronex 4 250
Cronex Quanta Ill LaOBr:Tm Blue Cronex 4 BOO
Cronex Quanta V LaOBr:Tm Blue Cronex 4 320
Gd202S:Tb Green Cronex 8 400
Quanta Detail YTa04:Tm Ultraviolet/ Cronex 4 100
Quanta Fast Detail YTaO.:Nb Ultraviolet/ Cronex 4 400

Kodak X-Omatic Fine BaPbSO., Blue XRP 32

yellow dye
X-Omatic Regular BaSrSO.:Eu, Blue Ortho G 200
neutral dye
Lanex Fine Gd202S:Tb, Green Ortho G 100
neutral dye
Lanex Medium Gd202S:Tb, Green Ortho G 250
yellow dye
Lanex Regular Gd202S:Tb Green Ortho G 400

Note: Dupont Cronex 4 film is now largely replaced by Cronex 7 and Cronex 10, and Kodak Ortho
G film by T-Mat G. The older films still accurately reflect relative intensifying screen speeds, which
is the purpose of this table.

Table 11-3. Speed Class System of Numbers from 100 to 1000

Number 100 125 160 200 250 320 400 500 640 BOO 1000
Logarithm 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.B 2.9 3.0

standardized. Published values are accu­ within the accepted range for diagnostic
rate for the conditions under which they radiology (usually considered to be density
were determined, but such conditions vary. 0.25 to 2 .0) . Let us consider two hypothet­
Until rigid standards (such as the ANSI ical films (Fig. 11-15), one a high-contrast
method1) are adopted by industry, assign­ film (film A) and one a lower contrast film
ing a particular film-screen system speed (film B). If the density recorded on the film
to one of the numbers in the 0.1log system is to remain in the range of 0. 2 5 to 2. 0, film
appears reasonable. There are more than A will be limited to a log relative exposure
1000 film-screen combinations on the mar­
ket today, so some "lumping" of an oth­ 'A :s

erwise enormous variety of expressions of

2 or
speed seems necessary. Some of you may
have also noticed that the mAs stations of !:::
some x-ray generators now use an 0.1 log
sequence of steps.
0 i

0 25�
Latitude I
03 0.9 1.5 2.1 2.7
Unlike average gradient and speed, film
latitude is not expressed in numeric terms.
Latitude refers to the range of log relative Figure 11-15 Film B has greater latitude than
exposure (mAs) that will produce density film A

range of 0.75 to 1.42, or a difference in log emulsion side. An identical emulsion on

relative exposure of 0.67 corresponding to each side of the base prevents this curling.
an actual ratio of 4.68 to 1 (antilog 0.67 = The photographic advantage of a double
4.68). Film B will remain in the designated emulsion is important only when the film
density values over a range of log relative is exposed with intensifying screens. We
exposure from 0.85 to 2.35, or a difference may assume that each emulsion receives an
in log exposure of 1.50 corresponding to identical exposure from each screen, be­
an actual ratio of 31.6:1. Film B is said to cause any filtering action of the front
have greater latitude than film A in that it screen that might act to decrease the in­
will accept a wider range of exposures. tensity of x rays reaching the back screen
Note that film B has greater latitude but is small.
less film contrast. Generally speaking, the Consider the case of a single-emulsion
latitude of a film varies inversely with film film in a cassette, which receives two ex­
contrast. There are two practical aspects to posures, log E1 and E2, and responds with
this concept of film latitude. For the tech­ densities D1 and D2• Film contrast for this
nologist exposing a film, the film with more exposure difference may be expressed as
latitude makes the exposure less critical; if
02- D,
he has picked the proper kVp for adequate Contrast = _ __::.
_ ___c.
_ .
log E2- log E,
penetration, he has more room for error
in his choice of exposure (mAs). Generally, If this same exposure is now used to expose
the radiologist is interested in high con­ a double-emulsion film with light from in­
trast, which means films of less latitude. But tensifying screens, each emulsion will re­
there may be situations in which a wide spond with densities D1 and D2• When two
range of subject contrast (such as in the films are superimposed, the resulting den­
chest) must be recorded, and in such cases sity is the sum of the densities of each film.
the film with the lower contrast but higher Therefore, when the double-emulsion film
latitude may produce a radiograph in is viewed, the densities of each emulsion
which many small changes in film expo­ are added, and the resulting total density
sures (i.e., subject contrast) can be re­ will now be D1 + D1 = 2DI and D2 + D2
corded. Such lower contrast but higher lat­ = 2D2. The exposures E2 and E1 have not
itude films are available to the radiologist. changed, but we have allowed two emul­
(We will discuss another aspect of exposure sions to respond to the exposure. The con­
latitude as it pertains to kVp and subject trast that the eye now sees is
contrast in Chapter 14).
202- 20, 2(02 - D,)
1og E2- 1og E, log E2- log E,
Double-Emulsion Film

Films used for routine radiography have Because log E2 - log E1 is the same for
photosensitive emulsion coated on both each exposure, we may compare contrast
sides of the base support. There is a phys­ without the log E2 - log E1 term:
ical and photographic reason for this. The
Single emulsion contrast = 02- D,
emulsion is applied to the base in liquid
form. When the emulsion dries, it shrinks
Double emulsion contrast = 2(02 - D,)
to about one tenth of its original volume.
Most of the decrease in volume causes a The double-emulsion film has produced
decrease in thickness of the emulsion, but twice the contrast of a single-emulsion film.
there is also a slight tendency to shrink in Obviously, the overall density of the dou­
area. If emulsion were put on only one side ble-emulsion film is increased, resulting in
of the base, shrinking of the emulsion increased film speed. For a film being ex­
would cause the film to curl toward the posed to x rays directly, a similar effect

could be produced by making a single­ peak wavelengths from some rare earth
emulsion film with a thicker emulsion. Be­ screens, such as Kodak Lanex screens, are
cause light photons are easily absorbed by produced by the terbium ion with about
the emulsion, however, only the outer layer 60% of its energy at about 544 nm (green
of the emulsion is affected by light from light). It is possible to extend the sensitivity
intensifying screens. This is one reason of film to the green wavelengths by coating
why x-ray film designed for exposure by the silver halide grains with a thin layer of
light from two intensifying screens (in a dye that absorbs the green light and then
cassette) has a thin emulsion on each side transfers this absorbed energy to the grain.
of the base, rather than a single thick emul­ Such green-sensitive film is called ortho
ston. film. Similarly, the silver halide grain can
be coated with a dye that absorbs red light,
EMULSION ABSORPTION and this is called a pan film (panchromatic:
To expose an x-ray film with intensifying sensitive to light of all colors). This is illus­
screens, it is necessary for the silver halide trated in Figure 11-17. When rare earth

grains in the film to absorb the light emit­ screens are used, an appropriate film

ted by the screen phosphor. The ability of should be used if one is to take advantage
the film grains to absorb light depends on of all the light emitted by the screen. Such

the wavelength, or color, of the light. a combination is the Kodak Lanex screen

Standard silver halide films absorb light and Ortho G film (Fig. 11-18).

in the ultraviolet, violet, and blue regions Please look back at Figure 9-12, which
of the visible spectrum. Used with calcium shows the spectral emission of LaOBr:Tm
tungstate or barium lead sulfate screens, (found in Du Pont Cronex Quanta III and

such films worked well because these phos­ Quanta V screens). Remember that this

phors emitted light that was absorbed by rare earth, activated by thulium, produces

natural silver halide (Fig. 11-16). Note that light to which natural silver halide film is

natural silver halide film does not absorb sensitive. Compare the twin peaks of 374
in the green and yellow portions of the vis­ nm and 463 nm in Figure 9-12 to the sen­
ible spectrum, where much of the light sitivity of silver halide as shown in Figure

from some rare earth phosphors is emitted 11-17. The more recently introduced yt­
(see Fig. 11-18). You will recall that the trium tantalate intensifying screens also
emit light in the ultraviolet and blue wave­
lengths to which natural silver halide films
exhibit maximum sensitivity. Some rare
earth phosphor screens use green-sensitive

w >­
z >

w <f>
/ ORTH0 7
_Jw "­
0: ��T=�-.---.----.---.-�-.- >
300 400 500 600 <l

0: �------.---�---l_,--�-
300 400 500 600 700
Figure 11-16 The spectral emission of barium WAVELENGTH in nm
lead sulfate and calcium tungstate intensifying
screens compared to the spectral sensitivity of Figure 11-17 Relative spectral sensitivity of
natural silver halide (not drawn to scale) natural silver halide, ortho, and pan films

>- 2 �<n
t:: Q <n
> <J) �
i= 1!2 ,_<n
2 >O:
w 2 >=>­
<llw ;;; ....
- u
u_ <J) ��
ww "-"'
> > ww
.... ....
<1 <1 ��
_j _j _J_J

ww �� '------'---"'--L---'"--'
0: 0:
300 400 500 600 700
300 400 500 600 VIOLET

Figure 11-19 Spectral sensitivity of natural
Figure 11-18 Spectral emission of Kodak La­ silver halide and ortho film compared to the
nex Regular screens and the spectral sensitivity transmission of an amber and a red safelight
of Kodak Ortho G film (the bars and curve are filter
approximations used for illustration only, and
no attempt to achieve accuracy was made)
opposite the emulsion. The main cause of
this crossover is incomplete absorption of
film, some use blue-sensitive film, and some
light by the adjacent emulsion. This un­
(e .g., Cronex Quanta V) may use either.
absorbed light passes through the film base
to reach the opposite emulsion. The cross­
Darkroom Safelight
over light is spread because of diffusion,
The use of ortho films requires that the scattering, and reflection caused by the film
correct darkroom safelight be used. For base and interfaces between the emulsions
many years an amber safelight (such as the and film base. Crossover exposure is a sig­
Kodak Type 6B filter) has been used with nificant contributing factor to unsharpness
blue-sensitive films. The amber safelight in film-screen systems.
emits light to which ortho film is sensitive, How important is this crossover light in
however, and will produce "safelight fog" terms of total film exposure? With green­
if used with such film. With ortho films a sensitive films, up to 40% of the total ex­
safelight filter shifted more toward the red posure is attributed to print-through. Blue­
is required (this removes the green light to sensitive emulsions (natural silver halide)
which ortho film is sensitive). Such a red are slightly more efficient light absorbers
safelight filter is the Kodak GBX-2 filter. and, with blue light systems, crossover is
Figure 11-19 illustrates the approximate responsible for less than one third of the
sensitivity of regular silver halide film and total exposure (about 30 to 32% with
ortho film as compared to the transmission CaW04 screens, and only 23% with
of an amber (6B) and a red (GBX-2) safe­ BaSrS04 screens).
light filter. The ideal way to reduce print-through
is to increase light absorption in the silver
Crossover Exposur� halide grains of the film emulsion. This
Crossover exposure, also called "print­ would improve image quality without re­
through exposure," occurs when a double­ ducing system speed. The original film de­
emulsion x-ray film is exposed in a cassette signed to reduce print-through had a light­
containing two intensifying screens. Ide­ absorbing dye coated on both sides of the
ally, each film emulsion would receive light film base. This anticrossover dye absorbed
only from the screen in contact with the light attempting to diffuse through the
emulsion. Crossover is the exposure of a base into the opposite emulsion, and de­
film emulsion to light emitted by the screen creased crossover exposure to only about

13% of the total exposure. System speed approximate indications of decrease in

was decreased by about 40% with this film. crossover. There are different ways of de­
At present, two technologies attempt to termining the fraction of crossover, and
reduce crossover exposure by increasing significantly different answers result with
light absorption in the film emulsion. different methods. One must not use the
These technologies involve: actual numbers to compare products. We
report some figures from the literature to
1. Matching screen light emission to sil­
illustrate the concept of reduction in print­
ver halide natural sensitivity
through, not as a means of deciding which
2. Changing the shape of the silver hal­
type of film or screen does the best job.
ide grains

In Chapter 9, we discussed yttrium tan­ TRANSPARENCY VERSUS PRINT

talate phosphor intensifying screens. The Why is a radiograph viewed as a trans­

ultraviolet/blue light emission of this phos­ parency rather than as a print, like an or­
phor is an excellent match for the natural dinary photograph? The density of a print
light-absorbing sensitivity of silver halide. is related to the amount of light reflected
For example, DuPont reports a reduction or absorbed by the paper. The density of
in print-through from 33% using calcium the maximum black of most photographic
tungstate screens to 19% using the same printing papers is between 1.3 and 1.7. A
film with yttrium tantalate screens. Print­ few papers give density values as high as
through can be decreased even more when 2.0 (glossy-surfaced paper gives the highest
a dye, matched to the light emission of the maximum black). Obviously, such a limi­
screens, is added to the emulsion. Using a tation on maximum density would be in­
film that incorporates this dye technology tolerable in radiology, in which densities up
reduces print-through to only 14% of total to 2.0, and occasionally greater, are com­
exposure. The film incorporating dye in monly encountered. This limitation is over­
the emulsion is a higher speed film, the come by viewing the radiograph as a trans­
result being no loss of system speed. parency, which permits use of density
A different approach to increasing light ranges up to a maximum of 3.0 or more
absorption in the film emulsion was intro­ for diagnostic radiology films and up to 6.0
duced by Kodak: tabular silver halide for industrial x-ray use.
grains. Tabular grains of silver halide have
a thickness that is less than one tenth their SUMMARY
diameter. This results in flat film emulsion The amount of blackening of an x-ray
grains that present a much larger surface film is expressed by the term "photo­
area to incident light photons when com­ graphic density." The most useful range of
pared to an equal mass of conventionally density in diagnostic radiology is 0.25 to
shaped grains. This large surface-area-to­ 2.0, although densities up to 3.0 are some­
volume ratio allows tabular grains to absorb times used.
significantly more light photons when com­ Analysis of the characteristic curve of a
pared to an emulsion containing an equal film provides information about the con­
mass per unit area of conventional grains. trast, speed, and latitude of the film. Film
Reduction in crossover is due to increased contrast amplifies subject contrast if the
absorption in the tabular silver halide average gradient of the film is greater than
grains. The reduction in the crossover ex­ 1. Film contrast will vary with the amount
posure fraction for conventional versus of exposure (density), the way the film is
tabular grain film is in the range of 29 to exposed (intensifying screens or direct ac­
30%, versus 15 to 19%. We must emphasize tion of x rays), and the way the film is de­
that these percentages are only useful as veloped.

Double-emulsion films produce greater 6. Lawrence, D.J.: Kodak X-Omatic and Lanex
screens and Kodak films for medical radiography.
contrast than single-emulsion films (assum­ Rochester, NY, Eastman Kodak Company, Ra­
ing light is used to expose the film). diography Markets Division, File No. 5.03, June
X-ray film is viewed as a transparency
7. Meredith, W.J., and Massey, J.B.: Fundamental
because of the greater range of density Physics of Radiology. Baltimore, Williams & Wil­
available. kins, 1968.
The film must be able to absorb the wave­ 8. Ostrum, B.J., Becker, W., and Isard, H.J.: Low­
dose mammography. Radiology, 109:323, 1973.
length of light emitted by the intensifying 9. Presentation Script: New Screen Technology.
screen. Prepared by Photo Products Department. Wil­
mington, Del., E. I. du Pont de Nemours & Com­
!0. Rao, G.U.V., Fatouros, P.P., and James, A.E.:
l. American National Standards Institute: Ameri­ Physical characteristics of modern radiographic
can National Standard Method for the Sensitom­ screen-film systems. Invest. Radio!., 13:460, 1978.
etry of Medical X-Ray Screen-Film-Processing 11. Seeman, H.E.: Physical and Photographic Prin­
Systems. New York, American National Stan­ ciples of Medical Radiography. New York, John
dards Institute, 1982. (Available as ANSI PH2.43- Wiley and Sons, 1968.
1982 from the American National Standards In­ 12. Sensitometric Properties of X-Ray Films. Roch­
stitute, 1430 Broadway, New York, NY 10018.) ester, NY, Eastman Kodak Company, Radiogra­
2. Bates, L.M.: Some Physical Factors Affecting Ra­ phy Markets Division.
diographic Image Quality: Their Theoretical Ba­ 13. Thompson, T.T.: Selecting medical x-ray film.
sis and Measurement. Washington, DC, U.S. Gov­ Part l. Appl. Radio!., 4:47, 1974.
ernment Printing Office, 1969, Public Health 14. Thompson, T.T.: Selecting medical x-ray film.
Service Pub. No. 999-RH-38. Part II. Appl. Radio!., 4:51, 1974.
3. Brixner, L., Holland, R.S., Kellog, R.E., Mickish, 15. Wayrynen, R.E.: Radiographic film. Contained in
D., Patten, S.H., Zegarski, W.: Low print-through the syllabus of the 1975 AAPM Summer School:
technology with rare earth tantalate phosphors. The Expanding Role of the Diagnostic Radiologic
SPIE V ol. 555 Medical Imaging and Instrumen­ Physicist, Rice University, Houston, July
tation, 1985, p. 84. 27-August 1, 1975, p. 112.
4. Doi, K., Loo, L.N., Anderson, T.M., and Frank, 16. Wayrynen, R.E.: Fundamental aspects of mam­
P.H.: Effect of crossover exposure on radio­ mographic receptors film process. In Reduced
graphic image quality of screen-film systems. Ra­ Dose Mammography. Edited by W.W. Logan and
diology, 139:707, 1981. E.P. Muntz. New York, Masson Publishing USA,
5. Huff, K.E., and Wagner, P.W.: Eastman Kodak 1979, pp. 521-528.
Company Report: Crossover and MTF charac­ 17. Weiss, J.P., and Wayrynen, R.E.: Imaging system
teristics of a tabular-grain x-ray film. Eastman for low-dose mammography. ]. Appl. Photogr.
Kodak Company, Rochester, New York !4650. Eng., 2:7, 1976.

12 Fluoroscopic Imaging

X rays were discovered because of their nym for fluoroscopy in the good old days).
ability to cause fluorescence, and the first A sheet of lead glass covered the screen, so
x-ray image of a human part was observed the radiologist could stare directly into the
fluoroscopically. Dr. Glasser, in his book, screen without having the x-ray beam
Dr. W. C. Rontgen, recounted: strike his eyes. Screen fluorescence was so
faint that fluoroscopic examinations were
To test further the ability of lead to stop the
rays, he selected a small lead piece, and in carried out in a dark room by a radiologist
bringing it into position observed to his who had dark-adapted his eyes by wearing
amazement, not only that the round dark red goggles for 20 to 30 minutes prior to
shadow of the disc appeared on the screen,
the examination.
but that he actually could distinguish the out­
Residents trying to learn fluoroscopic
line of his thumb and finger, within which
appeared darker shadows-the bones of his techniques were at an enormous disadvan­
hands.1 tage because it was impossible to see what
the professor claimed he was seeing on the
THE FLUOROSCOPE dim screen . The days of red goggles and
The first generation fluoroscopes con­ green screens are gone forever.
sisted of an x-ray tube, an x-ray table, and
a fluoroscopic screen (Fig . 12-1). The flu­ Visual Physiology
orescent material in the screen was copper­ In 1941, at a meeting of the Radiologic
activated zinc cadmium sulfide that emitted Society of North America in Chicago, Dr.
light in the yellow-green spectrum (giving W. Edward Chamberlain startled his au­
rise to the term "green screen" as a syno- dience when he announced that the light


Figure 12-1 Fluoroscope


reflected from the piece of paper he was The dim fluoroscopic images required
holding in his hand was 30,000 times use of rod vision, with its poor visual acuity
brighter than the illumination of a fluor­ and poor ability to detect shades of gray
oscopic screen.3 Radiologists had always (contrast). What was needed was a way to
known that the fluoroscopic screen was produce an image bright enough to allow
poorly illuminated, but no one before cone vision without giving the patient an
Chamberlain had ever bothered to meas­ excessive radiation exposure. The solution
ure it. How can we see anything with so came in the form of the x-ray image inten­
little light? The answer is found in the eye's sifier, developed in the 1950s.
amazing ability to adapt to low levels of
The retina contains two different types The components of an x-ray image in­
of light receptors, rods and cones. Cones tensifier are shown in Figure 12-2. The
(central vision) function most efficiently in tube itself is an evacuated glass envelope,
bright light, while rods (peripheral vision) a vacuum tube, which contains four basic
function best with low levels of illumina­ elements:
tion. Daylight (cone) vision is called pho­
1. input phosphor and photocathode
topic vision, and night (rod) vision is called
2. electrostatic focusing lens
scotopic vision. The differences between
3. accelerating anode
day and night vision are so great that hu­
4. output phosphor
mans can be considered to have two almost
totally separate vision systems. After an x-ray beam passes through the
The cones are concentrated very densely patient, it enters the image intensifier tube.
in the fovea at the center of the retina, and The input fluorescent screen absorbs x-ray
are sparsely scattered over the rest of the photons and converts their energy into
retina. The dense concentration of cones light photons. The light photons strike the
gives high visual acuity for direct vision, photocathode, causing it to emit photo­
and the sparse population in the remainder electrons. These electrons are immediately
of the retina contributes to daylight pe­ drawn away from the photocathode by the
ripheral vision. The cones are almost com­ high potential difference between it and
pletely blind to low levels of illumination. the accelerating anode. As the electrons
There are no rods in the fovea, so sco­ flow from the cathode toward the anode,
topic vision is entirely peripheral vision. they are focused by an electrostatic lens,
Also, the density of rods (and the inter­ which guides them to the output fluores­
connecting network of nerves) is less over cent screen without distorting their geo­
the remainder of the retina than the den­ metric configuration. The electrons strike
sity of cones in the fovea. The result is that the output screen, which emits the light
scotopic (rod) vision is less acute than pho­ photons that carry the fluoroscopic image
topic (cone) vision. Rods can be extremely to the eye of the observer. In the intensifier
sensitive to low levels of illumination. Rods tube, the image is carried first by x-ray pho­
are most sensitive to blue-green light, and tons, then by light photons, next by elec­
daylight levels of these wavelengths of light trons, and finally by light photons.
greatly reduce the sensitivity of rods to low
(night) illumination levels. Fluoroscopists Input Phosphor and Photocathode

had to "dark adapt" by wearing red goggles The input fluorescent screen in image
to filter out blue-green wavelengths for pe­ intensifiers is cesium iodide (Csl). The in­
riods of over half an hour to allow the rods put phosphor of older image intensifiers
to recover peak sensitivity before fluoros­ was silver-activated zinc-cadmium sulfide.
copy. Csl is deposited on a thin aluminum sub-



------------- 1 -------+

X-ray Electrons
Photons --- -- --------..
Electrons Light

-------fti"in=.: - -----,
Aluminum � Fluorescent
Layer Screen

Figure 12-2 Input phosphor and photocathode (A) and output phosphor of an image intensifier

strate by a process called "vapor deposi­ with zinc-cadmium sulfide to 0.1 mm with
tion." An interesting and useful character­ cesium iodide. The principal advantage of
istic of Csi is that during the deposition the thinner phosphor layer combined with
process the crystals of Csi grow in tiny nee­ needle-shaped crystals is improved reso­
dles perpendicular to the substrate. This is lution. The resolution of a Csi image in­
useful because the phosphor layer exhibits tensifier will be about 4 line pairs per mil­
minimal lateral light diffusion. When one limeter (with a range of 3 to 5 lp/mm).
of these needle-shaped crystals absorbs an Ideally, for maximum photoelectric ab­
x-ray photon and produces light, the light sorption, the K-absorption edge of a phos­
will be transmitted to the photocathode phor should be as close to the energy of
with little scattering. the x-ray beam as possible, provided the
Image quality is dramatically better with energy of the edge does not exceed that of
cesium iodide input screens than it was with the beam. An obvious problem arises in
the older zinc -cadmium sulfide screens. attempting to accomplish this ideal. Beam
Three physical characteristics of cesium io­ energy is a spectrum, a whole array of en­
dide make it superior; the vertical orien­ ergies, whereas the K edge is a single en­
tation of the crystals, a greater packing ergy or, at most, several energies, depend­
density, and a more favorable effective ing on the number of different absorbers
atomic number. Since cesium iodide can be in the phosphor. The mean energy of an
vacuum-deposited, it requires no inert x-ray beam is approximately one third of
binder, so more active material can be its peak energy, depending somewhat on
packed into a given space. The packing the kV p and filtration. Most fluoroscopy on
density of cesium iodide is three times adults is done at a peak energy of from 80
greater than that of zinc-cadmium sulfide. to 120 kVp, which is a mean energy be­
Phosphor thicknesses have been reduced tween 30 and 40 keV. Table 12-1 shows
comparably from approximately 0.3 mm the Kedges of some elements used in phos-

Table 12-1. Atomic Number and electrons produced are proportional to the
K-Absorption Edge intensity of the light (Fig. 12-3A).
Now we have to get the electrons to the
other end of the image intensifier tube and
Sulfur 16 2.5
make them maintain their relative position.
Zinc 30 9.7
Cadmium 48 26.7 We do this using an electrostatic focusing
Iodine 53 33.2 lens and an accelerating anode.
Cesium 55 36.0
Electrostratic Focusing Lens
The lens is made up of a series of posi­
phors. The energy of the K edge of cad­ tively charged electrodes that are usually
mium (26.7 keV) is quite good, but its plated onto the inside surface of the glass
chemical mates, zinc (9.7 keV) and sulfur envelope. These electrodes focus the elec­
(2.5 keV), are far from ideal. The Kedges tron beam as it flows from the photocath­
of cesium (36 keV) and iodine (33.2 keV) ode toward the output phosphor. Electron
are almost perfect. The more appropriate focusing inverts and reverses the image.
atomic numbers of cesium and iodine give This is called a point inversion because all
these screens a substantial advantage over the electrons pass through a common focal
those made of zinc-cadmium sulfide. Ce­ point on their way to the output phosphor
sium iodide input screens absorb approx­ (Fig. 12-2). Each point on the input phos­
imately two thirds of the incident beam as phor is focused to a specific point on the
opposed to less than one third for zinc­ opposite side of the output phosphor. For
cadmium sulfide, even though the cesium undistorted focusing, all photoelectrons
iodide screen is only one third as thick. must travel the same distance. The input
The photocathode is a photoemissive phosphor is curved to ensure that electrons
metal (commonly a combination of anti­ emitted at the peripheral regions of the
mony and cesium compounds). When light photocathode travel the same distance as
from the fluorescent screen strikes the pho­ those emitted from the central region. The
tocathode, photoelectrons are emitted in image on the output phosphor is reduced
numbers proportional to the brightness of in size, which is one of the principal reasons
the screen. The photocathode is applied
directly to the Csl input phosphor. Light
from the Csl passes directly into the pho­
tocathode. Older tubes had a thin light 1+---- Anode

transparent barrier between the input Evacuated Glass

phosphor and the photocathode. Light dif­ Envelope

fusion in this barrier reduced resolution. Focal Point

Let us review the process to this point.
A uniform beam of x rays has passed
through a patient and been attenuated by
Electrostatic Lens
the patient. This attenuated beam ofx rays
passes through the glass front of the image
intensifier tube and the thin aluminum
substrate of the input phosphor layer (Csl).
The Csl crystals produce light in propor­
Photocathode ond
tion to the intensity of the incident x-ray Input Fluorescent
beam. The light photons react with the Screen

photocathode. Photoelectrons are emitted

from the photocathode. The numbers of Figure 12-3 X-ray image intensifier tube

why it is brighter, a point to which we will

return later.
Accelerating Anode. The anode is lo­
cated in the neck of the image tube (Fig.
12-2). Its function is to accelerate electrons
emitted from the photocathode toward the
output screen. The anode has a positive
potential of 25 to 35 kV relative to the pho­
tocathode, so it accelerates electrons to a
tremendous velocity.
Output Phosphor. The output fluoro­
rescent screen of image intensifiers is sil­
ver-activated zinc-cadmium sulfide, the
same material used in first-generation in­
put phosphors. Crystal size and layer thick­
ness are reduced to maintain resolution in /
the minified image. The diameter of most
output screens ranges from about \{ to 1
in. Because the electrons are greatly accel­ Figure 12-4 Mirror optical system of an
image intensifier
erated, they emit more light photons from
the output screen than were originally
present at the input screen. The number limited by the mirror, making it difficult,
of light photons is increased approximately if not impossible, to palpate the patient.
50-fold. Also, only one observer can view the image,
A thin layer of aluminum is plated onto which is a serious disadvantage in training
the fluorescent screen (Fig. 12-3B) to pre­ beginning fluoroscopists.
vent light from moving retrograde Viewing the output of an image inten­
through the tube and activating the pho­ sifier (II) is done via a closed circuit tele­
tocathode. The aluminum layer is very vision chain in modern systems. But we
thin, and high-energy photoelectrons eas­ would also like to be able to record the
ily pass through it en route to the output image on photospot or cine film when ap­
screen. propriate. We could record the film image
The glass tube of the image intensifier is from the TV camera signal, but the re­
about 2 to 4 mm thick, and is enclosed in sulting image would be degraded by the
a lead-lined metal container. The lead lin­ television chain. It is better to expose the
ing protects the operator from stray radi­ film directly to the output phosphor of the
ation. image intensifier tube. To maintain contin­
The output phosphor image is viewed uous TV viewing while exposing the film
either directly through a series of lenses means that we must split the light from the
and mirrors, or indirectly through closed­ II output into two paths at the time of film
circuit television. A mirror optical system exposure. During routine fluoroscopy, all
is shown in Figure 12-4. As you can see, the light output of the II is directed to the
light travels a long distance, and it is re­ TV camera. When film mode is selected, a
flected and focused several times. This can semitransparent (often termed "partially
be done with only minimal loss of bright­ silvered") mirror is positioned in the light
ness. The mirror image is only visible in a beam, as shown in Figure 12-5. With this
small viewing angle. If the operator moves arrangement, most of the light (about 90%)
his head a few degrees to one side, the goes to the film camera, but enough can
image is lost. His freedom of movement is pass through the mirror to form a TV pic-



l � ' '

Lead lined '

\ '

intensifier tube \, /1

housing Film

Figure 12-5 Optical coupling between an image intensifier and viewing system

ture. Since the exposure level has been in­ The brightness gain is the ratio of the two
creased, 10% of the available light will still illuminations:
produce a satisfactory TV image. In the
intensifier luminance
filming mode, the exposure is not contin­ Brightness gain = -------­
Patterson B-2 luminance
uous so the TV picture comes on only dur­
ing the exposure time and shows the actual If the image intensifier is 6000 times
image being recorded by the film. We will brighter, the brightness gain is 6000. The
return to the topic of photospot and cine concept is easy to understand, and was
cameras in the next chapter. readily accepted by radiologists. Patterson­
In some systems, the image is coupled to type B-2 fluoroscopic screens, however,
the TV camera by a fiberoptic bundle (fiber vary from one batch to another, and de­
face plate). Fiberoptic coupling precludes teriorate at an unpredictable rate with time
filming directly from the output phosphor so brightness gain measurements are not
of the image intensifier tube. A fiber face reproducible. Because of this lack of re­
plate is a bundle of fine optically shielded producibility, the International Commis­
glass fibers (several thousand per mm2) that sion on Radiologic Units and Measure­
is a few millimeters thick. In the future it ments (ICRU) has recommended a second
may be possible to obtain films that are ac­ method of evaluation, called the "conver­
ceptable from the TV signal, and we pre­ sion factor," to supersede the older bright­
dict that all image tubes will be coupled ness gain method. The conversion factor
directly to TV cameras. is a ratio of the luminance of the output
phosphor to the input exposure rate:

. cd/m2
Convers1on factor
= -­


Two methods are used to evaluate the Output screen luminance is measured m

brightness gain of image intensifiers. The candelas (abbreviated cd, and defined as
first compares the luminance of an inten­ the luminous intensity, in the perpendic­
sifier output screen to that of a Patterson­ ular direction, of a surface of a l/600,000
type B-2 fluoroscopy screen when both are m2 of a black body at the temperature of
exposed to the same quantity of radiation. freezing platinum under a pressure ·of

101,325 Nt/m2 . .. ridiculous to remem­ output screen approximately 1 in. in di­

ber). Radiation quality and output lumi­ ameter. Image intensifiers with 12- to 16-
nance are explicitly defined, so the method in. diameter input screens are available. We
is accurate and reproducible. will briefly consider these big tubes shortly.
The brightness gain, often called "inten­ With a l-in. output screen, the minification
sification factor," of modern image inten­ gain is simply the square of the diameter
sifiers can easily reach 10,000 (values of of the input screen; that is, a 9-in. inten­
20,000 are possible). The conversion factor sifier has a gain of 81.
usually equals about 1% of the brightness The brightness gain from minification
gain, so a conversion factor of 100 is about does not improve the statistical quality of
the same as brightness gain of 10,000. Most the fluoroscopic image.The same number
specifications for image intensifiers will of light photons make up the image re­
quote the conversion factor. gardless of the size of the output screen.
The brightness gain tends to deteriorate For example, a 6-in.intensifier with a 2-in.
as an image intensifier ages. This means output phosphor has a minification gain of
that the patient dose with an old image in­ 9 (62 -;- 22 = 36 -;- 4 9), whereas the

tensifier tends to be higher than that with same intensifier with a l-in.output phos­
a new intensifier of the same type.Because phor has a gain of 36. The total light output
the deterioration can proceed at a rate of of both units is exactly the same, however,
about 10% per year, a periodic check of so the same number of photons make up
image intensifier brightness can be valua­ both images.The photons are compressed
ble. Unfortunately, an accurate check is together on the smaller screen and the
somewhat complicated. Some indication of image is brighter, but its statistical quality
image intensifier aging can be obtained by is not improved.
comparing the input dose level required Theoretically, brightness can be in­
for automatic brightness control operation creased indefinitely by minification. A 9-in .
with the dose level under the same condi­ intensifier with a 1/win. output screen
tions as when the intensifier was new. would have a brightness gain from mini­
The brightness gain of an image inten­ fication alone of over 20,000. Excess mini­
sifier comes from two completely unrelated fication produces a very small image,
sources, called "minification gain" and though, which has a definite disadvantage.
"flux gain." We will discuss them separately. Before the image can be viewed it must be
greatly magnified, which not only reduces
Minification Gain the brightness but also magnifies the fluor­
The brightness gain from minification is oscopic crystals in the output screen, re­
produced by a reduction in image size. The sulting in a precipitous drop in resolution.
quantity of the gain depends on the relative
Flux Gain
areas of the input and output screens. Be­
cause the size of an intensifier is usually Flux gain increases the brightness of the
indicated by its diameter, it is more con­ fluoroscopic image by a factor of approx­
venient to express minification gam m imately 50. For each light photon from the
terms of diameter: input screen, 50 light photons are emitted
by the output screen. In simplified terms,
Minification gain =

(�od 1)2 you may think of one light photon from

the input screen as ejecting one electron
where d1 is the diameter of input screen, from the photocathode.The electron is ac­
and d0 is the diameter of output screen. celerated to the opposite end of the tube,
Most x-ray image intensifiers have an input gaining enough energy to produce 50 light
screen from 5 to 9 in. in diameter and an photons at the output screen.

The total brightness gain of an image fluoroscopic image. Some light photons
intensifier is the product of the minification penetrate through the aluminum, pass
and flux gains: back through the image tube, and activate
the photocathode. The cathode emits pho­
Brightness gain = minification gain x flux gain
toelectrons, and their distribution bears no
For example, with a flux gain of 50 and a relationship to the principal image. These
minification gain of 81 (9-in. intensifier electrons produce "fog" and further re­
with a l-in. output screen), the total bright­ duce image contrast. Contrast tends to de­
ness gain is 4050 (50 x 81). teriorate as an image intensifier ages.


Contrast Applied to image intensifiers, lag is de­
Contrast can be determined in vanous fined as persistence of luminescence after
ways, and there is no universal agreement x-ray stimulation has been terminated.
as to which is best. The following is a de­ With older image tubes, lag times were 30
scription of the simplest method. A Y:;.-in. to 40 ms. With Csl tubes, lag times are
thick lead disc is placed over the center of about l ms. The lag associated with vidicon
the input screen. Disc size is selected to television tubes is now more of a concern
cover 10% of the screen; that is, a 0.9-in. than image intensifier lag times.
diameter disc is used for a 9-in. image in­
tensifier. The input phosphor, with the disc Distortion

in place, is exposed to a specified quantity The electric fields that accurately control
of radiation, and brightness is measured at electrons in the center of the image are not
the output phosphor. Contrast is the capable of the same degree of control for
brightness ratio of the periphery to the cen­ peripheral electrons. Peripheral electrons
ter of the output screen. Defined in this do not strike the output phosphor where
way, contrast ratios range from approxi­ they ideally should, nor are they focused
mately l0: l to better than 20: l, depending as well. Peripheral electrons tend to flare
on the manufacturer and intended use. out from an ideal course. The result is un­
Two factors tend to diminish contrast in equal magnification, which produces pe­
image intensifiers. First, the input screen ripheral distortion. The amount of distor­
does not absorb all the photons in the x­ tion is always greater with large intensifiers
ray beam. Some are transmitted through because the further .an electron is from the
the intensifier tube, and a few are even­ center of the intensifier, the more difficult
tually absorbed by the output screen. it is to control. Figure 12-6 shows a cine
These transmitted photons contribute to image of a coarse wire screen taken with a
the illumination of the output phosphor 9-in. intensifier. As you can see, the wires
but not to image formation. They produce curve out at the periphery; this effect is
a background of fog that reduces image most noticeable at the corners. This same
contrast in the same way that scattered effect has been observed in optical lenses
x-ray photons produce fog and reduce con­ and termed the "pincushion effect," and
trast in a radiographic image. The second the term is carried over to image tubes. The
reason for reduced contrast in an image distortion looked like a pincushion to the
intensifier is retrograde light flow from the guy who named it. Generally this distortion
output screen. Most retrograde light flow does not hamper routine fluoroscopy, but
is blocked by a thin layer of aluminum on it may make it difficult to evaluate straight
the back of the screen. The aluminum layer lines (for example, in the reduction of a
must be extremely thin, however, or it fracture).
would absorb the electrons that convey the Unequal magnification also causes un-



Figure 12-6 Test film of a wire screen (35 mm

cine frame) from a 9-in. image intensifier 611MODE

Figure 12-7 Dual-field image intensifier

equal illumination. The center of the out­
put screen is brighter than the periphery
shows this principle applied to a dual-field
(Fig. 12-6). The peripheral image is dis­
image intensifier. In the 9-in. mode, the
played over a larger area of the output
electrostatic focusing voltage is decreased.
screen, and thus its brightness gain from
The electrons focus to a point, or cross,
minification is less than that in the center.
close to the output phosphor, and the final
A fall-off in brightness at the periphery of
image is actually smaller than the phos­
an image is called vignetting. Unequal fo­
phor. In the 6-in. mode the electrostatic
cusing has another effect on image quality;
focusing voltage is increased, and the elec­
that is, resolution is better in the center of
trons focus farther away from the output
the screen.
phosphor. After the electrons cross, they
In summary, the center of the image in­
diverge, so the image on the output phos­
tensifier screen has better resolution, a
phor is larger than in the 9-in. mode. The
brighter image, and less geometric distor­
optical system is preset to cover only the
format, or size, of the smaller image of the
9-in. mode. In the 6-in. mode, the optical
system "sees" only the central portion of
the image, the part derived from the cen­
Dual-field or triple-field image intensi­ tral 6 in. of the input phosphor. Because
fiers attempt to resolve the conflicts be­ this image is less minified, it appears to be
tween image size and quality. They can be magnified when viewed through a televi­
operated in several modes, including a 4.5- sion monitor. The physical size of the input
in., a 6-in., or 9-in. mode. The 9-in. mode and output screens is the same in both
is used when it is necessary to view large modes; the only thing that changes is the
anatomic areas. When size is unimportant, size of the output image. Obviously, the 6-
the 4.5- or 6-in. mode is used because of and 9-in. modes have different minifica­
better resultant image quality. Larger tion gains. Exposure factors are automat­
image intensifiers (12- to 16-in.) frequently ically increased when the unit is used in the
have triple-field capability. 6-in. mode to compensate for the de­
Field size is changed by applying a simple creased brightness from minification.
electronic principle: the higher the voltage While we are discussing intensifier size,
on the electrostatic focusing lens, the more there is another point to consider. A 9-in.
the electron beam is focused. Figure 12-7 image intensifier does not encompass a

procedures. Remember that there is �n au­

tomatic increase in exposure rate m the

magnified viewing mode. Contrast and res­
olution will be improved when only the

1 central area of these large tubes is used,

and distortion will be minimized. These
tubes are large, somewhat bulky to use, and
Figure 12-8 Reduction of fluoroscopic field very expenstve.
size by an image intensifier
9-in. field in the patient. The x-ray image
Early fluoroscopy was accomplished by
is magnified by divergence of the beam
radiologists looking directly at a fluoro­
(Fig. 12-8). The intensifier sees a much
scopic screen. The image on the screen was
smaller field than its size would imply, an
only .0001 as bright as the image of a rou­
important point to consider when ordering
tinely viewed radiograph, so dark adapta­
a unit to perform a particular function.
tion of the eyes was required.
In the 1950s the image intensifier alle­
Large Field of View Image Intensifier
viated this situation by producing an image
bright enough to be viewed with cone vi­
The development of digital angiography
sion . The input phosphor of modern
was associated with a need for large image
image intensifiers is cesium iodide; the out­
intensifier tubes that could image a large
put phosphor is zinc cadmium sulfide
area of the patient. This is especially true
(green light). Brightness gain is the pr�d­
in abdominal angiography. Image intensi­
uct of minification gain and flux gam.
fier tubes with diameters of 12, 14, and 16
Imaging characteristics important in the
in. are now available to meet this need.
evaluation of image intensified fluoroscopy
Conventionally sized image intensifier
include contrast, lag, and distortion. Large
tubes (up to about 10 in.) are made with a
field of view image intensifier tubes are
glass envelope with a glass convex input
available to fill special needs, such as digital
window. Use of a glass input window on
and spot-film angiography. Most image in­
the new large tubes became impractical be­
tensifiers allow dual-field or triple-field
cause the glass would be so thick it would
absorb a significant fraction of the x-ray
beam (remember that these tubes are un­ REFERENCES
der a high vacuum) . These larger tubes l. Glasser, 0.: D r. W.C. Rontgen. Springfield, IL,
may have an all-metal envelope wit ? a ligh�­ Charles C Thomas, 1945.

weight, non-magnetic metal (alummum, ti­ 2. Medical X-Ray and Gamma-Ray Protection for
Energies up to 10 MeV. W ashington, DC, Na­
tanium, or stainless steel) input window. .
tional Council on Rad1atwn Protection and Meas­
These big tubes usually allow triple-field urements, 1968, Report No. 33.
3. Chamberlain, W. E.: Fluoroscopes and fluoros-
imaging, such as a 6-in., 9-in., and 14-in.
copy. Radiology, 38:383, 1942.
mode with a 14-in. diameter tube. The . .
4. Sturm, R.E., and Morgan, R.H.: Screen mtensl­
magnification factor with the small field fication systems and their limitations. Am. J.
Roentgenol. , 62:613, 1949.
can be helpful when fluoroscopy is used to
5. Thompson, T.T.: A Practical Approach to � od­
guide delicate invasive procedures such as ern Imaging Equipment. Boston, MA, Little,
percutaneous biliary and nephrostomy Brown and Co., 1985, pp. 81-126.

13 Viewing and Recording

the Fluoroscopic Image

Before the invention of the x-ray image video camera, where it is converted into a
intensifier, attempts at displaying the fluor­ series of electrical pulses called the video
oscopic image on television were only par­ signal. This signal is transmitted through
tially successful. The large fluoroscopic a cable to the camera control unit, where
screen required an elaborate optical sys­ it is amplified and then forwarded through
tem, and suboptimal screen brightness pro­ another cable to the television monitor.
duced a weak video signal. The develop­ The monitor converts the video signal back
ment of the image intensifier solved both into the original image for direct viewing.
these problems. Its small output phosphor Before discussing the individual com­
simplifies optical coupling, and its bright ponents of a television system, we will move
image produces a strong video signal. a little ahead of ourselves and describe the
nature of the video picture. An apprecia­
CLOSED-CIRCUIT TELEVISION tion of this will make the design of both
The components of a television system the camera and monitor easier to under­
are a camera, camera control unit, and stand. The television image is similar to the
monitor (Fig. 13-1). To avoid confusion in screened print shown in Figure 13-2. It is
nomenclature, we will use the terms "tel­ made up of a mosaic of hundreds of
evision" and "video" interchangeably. thousands of tiny dots of different bright­
Fluoroscopic television systems are always ness, each contributing a minute bit to the
closed-circuit systems; that is, the video sig­ total picture. When viewed from a distance
nal is transmitted from one component to the individual dots disappear, but at close
the next through cables rather than range, or with magnification, they are
through the air, as in broadcast television. clearly visible. The dot distribution is not
A lens system or a fiberoptic system conveys random or haphazard in a television pic­

the fluoroscopic image from the output ture. Instead, the dots are arranged in a

phosphor of the image intensifier to the specific pattern along horizontal lines,
called horizontal scan lines. The number
of lines varies from one television system

Control �m to another but, in the United States, most
fluoroscopy and all commercial television

Camero Unit
systems use 525 scan lines. To avoid con­
fusion we must clarify the meaning of tel­
evision lines. When a radiologist thinks of
lines, it is usually in terms of lines per unit
Figure 13-1 Components of a television sys­ length. For example, if a grid has 80 lines,
tem the unit is "lines per inch." Television lines


Figure 13-2 Dot picture and enlargement of a screened print

have only the unit of "lines," and no unit the face plate on its way to the target. The
of length. The 525 lines in most television signal plate is a thin transparent film of
systems represent the total number in the graphite located on the inner surface of the
entire picture, regardless of its size. The face plate. It is an electrical conductor with
lines are close together in a small picture a positive potential of approximately 25 V.
tube and spread apart in a large tube, but The vidicon target is functionally the
in both the total number is the same. most important element in the tube. It is a
thin film of photoconductive material, usu­
Television Camera ally antimony sulfide (Sb2S3) suspended as
The vidicon camera is the one usually globules in a mica matrix. In a plumbicon
employed for fluoroscopy, and is the only the photoconductive material is lead mon­
one that we will discuss in any detail. There oxide (PbO). Each globule is about 0.00 1
are several types of vidicons; one is the in. in diameter and is insulated from its
plumbicon, which we will mention. The neighbors and from the signal plate by the
vidicon camera is a relatively inexpensive, mica matrix. The function of the globules
compact unit. The essential parts of a vid­ is complex, but they behave like tiny ca­
icon camera are shown in Figure 13-3. The pacitors. After reviewing the other ele­
most important part is the vidicon tube, a ments in the camera, we will return to these
small electronic vacuum tube that meas­ globules and discuss them in more detail.
ures only 1 in. in diameter and 6 in. in The cathode is located at the opposite
length (larger tubes are sometimes used). end of the vidicon tube from the target and
The tube is surrounded by coils, an elec­ is heated indirectly by an internal electric
tromagnetic focusing coil, and two pairs of coil. The heating coil boils electrons from
electrostatic deflecting coils. the cathode (thermionic emission), creating
The fluoroscopic image from the image an electron cloud. These electrons are
intensifier is focused onto the target assem­ immediately formed into a beam by the
bly, which consists of three layers: ( 1) a control grid, which also initiates their ac­
glass face plate; (2) a signal plate; and (3) celeration toward the target. The cathode­
a target. The only function of the glass face heating coil assembly with the control grid
plate is to maintain the vacuum in the tube is called an "electron gun" because it shoots
(remember, an electron beam must travel electrons out of the end of the control grid.
in a vacuum). Light merely passes through As the electron beam progresses down the

I. SiGlganalss fplaceateplate
C o i l� Target 2.

n t G ��==D=ef�le c:ti=n g Coi �·§l :; _j


r-::C�o�;�ro�I� �r=di ��:= ; ������ � � •

�ElBeamectr n
System c al
Figure 13-3 Vidicon camera
Video Signal
tube, it moves beyond the influence of the This coil extends almost the entire length
control grid and into the electrostatic field of the tube and creates a constant magnetic
of the anode. The anode has a positive po­ field parallel to the long axis of the tube;
tential of approximately 250 V with respect this field keeps the beam of electrons in a
to the cathode. The electrons are acceler­ narrow bundle. The electrons progress
ated to a relatively high velocity, but they down the tube in a series of oscillating spi­
are still low energy electrons (about 250 rals, and strike the target as a finely focused
eV). The anode extends across the target beam (Fig. 13-4A).
end of the tube as a fine wire mesh. The The electron beam is steered by variable
wire mesh and signal plate form a uniform electrostatic fields produced by two pairs
decelerating field adjacent to the target. of deflecting coils that wrap around the vid­
The signal plate ( + 25 V) has a potential icon tube. Vertical deflecting coils are
of 225 V less than that of the wire mesh
( + 250 V), so electrons should flow from
the signal plate to the wire mesh. The elec­
trons from the cathode are accelerated to 1·· ···· · ··· · · · · · ·· . . . . . 1
......... . ... . . . . . . ............ ... ... ....... ........ ..................... ...................................... .... ......... ........................... .....

relatively high velocities, however, and ELECTRON BEAM"'-...

they coast through the decelerating field Spiral
like a roller coaster going uphill. By the
time they reach the target, they have been
slowed to a near standstill (they are now FOCUSING COIL/
25-eV electrons). The decelerating field
also performs a second function: it
straightens the final path o f the electron
beam so that it strikes the target perpen­

10.{07##$$�+ 1
dicularly. ELECTRON BEAM
Because the electron beam scans a fine
mosaic of photoconductive globules, it is
===:'== ______ _..

Electric Field---,
critical that the electron beam not spread
out as it goes through the tube. This is ac­ �AI]
complished by an electromagnetic focusing
coil that wraps around the vidicon tube. Figure 13-4 Focusing and deflecting coils

shown in Figure 13-4B. By alternating the When a globule absorbs light, photoelec­
voltage on the coils, the focused electron trons are emitted (Fig. 13-SB). The elec­
beam is moved up and down to scan the trons are immediately attracted to the an­
target. The other pair of coils moves the ode and removed from the tube. The
beam from side to side along a horizontal globule, having lost electrons, becomes
line. All four coils, working together, move positively charged. Since the globule is in­
the electron beam over the target in a re­ sulated from its surroundings it behaves
petitive scanning motion. like half of a tiny capacitor, and draws a
current onto the conductive signal plate.
Video Signal The current that flows onto the signal plate
Now that we have discussed the physical is ignored, or clipped, and is not recorded
makeup of the camera, we can return to (Fig. 13-5C). Similar events occur over the
the critical target end of the tube and the entire surface of the target. A brighter area
formation of the video signal (Fig. 13-SA). in the light image emits more photoelec­
trons than a dim area, and produces a
Signal Plate stronger charge on the tiny capacitors. The
result is a mosaic of charged globules that
Glass store an electrical image that is an exact
Globule Face Plate replica of the light image focused onto the
The electron beam scans the electrical
image stored on the target and fills in the
holes left by the emitted photoelectrons,
thus discharging the tiny globule capaci­
Electrons tors. After the capacitors are fully dis­
charged (no more positive charges are left),
no additional electrons can be deposited in
the globules. It was indicated earlier that
the electrons in the electron beam were re­
duced to low energy electrons before they

' entered the target. There are two reasons
Image -= ' for this. Reason one is that we want no
{ Clipped Sional
electrons to enter the target after the pos­
itive charge has been neutralized. The sec­
ond reason is that the electrons should not
have sufficient energy to produce second­
Electron ee: · ary electrons when they do enter the glob­
Beam 8- ,
ules. Of course, high energy secondary
electrons would be able to neutralize the
positive charge in other globules and de­
grade the image. Excess electrons from the
Signal Plate
scanning beam drift back to the anode and
,. ..
Target - · are removed from the tube. When the elec­
Globule trons in the scan beam neutralize the pos­
Face Plate
Mica -TI@B'illilliLJ-::
.J itive charge in the globules, the electrons
on the signal plate (Fig. 13-5D) no longer
have an electrostatic force to hold them on
the plate. They will leave the plate via the
Figure 13-5 Formation of the video signal resistor. These moving electrons form a

current flowing through a resistor, and one line at a time, so here, too, we have
therefore a voltage appears across the re­ scan lines.
sistor. This voltage, when collected for each The readout process reminds us of a fire
neutralized globule, constitutes the video brigade with water buckets. The brigade
signal (Fig. 13-5£). The globules are not passes the buckets along a line of men until
all discharged at the same time. Only a the water reaches the end of the line. This
small cluster, a dot, is discharged each in­ comparison is pretty good, but we would
stant in time. Then the electron beam have to make one small change. Instead of
moves on to the next dot in an orderly se­ the buckets being passed along, we would
quence, discharging all the globules on the have to dump the water from one bucket
target. The result is a series of video pulses, into the next. A problem comes up because
all originating from the same signal plate there is already water in all the buckets. In
but separated in time. Each pulse corre­ the fire brigade then, we would need empty
sponds to an exact location on the target. buckets between each bucket containing
Reassembling these pulses back into a vis­ water. So in the CCD, there are empty
ible image is done by the camera control spaces into which the charge may move.
unit and the television monitor. The motion of the charge is controlled by
changing the depth of the charge buckets.
Charge-Coupled Device TV Camera To collect charge, the buckets must have a
A charge-coupled device is usually writ­ deeper potential well (bucket), than the
ten and spoken of as a CCD. A CCD is a next well position. The charge may be
semiconductor device that can store charge passed to the next well position by making
in local areas and, on an appropriate signal that position have a deeper potential well
from the outside, transfer that charge to a than the bucket containing the charge. It
readout point. In conjunction with a pho­ is important to realize that electrons from
toelectric cathode, the CCD makes a very one bucket can never mix with electrons
nice TV camera (or video camera). The from another bucket.
CCD camera forms a picture in much the CCD cameras need no readout electron
same way as a vidicon: the light photons beam (or controlling coils), and can be
from the scene to be imaged are focused made shorter than the vidicon tube. In ad­
on the photoelectric c