Note: Descriptions are shown in the official language in which they were submitted.
3~
, LARGE CAPACITY, LARGE AREA
VIDEO_IMAGING SENSORS
Field of the Invention
The present invention relates to scanning image
tubes especially those employed in video cameras. It is
; directed particuLarly to those devices which benefit from
or require a sensor-target to have a large area capacity
; to perform optimally in an application such as to image
the chest or abdomen of an adult with X-rays.
-
10 Background of the Invention
In the past, video camera tubes have been desig-
nated "low velocity" or "high velocity". Low velocity
tubes have typically featured better detective quantum
efficiency and contrast than high velocity tubes. High
15 velocity tubes have, on the other hand, typically featured
better lag performance (less lag) and spatial resolution
and permitted the use of larger capacity layers in the
sensor-target than low velocity tubes. Such characteris-
tics are discussed in 1. the Electronics Engineers'
20 Handbook, Second Edition, edited by Donald G. Fink, 1982;
2. Television Camera Tubes: A Research Review by P.K.
; Weimer in Advances in Electronics & Electron Physics, Vol.
XIII, pp. 387, Academic Press, New York and London, 1960;
3. Photoelectronic Imaqing Devices, Edited by L.M.
25 Biberman and S. Nudelman, Plenum Press, New York, 1971;
i 4. Electronic Image Storage by B. Kazan and M. Knoll with
contributions by ~ittarth, Academic Press, New York and
London, 1967; 5. The High Beam Velocity Vidicon by
J. Dresner R.C.A. Review, 305, June 1961; and 6. Advances
30 in Image Pickup & Display, Ed. by B. Kazan, Volumes I
II, 1974, 1975 Academic Press, New York and London. The
desirable features of both the low velocity tube and of
the high velocity tube have not, until the present inven-
tion, been incorporated into a single tube. It should be
35 noted that the low velocity tube currently is the only
'I
~L'~Z5~3
tube used in practice due to its higher efficiency and
contrast.
Two classes of scanner image tubes are currently
in use. The first operates "without" gain and is exempli-
5 fied by the vidicon, Plumbicon, Chalnicon, Saticon,Newvicon and Silicon vidicon. The second operates "with"
gain and is used in situations where there is insufficient
light to operate "without" gain. These tubes are exempli-
fied by the image orthicon, SEC and SIT tubes. The unique
10 feature of this latter group is the incorporation of a
front end intensifier structure designed to provide image
charge multiplication. In both classes of tubes, it is
observed that there is included an imaging section and an
electron beam scanning section. It should be noted,
however, that in the "no gain" tube, the imaging section
comprises simply a disc-like layer of material. It serves
the dual function of being a photon sensor and a target on
which to store a layer of electronic charge. For this
reason, it i9 referred to herein as the sensor-target.
In a "low velocity," no gain tube, the electron
beam scans the inside surface of the sensor-target in a
raster scanning fashion. It scans adjacent parallel lines
of the sensor-target one after another. The electrons
arrive at the sensor-target surface with low energies, and
2sin particular, too low to cause any secondary electron
emission. In the process, it deposits electrons uniformly
across t'ne surface and drives it to approximately the
potential of the electron gun cathode which is usually at
or near ground. Secondly, the electron beam generates a
30 time varying video signal as it scans through the raster.
This results from modification of the charge on the sensor
target as a result of projection of the optical image onto
the input sensor surface of the image tube. The process
occurs on a successive pixel by pixel basis as electrons
35 lost on the target surface through image formation are
replaced by the scanning electron beam.
The imaging section incorporates a medium with two
functions; a sensor responsive to the incident radiation
to be imaged and an insulating target having a resistivity
40 on the order of 10l2 ohm cm. The high resistivity is
P1~
;
essential to maintaining electron charge storage and
s immobility on the inner scanned surface of the sensor-
target during a raster scanning period. On exposure of
the sensor-target to incident radiation, from an image of
5 an object, there results a flow of charge throug'n the
medium such that electrons are lost from the scanned
surface. Accordingly, electrons are lost across the sur-
face in numbers proportional to the changing intensity of
irradiation comprising the incident image. The resultant
10 electronic image is "readout" by the scanning electron
beam.
' The tubes "with gain" have a forefront intensifier
type structure which separates the sensor from the target.
As a result, the sensor changes from a vidicon photocon-
15 ductor to a photoelectron emitter.
In operation, photoelectrons generated from the
sensor's absorption of incident irradiation are accele-
rated in the tube vacuum by the intensifiers electron
optics and are imaged onto the target's outer surface.
20 The electrons strike that surface with sufficient energy
to cause charge multiplication and a loss of a propor-
tionate number of electrons stored on the targets inner
surface. As a result, the scanning electron beam must
replace more stored electrons per absorbed photon than in
25 the "no gain" tube, and the video signal is amplified.
The scanner image tubes and the prior art have
experienced various problems, especially where large tar-
get area i8 required. Conventional low velocity tubes in
such applications suffer when the high beam resistance is
30 coupled to the large capacity to result in undesirable
excessive lag. Large capacitance can result from
increasing the size of the target, increasing the mediums
dielectric constant and/or decreasing its thickness.
In general, numerous factors must be considered if
35 the scanner image tube is to perform effectively and
optimally. Besides the sensor-target's capacity and
distributed capacity, other factors include the energy of
the electron beam, its resistance and current, as well as
the modulation transfer function (MTF) of the tube's
40 imaging components. The latter depends on such factors as
3~3~
target thickness, the lateral displacement of charge
stored on the sensor-target inner surface during a raster
period, and beam current disc'narge characteristics.
Moreover, a high detective quantum efficiency
5 (DQE) of preferably 100~ is a desired feature for an opti-
mal scanner image tube. A DQE of 100~ corresponds to a
sensor having a quantum efficiency of 100%; output noise
- limited by the photon noise at the input where the image
is projected initially--i.e. photon noise, should dominate
lOother electronic sources of noise: and the MTF high enough
to assure that the dynamic range matches the imaging
requirement throughout the necessary spatial frequency
spectrum.
For large area X-ray imaging, conventional, low
15 velocity beam scanner tubes have failed to meet the needs
-I of diagnostic radiology. As an alternative approac'n, x-
ray intensifiers have been developed with diameters up to
22 inches. Such intensifiers are than optically coupled
to a conventional sized TV camera. These systems are used
20in fluoroscopy and diagnosis. They suffer from intrinsi-
; cally poor spatial resolution compared to that of the x-
ray sensor. This results from multiplication of component
MTF's~ Poor spatial resolution manifests itself medically
in two ways, i.e., in reduced diagnostic performance and
25increased dose to the patient. These deficiencies must be
overcome.
3 The operation of a "high velocity" beam video tube
differs from the "low velocity" beam in that the electrons
of the scanning beam are made to strike the inner surface
30of the sensor-target with sufficient energy to cause
secondary electron emission. In an ideal tube, the elec-
tron bombarded surface gives off secondary electrons,
which are collected by "collecting" electrodes. In this
process the scanned surface becomes positively charged
35because of the lost secondary electrons, and the surface
voltage rises until at equilibrium, the scanning beam is
essentially equivalent to the collected beam.
When the sensor is then exposed to incident light,
there is a loss of positive charge in a manner that causes
40an electronic image reproduction of the incident irradia-
tion image. The scanning beam then is able to replace the
positive charge lost througtl secondary electron emission
until the equilibrium potential i5 reached. In the
process of replaciny lost charge, it generates a corre-
sponding video signal on a pixel-by-pixel basis whicn
effectively creates the time varying video signal that is
similar to that from the "low velocity" tube.
The Iconoscope, in the United States, and its
counterpart, the Emitron, in Great Britain, were the first
video tubes to incorporate the eatures of "high velocity"
scan and the principle of charge storage.
The Iconoscope, however, suffered from the low
sensitivity and poor efficiency resulting in large part
from poor collection and redistribution effects of
15 secondary electrons.
A "high velocity" tube with a photoconductive
sensor target was demonstrated in experimental tubes in
the early 1950's. The purpose was to develop a new con-
cept which would overcome the problems of lag and limited
sensitivity. It was discovered that under certain condi-
tions of operation, the tube could be made to operate with
a high capacity target and provide improved lag per-
formance. Furthermore, the higher capacity (thinner)
targets also offered superior spatial resolution, and in
theory, were expected to provide superior quantum effici-
ency. Redistribution effects were reduced in that the
photoemitted electrons did not exist. The secondary elec-
trons generated during scanning were less of a problem in
one mode of operation and worse in the other. Shading
continued to be bothersome, but again on a reduced scale
due to better collector design. nevertheless, the
approach was dropped as the low velocity vidicon-type
tubes were irnproved to the point where they met industrial
and broadcast requirements.
However, applications exist that require the use
of large capacity targets. These cannot be carried out
with the conventional "low velocity vidicons" where target
capacity and stray capacity must be limited. The "hig'n
velocity" approach can serve as t'ne basis for a new
invention which offers the opportunity to provide a new
i7
imager able to combine the best features of the "high and
low" velocity tubes.
Diagnostic radiology, for example, requires large
area imaging with attendant sensor-target capacity and
- 5 distributed capacity both prohibitively large with "state
of the art" video tubes.
Solid state sensor panels employed in X-ray
imaging are under development but suffer from other dis-
3 advantages. Some such sensor panels provide that readout
10 be performed on the same side of the panel that is
exposed. Moreover, the panel undergoes a cycling of
voltages to respectively effectuate charging, reading,
writing, erasing, and recharging. Accordingly, the panel
must be transported from one station to another, rendering
15 rapid imaging impossible. When such a system is auto-
mated, a large mechanical transporter is incorporated to
move the panel from the exposure platform to its readout
station. Disadvantages include cost, space occupied by
the system, and the time involved to complete the process.
20 Summary of the Invention
J
The present invention improves over prior art
- scanner image tubes by overcoming various disadvantages
and shortcomings set forth hereinbefore and by incorpo-
25 rating features that provide for more optimal operation.
To achieve such ends, the present invention has as
an object reducing, if not eliminating, the negative
effects of beam resistance and capacitance in a low velo-
city image tube.
Moreover, the present invention has as another
object the realization of advantages relating to both low
velocity beams and high velocity beams in a scanner image
tube. That is, the present invention provides for good
lag performance, for good spatial resolution, and for the
35 use of large capcity layers in the sensor-target as well
as providing for a high DQE and good contrast.
It is another object to provide a scanner image
tube useful in radiology, especially in applications
7 ~2~
requiring a large area sensor--such as imaging a human
chest. In this regard, a scanner image tube is provided
that can be operated in near rçal-time or in snap shot
applications. In diagnostic radiology, the large majority
of X-ray pictures are simply "snap-shots" as exemplified
by the simple chest radiographO Most procedures in angio-
graphy require repetition rates up to 7.5 frames/sec.,
while studies of the adult heart and coronary arteries can
require real time imaging, i.e., 30 frames/sec.
: 10 Yet another object of the present invention is to
significantly reduce the effective capacitance of large
area imaging tubes by employing transparent, parallel
stripe signal electrodes arranged side by side, parallel
to the tube's scanning raster.
Still another object of the present invention i5
to reduce the beam resistance and capacitance of large
area imaging tubes.
It is still a further object to minimize the
magnitude of distributed capacity associated with pre-
amplifier noise. This capacity increases with the area ofthe sensor-target in conventional "low velocity" and "high
velocity" video tubes. By dividing the signal electrodes
into stripes, whose length may be the length of a scan
line, and whose width can encompass one or more scan
lines, the area and distributed capacity associated with
an individual preamplifier attached to a stripe can be
sharply reduced. This approach requires multiplexing or a
separate preamplifier per electrode tripe for optimal
performance with the total given by the number of raster
lines divided by the number of raster lines per stripe.
It is yet a further object of the invention to
achieve high sensitivity in a tube having a high velocity
beam. Prior high velocity beam tubes have produced
secondary electron emission that result in the back-
scattering of some of these emitted electrons back ontothe sensor during scanning, causing image quality degrada-
tion. The present invention can employ a high velocity
beam without such attendant degradation.
It is still yet another object to optimize sensor
quantum efficiency by permitting a thin layer of, for
i
i
. 8
3 example, crystalline antimony trisulfide to be used as a
sensitive sensor--the high capacity and high lag related
thereto being compensated for in the invention, particu-
; larly by providing effectively a relatively low beam
resistance. Lag, it is noted, arises from the failure of
5 the beam to return t'ne surface of a target to the charged
potential after a single scan. In viewing an image,
undesirable smear results from lag.
'! It is still a further object to provide readout
, and recharging during beam scanning and to provide many
s 10 electron5 for recharging than are required.
; It is yet a further object to provide, in various
embodiments of the invention, charge multiplication or
gain as needed using properties of the source of the
scanning beam and the target. The invention thereby
15 directs a more copious flow of electrons to effect read-
out and recharge of the target than is produced by the
source.
In one embodiment, these and other objects are
achieved by a scanner image sensor target including a
20 first electrode; a second grounded electrode, said first
- electrode being positive relative thereto; a first solid
state layer; a second layer; and means for a raster
scanning beam of irradiation; w'nerein said first layer is
sandwiched between said first electrode and said second
25 layer; and wherein said second layer is sandwiched between
said first layer and said second electroae, and wherein
said raster scanning means directs irradiation beams into
said second layer through said second electrode, said
second layer generating electrons which are conveyed
30 toward the interface with the first layer for storage
thereat; the electrons stored at the interface forming an
electron layer displaced from said second electrode, the
electrons of said electron layer being combinable with
holes generated in said first layer in response to imaging
35 irradiation passing through said first electrode into said
first layer; the combining of holes from said first layer
witn electrons from the said second layer at the interface
forming an electronic image thereat. Accordingly, an
imaging pattern of irradiation in a given spectral ban
33~7
g
strikes the first layer in the image section whic'n gene-
' rates free electrons free electrons and holes. The holes,
in a pattern analogous to the pattern of irradiation,
drift toward the interface under the influence forming an
5 electronic image thereat. Accordingly, the free electrons
drift away from the interface toward the first electrode.
The holes combine with electrons at t'ne interface to form
an electronic image thereat. Scanning with the beam
i results in readout and recharging at the same time.
Hence, it is an object to permit image acquisition
and processing simultaneously or consecutively, as
desired, at a single station. This structure avoids the
negative effects of charge redistribution from both
photoemission and secondary electron emission which doomed
15 the success of the Iconoscope for such uses.
It is noted that the two electrodes represent
plates of a capacitor and the first layer and the second
layer from the dielectric therebetween. The second layer
^; is subject to local charge multiplication where a scanning
20 beam of electrons strikes and generates more electrons (or
electron-hole pairs). Hsnce, a beam striking the second
layer can produce almost a s'nort circuit between the
second electrode and the interface along a path determined
by the beam--resulting in a low beam resistance. It is
25 thus an object of the invention to increase tube per-
formance and adaptability by reducing beam resistance
s alone and in conjunction with reducing capacitance by the
utilization of striped electrodes.
It is to be noted that the combination of layers
30 and electrodes as described above, creates a structurewhich permits electron charge storage required for image
formation to coccur at the interface of the two layers.
This location is displaced from its position in a conven-
tional video tube, and is intrinsic to this invention.
35 Accordingly, the structure which provides the Displaced
Electron Layer in the Sensor-Target will be referred to as
DELST.
Moreover, it is observed that the beam can cause
sufficient charge multiplication in the second layer along
40 a path between the second electrode and a pixel at the
to
interface--to provide practically a short circuit there-
between. It is, therefore, an object to permit an
increase in the capacity of the sensor target by effectu-
ating charge multiplication up to breakdown in the second
5 layer in response to the irradiating thereof by the beam.
This mechanism is designed to permit use of the high area
capacity sensor-target with minimal lag.
It is a further object of t'ne invention to provide
a scanner image device that can be scanned either by a
10 laser or by a high velocity electron beam, eac'n beam
passing through a beam section electrode and into a layer
of selected material. The selected material is responsive
to the beam irradiation and generates electrons that
charge or rec'narge an electron layer displaced from the
15 beam section electrode. PreEerably, the selected material
is subject to local avalanche breakdown triggered by the
beam on a pixel-by-pixel basis. one breakdown provides a
substantial short circuit through a pixel across the layer
- of selected material. It ensures an adequate supply of
20 electrons for all applications.
It is still a further object of the invention to
reduce or avoid the need for charge multiplication in
those applications where the laser beam scanner provides
sufficient generation of charge for storage and signal
25 readout.
It is yet a furt'ner object of the invention, when
using a scanning laser beam to avoid the vacuum require-
ment of the electron beam scanning tube, by system design
whereby the sensor-target is made as a stand-alone compo-
30 nent, as is the laser beam scanner.
It is still yet another object, in variousembodiments, to provide gain to charges approaching the
interface from both the image section and the beam section
of the scanner image sensor-target of the invention to
35 increase the signal-to-noise ratio and enhance other
characteristics of the tube.
It is still a further object of the invention to
minimize the problem of large area capacity relative to
associated electronic noise by selective electrode geo-
40 metrical configurations.
~2~ 3~
. 11
It is yet a further object to provide a relativelylow-cost scanner image sensor-target employable in X-ray
environments and in contexts ranging from high energy
applications of particle and photon radiology to low
5 energy uses in the visible and infrared spectral regions.
It is furthermore an object to incorporate suffi-
cient photoconductive gain in a sensor target to increase
the video signal level in a video imaging system. Such
increases can be used in low light level applications and
10 in overcoming electronic noise problems.
It is a principal object to minimize inter-
electrode, or shunting capacity. In a super high velocity
tube (in the kilovolt range) using video display type
electron-optics, the conventional field mesh and suppres-
15 sor mesh of low velocity beam operation are omitted. In alaser-scanner device, all the electrodes of a typical
video tube are eliminated.
It is another principal objective to reduce the
requirements on the sensor thereby to avoid the limita-
20 tions of space charge limited performance. This ismanaged by requiring the sensor to detect incident radia-
tion and if necessary provide gain, while charge storage
at the interface is required of the target and not of the
sensor.
25 Brief Description of the Drawings
Figure l is a side-view illustration of a first
embodiment of a scanner image tube employed in the
invention.
Figure 2 is a front view illustration of a sensor-
30 target employed in the embodiment of Figure l.
Figure 3 is an illustration of stripe electrodeseach connected to respective circuitry in accordance with
the invention.
Figure 4 is an illustration showing a plurality of
35 stripe electrodes being scanned by a fan beam according to
the invention.
~2~
12
it Figure 5a is an illustration showing reduction in
positive charge along a stripe electrode when exposed to
input radiation. Figure 5b is an illustration showing
secondary electrons, generated during readout, being drawn
5 away from the stripe which has undergone a reduction in
positive charge.
Figure 6 is an illu.stration of an embodiment of
the invention including a displaced electron layer sensor-
target (DELST).
Figures 7, 8 and 9 are illustrations showing the
effects of impinging input irradiation and a scanning beam
on a DELST. Figure 7 shows the beam striking one surface
of a sensor-target, and Figure 8 shows radiation impinging
v the opposite surface of the sensor-target. Figure 9 shows
15 readout.
Figure 10 is an illustration of a modification to
the DELST shown in Figure 6.
Figure 11 is an illustration of an embodiment of
the invention including a DELST and an optical scanning
20 beam.
Figure 12 is an illustration of a prioximity
focussed intensifier device employed in the invention.
Figure 12a illustrates a generalized arrangement
that utilizes the channel multiplier of Figure 12.
-I 25 Figures 13 and 14 are illustrations showing
embodiments of the invention including channel
v multipliers.
Description of the Invention
The essence of the invention is to provide a large
30 capacity video type imager that is able to function in a
manner that overcomes the customary limitations imposed by
large target and large distributed capacities.
One version mazes use of parallel stripe elec-
trodes spread across the sensor-target surface to offset
35 large distributed capacity and in a further preferred
embodiment, a "high velocity" beam is used in overcoming
the problem of large target capacity. The stripe elec-
13
trodes are particularly important when the cross-sectional
area of the sensor is required to be large, as in the case
of diagnostic radiology. The invention offers Ereedom of
design even in the case of small dimension sensor applica-
tions of using 'nigh dielectric constant materials for the
A sensor-target to increase target capacity. The advantage
derived is in the target-sensor being able to store a
larger charge, and thereby offer the opportunity to
increase the dynamic range of the imager.
The combination of a scanning beam with stripe
electrodes provides the means for achieving maximum
spatial resolution for the large dimension sensor, since
readout occurs at the primary sensor-target. It avoids
the intrinsic limitations of conventional systems used,
for example, in diagnostic radiology. There, X-ray
intensifiers are designed with a large degree of demagni-
fication, and then have their output optically coupled to
a video camera tube. The result is a substantial loss of
resolution to the extent where toe system is limited
principally to fluroscopy and to intravenous angiography.
A purpose of this invention is to provide a large area X-
ray video imager able to meet the needs of diagnosis as
distinct frorn fluoroscopy, for most of the procedures used
in the practice of diagnostic radiology.
Another version uses super "high velocity" elec-
trons (kilovolt energies) and video display type electron-
optics. This type of operation eliminates much of the
usual interelectrode capacitance and minimizes the total
distributed or shunt capacity. It effectively reduces
"gain" requirements on the sensor and also reduces, if not
eliminates, the need for stripe electrodes. Furthermore,
the effective RC time constant is reduceable by offsetting
large target capacitance with a dramatic decrease in beam
resistance.
Still, another version uses a laser scanner to
generate photoelectrons in a semiconducting target for
charging and recharging a storage surface. It effectively
eliminates interelectrode capacity, provides reauced beam
impedance, and permits operation with a large target
capacity.
14 3~
The "super high velocity" scanning electron beam
and the laser scanner versions use a basic two layer
structure, comprising a sensor layer adjacent to a target
layer. In one mode of operation, the target layer must
5 have the high resistivity associated with video charge
storage as exemplified by SEC and EBIC targets in conven-
tional "low velocity" scan tubes. Positive charge created
during the scanning process is stored on the target sur-
face adjacent to the inner surface of the sensor. The
10 sensor layer must absorb incident radiation and convert it
into representative charge which is transported to the
interface between the sensor and target, where it dis-
charges proportionately, the beam induced stored charge.
15 Since the stored charge layer is displaced from the outer
(scanned) surface where it resides in a typical video tube
operation to the inner surface at the interface, the
structure is referred to as a "Displaced F,lectron Layer
Sensor-Target," abbreviated as DELST.
In the DELST structure, there is opportunity for
sensor photoconductive gain. This is possible because the
conditions of "space charge limited" performance can-be
avoided in the sensor when it does not have to sustain
stored charge.
1. Operation with "Low" and "High" Velocity
Video Tubes
Referring to Figure 1, a first embodiment scanner
image tube 100 according to the invention is illustrated.
The tube 100 has essentially the same structure for opera-
30 tion as a "low velocity" or a "high velocity" vidicon.
The front end comprises a supporting member 102 on which
is deposited a conducting layer electrode 104 and a photo-
conducting layer 106. When the incident radiation 130 is
light, the front end member 102 is glass and 104 is a
35 light transparent, electrically conducting layer suitable
for applying a bias voltage from source 122 to the outer
surface of the photoconductor 106. The photoconductive
layer 106 is selected to be responsive to the light. When
the incident radiation is X-ray, the structure 102 can be
metal known to be transparent to the radiation. In the
simplest configuration, the metal combines the role of the
supporting member 102 and the transparent electrode 104.
The p'notoconductor 106 is then selected to be responsive
5 to X-rays. The materials sensitive to the other known
types of radiaton (gamma, alpha, ionic, cosmic, beta,
neutrons, etc.) are well known so that the tube may be
made responsive to substantially, if not all, of the
various types of radiant energy. The sensor-target 106
10 comprises an efficient absorber of X-ray radiaton, CsI,
which efficiently converts the incident X-ray into light
emission. The light, in turn, is absorbed by an adjacent
layer of light sensor photoconductor with a transparent
conducting electrode at the interface between the light
emitter and the photoconductor to apply the bias voltage
for the photoconductor.
The electrode 104 is connected to a conventional
video readout circuit 120 that includes a voltage source
122 which biases the electrode 104 relative to ground and
a capacitor 126 which carries the video signal to a
preamp. The resistance 124 serves as a load resistor to
develop the video signal.
Also shown in Figure 1 is an electron gun cathode
110 which provides electrons to form an electron beam 114
and which is directed by electron-optics 116 to scan a
raster on the inner surface of the photoconductor 106.
Depending upon the bias voltages applied through Vt to
electrode 104, Vc to the cathoae 110 and that applied to
the grid 118, the tube can operate either as a "low
velocity" or a "high velocity" vidicon.
The biases for "low velocity" operation are
selected so that the energy of the beam electrons is too
low when arriving at the inner-scanned surface of the
photoconductive layer 106 to cause any secondary electron
emission. Typical for a vidicon would be the electrode
bias 122 set at 30 volts and the gun cathode bias Vc set
at ground. The grid 118 is called the "decelerating grid"
and has the function of slowiny down the electrons so that
their arrival of 106 will be with energies too low to
cause any secondary electron emission. Accordingly, its
16 7
voltage is positive with respect to ground, and typically
about lOOOV.
In operation, electrons deposited on the inner
surface of 106 cause its voltage to become essentially the
same as the cathode described above as ground.
The biases for "high velocity" operation are
selected to ensure that tube performance is governed by
secondary electron emission. An example of bias settings
used by Dresner (see reference) is a bias voltage of
Vt=~300 volts for electrode 104, the electron gun cathode
110 set at Vc = volts (ground) and the grid 118 (Vg)
set at 320 volts. This is designed so that the beam 114
electrons have sufficient energy to cause secondary elec-
trons to be emitted from the scanned surface of the photo-
15 conductive layer 106. They are collected by the grid 118,and accordingly, it is called the "collector" grid.
In operation, the scanned surface of layer 106
acquires a positive equilibrium potential such that the
net target current equals zero. The equilibrium voltage
20 VO is determined by the collector voltage Vg and lies a
- few volts above it. For a fixed Vg, the potential across
layer 106 depends upon Vt applied to electrode 104.
Accordingly, the photocurrent can be made to flow in
either direction according to whether Vt is larger or
25 smaller than the equilibrium potential VO.
In dealing with low and high velocity vidicon
performance, particularly where large values of target and
distributed capacities apply, it is prerequisite that
their operation must be properly modified for optimum
30 performance. The case of X-ray imaging for diagnostic
radiology is compared with conventional light imaging to
point out the difficulties in operation, and the new
device conception to overcome these difficulties.
Reference to Figure 2 shows the circular sensor-
35 target 206 identified in Figure 1 as 106. Traced on thistarget is a raster, whose vertical and horizontal dimen-
sions, are taken equal (a square) for simplicity. The
typical sensor-target capacity for a small video tube
using a low velocity vidicon target using antimony tri-
40 sulfide is on the order of 1000-4000 picofarads, although
-
,f~63~
17
special tubes can be found with larger values. Assume for
simplicity that a l"xl" raster is used and that t'ne target
capacity is 1000 picofarads. If this tube were enlarged
to provide a 16"x16" raster to meet the needs of diag-
5 nostic radiology, the capacity would grow in proportion tothe increased area and be as large as 256,000 picofarads.
Correspondingly, the distributed capacity can be
expected to grow, at worst, in a manner proportional to
the increased area. The distributed capacity of the small
10 size "low velocity" tube is in the range of 2-25 pico-
farads, depending upon the manufacturer. Assuming a
favorable value of 5 picofarads for the small tube, the
large area tube could have a distributed capacity of about
1,280 pico~arads.
The implication of these large capacitance values
can be discerned by the associated capacitive lag and the
preamplifier noise. The scanning beam, instead of acting
as a constant current source, acts as a resistance of the
order of 107 ohms. Assuming a target capacity of 1000
picofarads for the small tube, the RC time constant
20 becomes 107 x 1000 x 10 12 = .01 secs. Since the raster
period at 30 frames/sec = .03 secs., the RC time constant
is clearly usable.
However, for the large diameter tube, the capacity
grows to 256,0Q0 picofarads and the RC time constant to
2.56 secs. this value causes the tube to be excessively
laggy for use in conventional video imaging and even in X-
ray imaging using snap-shot operation.
The distributed (interelectrode-shunt) capacity is
related to preamplifier noise through the expression
iN (PREAMP) = 4kT. 4~ .Re.Cd2.~f3 (1)
=1.38xlO 23 Joules/kO
where k = Planck-Boltzmann Constant
t = Absolute Temperature = 300
Re = Equivalent Noise Resistance of
18 ~2~ 3~
First Stage Preamplifier Resistance
Cd = Distributed or Stray Capacity
to Ground From ~iynal Leads
coupled primarily to ELECTRODES
of = Electrical Bandwidth
Assume Re = 40 ohms
Cd = 5 picoforads for the small tube
af = 107 hertz
which leads to iN PREAMP)= .264 na.
10 When Cd = 1280 picofarads for the large tube is used,
the preamplifier noise current grows by a factor of 256
and the preamp noise becomes 67.6 na. A "state of the
art" preamplifier has been reported for the small sized,
low velocity tube with a current noise as low as 0.5 na.
15 Good preamplifiers are available at about 1 na. Relative
to the 1 na value, the large area tube illustrates a value
almost 70 times larger and a corresponding reduction in
the signal to noise ratio.
Clearly, the excessive lag and preamp~fier noise
20 must be eliminated for video tubes to be applied, for
example, to diagnostic radiology.
The "'nigh velocity" team reduces the effect of
target capacity on lag. The reason can be explained in
terms of tube conductance. For the "low velocity" beam,
25 where the sensor-target voltage is small, the current
flowing into the target (It) is given by
bVt
It = a exp (2)
where b = e/kT
T = effective temperature of the cathode
responsible for the energy distribution of
electro-ns in the beam. It is in excess of
1000K.
e = electronic charge
k = the Boltzman constant
-
37
19
The requirement that capacity lag be reduced is
equivalent to requiring that the target capacity be
decreased and that the constant b increased. The latter
is equivalent to requiring a large value of the beam
5 conductance near the equilibrium potential of the scanned
surface. The beam conductance is given by
dIt = a b expbVt = bIt (3)
dVt
where It is limited since Vt is small in "low
velocity" operation
b can only be increased by reduction of the
effective temperature T.
; Recent new designs in electron guns have succeeded in
reducing the effective temperature T, but not to the
extent necessary for operation with the large capacities
- 15 inherent in our applications.
In the case of "high velocity" operation, Dresner
shows that the beam conductance is given by:
dIt = Ib N(v) dv (4)
dvt ~Vt v
where Ib = beam current
= secondary electron emission coefficient
of the sensor-target
N(v) = the energy distribution of the secondary
electrons produced at the sensor~target
This expression can be simplified to
25 _t = Ib N
dVt
~L~4;~3~-f'
.
where N represents the number of electrons effectively
; collected.
It is clear that this type of operation i8 not
inhibited by a low target voltage and that the beam con-
5 ductance increases linearly with beam current. The corre-
- spondingly causes a decrease in lag, where compared to the
"low velocity" case, to a first approximation the capaci-
tive lag is independent of beam current.
; The dependence of distributed capacity on the area
10 of the sensor-target is dealt with in accordance with the
present invention by using stripe electrodes as illus-
trated in Figure 3. The raster lines 302 are shown simply
as parallel horizontal lines. The shaded stripes 304
represent electrodes placed one next to the other, spaced
15 so that their separation distance is appreciably less than
the vertical dimension of a pixel. Attached to each
electrode is a preamplifier circuit exemplified by the
load resistors 324, the target bias sources t and the
coupling capacitors 326. The distributed capacity is now
20 associated with the area of each electrode stripe and the
wiring to the preamp. The length of each stripe as shown
in Figure 3 is slightly longer than the length of a raster
line. The width encompasses as many raster lines as
device capacity and design allows. This number may differ
for "low" versus "high" velocity operation because of a
"secondary electron" redistribution problem inherent to
conventional "high velocity" tubes.
Limiting the distributed capacity for each stripe
to that of the 1" x 1" area sensor-target referenced
earlier, means restricting each stripe to having the same
1 square inch area. Thus, for a length of sixteen inches,
the width must be restricted to 1/16" or 1.56 mm. A 500
line raster spread over 16 inches, would present 31.25
lines/inch, and therefore about 2 raster lines per elec-
trode stripe. For 1000 and 2000 TV line rasters, thenumber of lines per stripe would grow to 4 and eight,
respectively. With this approach, the number of preampli-
fiers depends upon the number of stripes given by the
vertical raster, height divided by a stripe width. For a
'I 21
j:
j 16" raster height, and a l/16" stripe width, the number of
preamplifiers equals 256.
A vertical set of stripes may be employed in place
of the horizontal stripes. Such an arrangement would
, 5 require the use of digital to analog conversion and
.; storage so that the picture can be reassembled at the end
of each frame.
` Using stripe electrodes to minimize distributed
capacity is achievable at the expense of adding a large
I 10 number of preamplifiers. Although, in view of modern
integrated circuit techniques such numbers are acceptable,
this number can be reduced by exposing each stripe
separately, reading out the rastar lines beneath each
stripe, and then repeating the process for successive
15 stripes. This is a procedure compatible with diagnostic
radiology where an X-ray fan beam parallel to the stripes
can be made to rotate so that each stripe is exposed and
readout in succession, as shown in Figure 4. This
approach permits replacing preamplifiers by switches or
20 multiplexing to the extent that any increase in dis-
' tributed capacity can allow. Shown in Figure 4 are the
electrode stripes 404, their separation 408, the sensor-
target shown here as one layer 406, the X-ray source 402
and the fan beam 410. Although for simplicity the fan
25 beam is shown in rotation, other approaches apply such as
translating the X-ray source and fan beam along the length
of the sensor-target, or for a fixed position fan beam and
X-ray imager translating the object. Not shown in Figure
4 is the overlaying support structure shown in Figure l as
30 102. The scanner 412 is identified as the means to charge
the inner-scanned surface of the sensor-target.
The invention described to this point can apply
equally well to the iarge area "high" or "low" velocity
tubes. The sensor-target stripe capacity and the dis-
35 tributed capacity can be made essentially the same as for
those found in conventional, sTnall dimension video tubes.
Thus, this invention demon.strates that the "low velocity"
electron beam scanner can be used for X-ray diagnostic
radiology, when coupled to this invention comprising elec-
40 trode stripes with associated preamplifiers and can be
22 7
further enhanced by using a translating fan beam and
associated switching or multiplexing circuits.
However, the high velocity beam offers the
opportunity for increased target capacity without a
5 corresponding increase in distributed capacity. This
approach permits extending the dynamic range of the video
sensor, since such range depends upon the magnitude of the
charge that can be stored on the target, and the ability
of the electron gun to read out totally that charge. The
10 "high velocity" tube offers this improvement over the "low
velocity" tube and without suffering additional lag.
Thus, for example, a porous layer of trisulfide target of
1000 picofarads could be replaced by a thin amorphous
layer of lO,OO0 picofarads permitting a corresponding
lS increase in the magnitude of the signal that can be
managed. Furthermore, the thinner layer offers improved
spatial resolution from the target with increased sensi-
tivity. The principal disadvantage of the "high velocity"
tube is in the redistribution of the secondary electrons.
20 When secondary electrons emerge from a surface, they
emerge in a range of velocities and angles according to
the expression
2md
e(Vt-vc) v COS~sin~ (6)
where: R is the distance-traveled from the point of
emission on the surface of the target to the point of
landing, if not collected by the mesh
d is the target to mesh spacing
m is the electron mass
e is the electron charge
v is the electron velocity
is the emergent angle of an electron trajectory
relative to the normal to the surface
Vt is the potential of the scanned surface of the
sensor-target
Vc is the potential of the collector
It is clear that many electrons return to the surface and
as a result diminish the quality of any electronic image
stored on the scanned surface. If the device is operated
pi
23
in darkness and equilibrium conditions prevail, then a
relatively constant secondary current reaches the collec-
tor equal to the beam current and a fairly constant
secondary flow of secondary electrons falls back on to the
5 surface of the sensor-target. In this case, the signal
I'' current is zero excepting for non-uniformities in charge
distrubution associated with shading, target surface
defects and other possible spurious signals. On exposure,
the charge distribution is modified dependent upon the
10 photoconductive response, pixel by pixel, in accordance
with the projected incident image. Thus, the potential of
each individual pixel is shifted accordingly. As each
., pixel is scanned, the number of secondary electrons
s generated varies according to each pixel potential, being
15 more or less according to the degree of potential change
experienced by the sensor-target. As the scanning re-
charges successive elements, the capacitive output signal
current is largely neutralized by the return of the
secondaries to other target areas, thus limiting the out-
20 put signal to 25~ of the level possible if all secondaries
were collected.
Another problem is associated with collector geo-
; metry which in the conventional "high velocity" tube leads
; to a serious degree of shading.
It is possible, furtl~ermore, to have secondary
emission from the collector mesh made deliberately domi-
nant over that from the target. This results in mesh con-
- trolled range of secondary electrons being reduced to a
value of R/2 and in improved contrast. Further reduction
30 in R is possible by close spacing of the collector mesh.
- The problem of redistribution effects in large
part were dealt with by operation with mesh controlled
secondary emission and close spacing between a fine mesh
and the target surface.
With the approach described in the present
invention using stripe electrodes, additional means for
minimizing redistribution effects is possible. For
example, the raster lines adjacent to three stripes can be
scanned in a manner designed to erase charge stored as a
40 result of prior imaging or redistributed secondary elec-
24 3~
, .
` trons. When the inside stripe is exposed by a fan beam,
then a new imaging charge distribution occurs along the
: lines within the stripe which results in the potential
along the lines being depressed relative to these in the
5 two surrounding stripes. Accordingly, there is a strongelectric field set-up between the lines within the center
stripe and those within the neighboring stripes. Then on
scan readout of the lines within the exposed center
stripe, the secondary electrons are drawn off either
10 toward the collector or toward the adjacent strip and few
, electrons fall back on the lines within the exposed
- stripe. As a result, the signal readout is optimized by
minimizing the dramatic signal loss associated with charge
redistribution in conventional operation of a high velo-
15 city tube. The approach is illustrated in Figure 5 with
three stripes 502, the sensor-target 504 and four scan
lines per stripe 506. Each scan line is represented by a
shaded region supporting a charge distribution. The cen-
ter stripe shows a reduced positive charge distribution
20 because of exposure in Figure 5a by X-rays 510.
Accordingly, there is an electric field set up between the
center stripe and its neighbors to draw away any secondary
electrons from the center stripe generated during read-
- out. This is shown in Figure 5b where the electron beam
25 512 scans the first of four lines with a stripe. The
trajectories of the secondary electrons are shown going to
the collector grid 508, the lines in adjacent stripes and
a few falling back on a line within the exposed stripe.
Not shown is the structure supporting the sensor-target
30 corresponding to 102 in Figure 1.
The procedure to expose and read out a full raster
involves repeating the process shown in Figure 5 in
sequence until all stripes are individually exposed and
read out.
An estimate of the total cycle time using this
procedure can be provided for diagnostic radiology.
Assume 256 stripes with each stri2e exposed for 3 msec, so
that the total time required to expose a 16 "long" surface
scanned by a translating fan beam equals .758 secs.
40 Assume also that each stripe and, therefore each line is
scanned by the electron beam three times for the cycle
required to minimize the effect of secondary election
redistribution. Let the dwell time of the beam on a pixel
be the same as for commercial television, that is about
10 7 sec. Then, for the 500, 1000 or 2000 line rasters,
respectively, readout time equals .075, .300 and 1.20
. seconds. If, to these numbers, is added the total expo-
sure time of 0.768 seconds, tne total times to acquire an
image for viewing and/or processing become 0.84, 1.07 and
10 1-97 secs respectively. nose values are vexy favorable
for most of diagnostic radiology compared to other
mechanical scanning methods being attempted today and,
furthermore, they would be accomplished with a far
` superior modulation transfer function.
Another procedure for reducing the effects of
redistruhution is to switch biases on the stripes. In
this case, the stripe to be exposed is made to be negative
compared to the others. The sequence of events is then
erase, expose and read out a stripe. During readout, the
20 secondary electrons generated are collected by the mesh
and the surrounding stripes. The magnitude of the signal
is then determined by the exposure reduced only by the few
electrons that might fall back on exposed lines.
One can readily conceive of other scan procedures
25 using more or less stripes and more or less lines per
stripe. They all fall within the essence of this inven-
tion designed to optimize the drawing away of secondary
electrons from the exposed stripe during a readout.
It is clear that the high velocity tube operating
30 with stripe electrodes offers the advantages of
1. Minimal lag
2. Optimum spatial resolution in the sensor-target
3. Increased dynamic range
4. High Detective Quantum Efficiency (DQE)
5. Electronic Noise for the large diameter tube can
be the same as for the small dimension,
conventional low velocity tube.
6. High signal levels. my generating a video signal
at the sensor-target using the combination of a
beam scan with the stripe electrodes, the
33~
26
modulation transfer function approaches that of
the sensor-target.
Finally, the technique of exposing a stripe with a
short burst of radiation followed by immediate readout
f 5 shortens the time requirement for sensor-target storage.
This offers an opportunity for wider choice of sensor-
, target materials with somewhat lower resistivities com-
3 patible with the shorter storage time required. This
could make easier the possibility of acquiring a sensor-
lo target offering photoconductive gain.
The number of stripes and associatea preamplifiers
, is based on the assumption that the distributed capacity
,3 grows in proportion to the area of the sensor target.
This is not necessarily the case in comparison with target
15 capacity which clearly grows directly proportionate to
area. With "high velocity" operation, considerably more
target capacity can be tolerated, since the price of
"increased lag" does not apply as it does for "low
velocity". The price for large distributed capacity is
e' 20 incrased electronic noise associated with the preampli-
; fier. If, for example, the distributed capacity can be
made to grow slowly with area, then the stripe widths can
be increased, the number of stripes and correspondingly
the number of preamplifiers reduced.
Gain before readout remains an important factor
in determining ultimate design of an X-ray sensitive video
camera tube. At present, the principal mechanism for
obtaining such gain is with a luminiscent layer such as
CsI, which generates about 1000 light photons per absorbed
30 X-ray photon. In theory, based on energy considerations
of the light photon versus the X-ray photon, one mig'nt
anticipate that ideal conversion would offer 15 to 20
times more light photons. The consequences would be a
similar improvement in signal and the signal to noise
35 ratio (relative to distributed capacity induced noise).
Correspondingly, an X-ray sensitive photoconductor
offering a similar range in gain from photons directly
into charge carriers would provide the same advantages in
signal and the signal to noise ratio. With such improve-
40 ment, the design can trade off more raster lines per
27
.
stripe (that is, wider stripes) and fewer preamplifiersversus some loss of image contrast due to recording an
increased number of scattered events. Channel multipliers
may also be employed to reduce the number of stripes
5 required.
- Another approach to relievlng the problem of dis-
tributed capacity is to improve the signal by adding
photoconductive gain in the sensor-target. none has been
evidence of such gain in tubes such as the Chalnicon and
10 Newvicon. However, these designs did not require high
gain for their application. Earlier attempts to incorpo-
rate gain were unsuccessful for operating under the condi-
tion of "space charge" limited currents. Recent
experience has shown the existence of high gain, high
15 resistive photoconductors suited to video applications.
These permit lifting the signal level of protons as high
as hundreds. T'neir performance will be described in
detail in the sections below related to operation with
DELST devices.
Increased signal, however, provides a problem for
electron guns in low velocity tubes, which are not
designed to provide many micoramperes of signal. "Hiqh
velocity" gun operation could be more easily adapted to
the requirements imposed by such high signal currents.
25 Nevertheless, the "overwhelming" of distributed capacity
induced noise by such large signal currents permits
increasing the width of stripes and reducing their number,
'' possibly for some applications to a single electrode.
The "high velocity" tube discussed above has been
30 described with the usual raster dimensions of a conven-
tional video tube. However, the principle of operation
applies even better to a line sensor shape, as compared to
an area sensor-raster shape. A line sensor for example,
can be matched to an X-ray fan beam of radiation. The fan
35beam and line video sensor can then be translated in
unison in a direction perpendicular to the plane defined
by the fan beam and line sensor, to provide an image
covering the area defined by the length oE the sensor and
the distance it is moved. This tec'nnique is well
40established in diagnostic radiology, but suffers in part
2~ 37
from the nature of the sensors used in that application.
Both the "high and low" velocity beam operation can
?, operate with a long sensor, whose width is determined by
resolution requirements and the number of raster lines
'I 5 required for an application
; The simplest case is a single raster scan line
alo-ng the sensor length. In the low velocity case, the
device will have low shunting capacity and target capacity
limited by the usual requirement on the RC time constant
10 and preamplifier noise. In the high velocity case, the
secondary electron redistribution effects become
negligible by proper design of the collecting mesh and the
device offers the opportunity to use a muck larger target
capacity, for superior dynamic range. Additional raster
15 lines can be managed within the capacity limits imposed by
an application. The advantages gained relate to decreased
power requirements imposed on the radiation source (as for
X-rays), and more speed in acquiring an image.
l`he line scanner also offers options in its shape.
20 It can be straig'nt or curved. The latter could have
- applications, for example, in X-ray radiology to image a
cylindrical object with a point X-ray source as in non-
destructive testing. on example in medicine is to X-ray
the breast with close proximity to the chest wall.
- 25 This approach to a line scanner applies with equal
validity to versions described below for DELST WITH A HIGH
VELOCITY SCANNING BEAM and DELST WITH A LASER SCANNING
BEAM.
2. Operation with DELS'r and a HIGH
- 3~ Velocity Scanning Beam.
A device that can offer the advantages of the
"high" and "low" velocity beam tubes without secondary
electron re-listribution effects, is the displaced electron
layer sensor-target. It san be designed to function with
35 either an electron beam or a laser beam.
Referring to Figure 6, a first embodiment dis-
placed electron layer sensor-target (~ELST) 600 according
to t'ne invention is illustrated. It includes a sandwich
;
29 ~2~
structure 602 formed of a first electrode 604, a first
layer 606, a second layer 608, and a second electrode 610.
' The first layer 606 is sandwiched between the first elec-
trode 604 and t'ne second layer 608. T'ne second layer 608
- 5 is sandwiched between the first layer 606 and the second
: electrode 6]0. Between the -first layer 606 and the second
'- layer 608 is an interface surface 612.
I, The first electrode 604 is connected to a conven-
tional readout circuit 620 that includes a source 622
10 which positively biases the first electrode 604 relative
to ground an'd a capacitor 623 which carries the vidao
signal to a preamp. The second electrode 610 is connected
to ground.
Also shown in Figure 6 are the exposure side of
15 the DELST indicated by imaging irradiation 630 directed to
the electrode 604, and the charge-read side of the DELST
indicated by a scanning beam 650 directed to electrode
~610. The operation of DELST can be decribed in the
following steps:
1. Use a scanning beam of particles or photons
to penetrate electrode 610 and cause the generation of
electrons in layer 608. Under the influence of the
'; applied electric field due to the bias 622, the electrons
flow to the interface surface where they are stopped and
25 stored.
2. Expose the DELST with imaging irradiation 630
directed at and transmitted through electrode 604, to be
absorbed in layer 606. The absorption process leads to a
conversion of the imaging irradiation into charge carriers
30 within layer 606 and under the influence of the applied
electric field to a depletion of charge carriers stored at
the interface surface 612.
3. Use the scanning beam to replace the missing
electrons at the interface 61~, and in so doing generate a
35 video signal picked up by the preamplifier 620.
Of particular importance in the DELST structure
is the addition of layer 608, whic'n makes possible over-
coming the disadvantages of low velocity tubes. This
occurs because this layer 606 permits operation with a
40 scanning beam of high velocity particles or an intense
37
..
!~
scanning beam of photons, without suffering the image
degradation of the prior art. In effect the electron
storage surface 612 is now buried between layer 606 and
608 and effectively isolated from the source of any
t 5 scanner electrons, or of unwanted secondary electron emis-
sion.
The layers 606 and 608 must have resistivities
sufficiently high to permit cnarge storage, i.e. on the
order of 1012 ohm cm, and is a requirement similar to t'nat
10 for vidicon type tubes. Various multiple layer structures
with non-ohmic heterojunctions have evolved over the years
to accommodate this requirement as found in the C'nalnicon,
Saticon and Newvicon. Such junctions can readily be
incorporated in layers 606 and 608 if desired.
Electrode 604 is transparent to the incident
radiation comprising the image of some object. If the
irradiation is light, then the electrode 604 must be
'I transparent to light. If it is X-rays, then electrode 604
!~ must be transparent to these X-rays. Similarly, for
20 energetic particles, such as alphas, betas and neutrons.
Layer 606 correspondingly must be responsivs to
the nature of the incident radiation. If the irradiation
is in the visible range, it must be a photoconductor
~3 responsive to light. Similarly, it must have a spectral
25 response matched to the spectra of the irradiation
. throughout the spectra of interest, which can range from
high energy gammas through the ultraviolet, visible and
into the near infrared. As one moves to sufficiently long
wavelengths in the infrared, this approach must be modi-
30 fied to accommodate the sensors' lower resistivity by
providing cooling as needed. Nevertheless, even infrared
responsive devices are derivable from DELST sensor-
targets. Appropriate materials for layer 606 could
include, for example, properly doped germanium or silicon.
The scanner beam 650 as shown in Figure 6 can be
derived from high velocity particles or photons. Parti-
cles would be used in combination with a layer 608
selected to provide charge multiplication. This would
permit a relatively low beam current of high velocity
40 particles to cause a large current scanning beam to flow
'
' 31 37
.,
in the layer 608, suitable for charging surface 612 and
s readout with low lag. Most commonly, the particles in
the beam would be high velocity electrons with sufficient
energy to cause the required charge multiplication in
- 5 layer 608. however, energetic beams of other particles
; such as ions and alphas could also be used for scanning.
When the scanning beam is derived from photons,
their source could be from devices such as flying spot
scanners using incoherent sources of radiation, or
10 coherent laser raster scanners. The essential requirement
is that the -flux of photons be sufficiently intense to
` generate sufficient numbers of electrons in layer 608 so
that the charging and reading out functions can be managed
properly. When using photon scanners, the layer 608 is
15 made of a photoconducting material whose spectral response
is matched to the beam radiation spectrum.
It should be noted that the roles of the electrons
and charge carriers can be reversed. Thus, if the polar-
ity of electrodes 610 and 604 are reversed then the elec-
20 trons are conducted through layer 60~ and the charge
carriers are conducted through layer 608.
An example that can serve to illustrate a DELST
structure and its operation is shown in Figures 7, 8 and 9
for imaging with lignt irradiation and scanning with a
25 high velocity electron beam.
When the scanning beam projects energetic elec-
trons through electrode 710, the layer 708 comprises a
charge multiplication layer. Accordingly, the layer 708
generates more electrons than are incident thereupon.
30 Various materials and mechanisms are known which
provide this charge multiplication effect. One mechanism
i5 referred to as electron bombardment induced conduc-
tivity (EBIC). Typical EBIC materials include semi-
conducting glass, magnesium oxide, and silicon. When such
35 materials are struck, or bombarded, by electrons of high
enough energy, electron-hole pairs are generated witch
exceed the number of incident electrons. A second
mechanism relates to seocndary electron conductivity
(SEC). Potassium Chloride (KCl) is a material related to
40 this mechanism, which is embodied in known SEC tubes.
32 3'~
:
Another type of known charge multiplier is the channel
multiplier commonly found in second generation image
- intensifier tubes. Although different from each other,
these various known mechanisms may be employed in the
5 second layer 708 to provide charge multiplication.
The various charge multipler mechanisms, it is
; noted, have been used in the image section of conventional
low velocity scanner image tubes that incorporate gain.
nose tubes amplify or intensify the input image, as
10 described above in the background section.
-I The ideal multiplier layer 708 would simply
provide a short circuit between electrode 710 and the
- interface surface 712. In Figure 7 is shown the flow of
multiplied electrons in an SEC type target, 708, which,
15 under the influence of the electric field, drift toward
the interface where they are stored. Optimally, as the
high velocity beam electrons 750 penetrate layer 708 by
; passing through electrode 710, they trigger an avalanche
of electron flow, affecting a short circuit flow toward
20 the interface 712. Each picture element (pixel) on the
interface 712 charges with electrons successively as the
beam scans through a raster with each avalanche termi-
nating as the voltage drop between 710 and 712 becomes too
small to sustain breakdown. This would ensure the maximum
number of electrons available in minimizing time for
- storage and signal readout. However, such an abundance
- might not be necessary for the dynamic range required in
most imaging applications. Accordingly, EBIC and SEC type
materials offer gain extending over a range from 2 to 3000
that might well cover the range needed in general. Figure
7 is designed to show the beginning of electon deposition
at the interface 712 as the raster scan is initiated.
Note that although layer 708 is responsive to the high
energy beam of electrons 750, it is not responsive to
input imaging irradiation 730 passing through electrode
704 and into layer 706. This scanning process results in
a uniform deposition of electrons at the interface surface
712.
- The input layer 706 is similar to the photoconduc-
; 40 tive sensor-target of the vidicon type tubes. It can have
-
33 ~Z,~3~
:C
!. any of the detailed structllres exemplified by the Vidicon,
Saticon, Chalnicon, Newvicon, and Silicon Vidicon. On
; absorption of incident imaging radiation, electrons and
:~ holes are created which drift in opposite directions due
5 to the internal electric field established by the bias
voltage 622. This is illustrated in Figure 8. There is
shown a c'narged interface 814 and the movement of holes
drifting toward the interface 812. On arrival, they
- remove stored electrons by recombination. This results in
10 the uniform distribution of electrodes at 812 being modi-
-I fied to form an electronic image reproduction of the
optical image. The electrons generated within driEt to
electrode 804 and are removed from layer 806. They do not
in this process contribute to the video signal.
Readout of the electronic image 916 is il lus-
- trated in Figure 9, where the high velocity electron beam
950, in collaboration with the multiplier layer 908, is
shown replacing the interface layer of electrons removed
in the imaging process. This is done on a pixel-by-pixel
20 basis during a raster scan, and results in a video signal
being picXed up by the preamplifier circuitry 920.
Note that throug'nout the description of this
example, the input electrode 904 is electrically conduc-
tive while being transparent to the imaging light.
25 Furthermore, the scanned electrode 910 is also elec-
trically conducting while permitting of the scanning
electrons into layer 908.
In the description of this example, the process of
imaging and readout has been described in sequence. In
30 fact, the process can be managed either in sequence or
concurrently, similar to these possibilities for conven-
tional video tubes. The forming of the electron layer may
be thought of as charging or recharging the target uni-
formly to a predefined equilibrium voltage.
Finally it is clear that electrons generated in
layer 908 are deposited on the interface surface 912.
Ideally they do not penetrate layer 906, at least to the
extent that significant video signal is lost. This is
managed by the selection of materials in layers 906 and
40 908 ana their treatment during deposition so as to form a
34 0
blocking layer of t'ne surface 912 preventing movement of
electrons from layer 908 into layer 906. This includes
the possibility of the blocking layer at 912 being formed
of materials other than those found in layers 906 and 908.
` 5 It constitutes a separate, recognizable layer formed
specifically to block the flow of electrons and optimize
their recombination with holes generated in layer 906.
Such layers are well known in the art.
The electron density pattern is read out over line
10 924 as the beam 950 scans. This is shown in Figure 9.
-s Specifically, when the beam generates electrons which are
- directe-l toward a portion (i.e. pixel) of the interface
912 where no electron-hole combination has occurred, there
is no variation detected a-t electrode 904. The voltage at
15 t'ne line 924 does not change. When the generated elec-
trons recharge a portion (or pixel) that has lost elec-
trons due to recombination, a surye of current--depending
on the level of recombination--is detected at line 924.
Further examination of the sandwich structure 902
20 reveals that the first layer 906 and the second layer 908
together represent the dielectric between two capacitor
plates--namely the two electrodes 904 and 910. Interposed
- between the two plates (i.e. the two electrodes 904 and
910) is the electronic image layer at the interface 912--
25 displaced from the second electrode 910. Between the
interface 912 and the second electrode 910 is a voltage
drop av. By selecting the second layer 908 of a material
that undergoes avalanche breakdown or other such elec-
tronic multiplication when bombarded by electrons from a
30 beam, the voltage drop av approaches zero between the
second electrode 910 and the interface 912 where breakdown
has occurred. The minimum for TV is determined by the
voltage at which avalanching or c'narge flow can no longer
be sustained.
In accordance with the invention, the electronic
breakdown is localized. Specifically, breakdown occurs
where the beam 950 causes electrons to be generated at a
particular time. Hence, as the beam 950 strikes t'ne
second layer 908, avalanche breakdown occurs along a path
40 in the second layer 908 from the second electrode 910 to
~4~
the interface 912 along which electrons are generated and
flow to the interface. Such a pat'n represents a sub-
stantial short, and the localized resistance can
momentarily approach zero.
With a low intensity, high velocity beam 950,
successive portions of the interface 912--w~ereat target
charge is or is not stored-- can be simultaneously read
out and recharged as required.
In essence, the present invention displaces the
10 image forming electron layer from the equivalent of the
second electrode 910--where it typically is positioned--to
the interface 912 and achieves charging (or recharging)
' not with electrons from a low current beam subject to high
beam resistance but rather with a larger flow of current
15 generated from layer 908 electrons that avoid the source
of conventional beam resistance.
In that beam resistance combines with sensor
; capacity to determine lag, the reduction in beam rPsis-
tance permits the sensor-target capacity to be increased
20 without adversely affecting lag. The increased sensor
- capacity may be reflectea in a larger area sensor and/or a
a thinner sensor. The larger area sensor permits use of
the present invention in a broad range of applications.
The increased thinness permits (a) extended dynamic range
25 where the sensor can support increased charge density and
(b) improved spatial resolution.
In either case, stripe width and stripe separation
may be employed in further overcoming capacitance.
With the bias polarity shown in Figures 7, 8 and
9, electron flow is always from the target layer toward
the sensor layer. Accordingly, the raster-caused, charge
storage resides on the inner surface of the sensor at the
interface, or in a special blocking layer. When charge is
stored on the sensor's surface, the sensor material must
have a high resistivity in the order of 1012 ohm-cm that
is typical for video operation. In this mode of opera-
tion, the sensor rnust verve to detect incident radiation
and store charge. Because of space charge limitations, it
cannot offer gain.
36 ~2~ l
Reversing the polarity of the bias however,
changes the situation dramatically. The high velocity
beam scan now results in charge depletion since multiplied
electrons now flow out of layer 908 through electrode 910.
-5 If the target performs as an SEC layer, for example, elec-
tron flow induced by the "high velocity" electron heam
flows away from the interface, causing the inner surface
of t'ne target to become positively charged. If the sensor
is an n type material such a CdS, the absorption of
10 radiation causes electrons to flow toward the interface
and discharge the positive surface of the interlace. In
this mode of operation, the sensor need not store charge
and can have a somewhat reduced resistivity. It now has
the opportunity to also provide gain, since space charge
15 limitations need not prevail.
Alternative, flexible schemes exist for applying
stripes and their functions. In one possible arrangement,
-biases can be applied to the stripes, with the video
-signal picked-off the opposite surface electrode. This
'20 offers the advantage of individual bias control, which can
be important for optimization in obtaining uniform
response, as well as unusual applications. Another
arrangement has stripe electrodes on each side of t'ne
DELST target, so that individual bias and preamplifier
25 stripes are paired to ensure optimum performance. In all
arrangements described above, the stripes can be placed on
-either the sensor or target surface, being required only
to be transmissive to the irradiation at their surfaces.
Referring now to Figure 10, a modification is
30 applied to the D~LST in 602. Specifically, a light
emitting X-ray sensor such as a cesium iodide layer 1000
is positioned between input X-radiation and the first
electrode 1004 for another embodiment of DELST 1002. The
cesium iodide layer 1000 converts X-ray photons into light
35 photons which pass t'nrough the first electrode 1004 to
strike the first sensor-target layer 1006. As in the
previously describe,1 electron scanner image tubes, the
second layer 1008 is a charge multiplication layer that
forms an interface region 1012 with the first layer 1006.
2~)3~
37
An electronic image forms at the interface 1012-- -
displaced fron the second electrode 1010--as described in
the previous emhodiments.
In a specific example of an electronic scanner
-5 image tube according to the invention, the electron beam
!has a comparatively low current of one microamp, and the
charge multiplicaiton layer has a gain of 200. T'ne target
is a conventional material having protions, or pixels,
thereof which discharge within 10 7 seconds. With such
parameters, the number of electrons n involved in current
flow i over a time t can be calculated from the equation:
- i = n e (7)
t
That is,
i n 1o-6 amp 6.7 x 1012 electrons/sec
e t 1.6xlO-1 amp/electrons/sec (8)
Over the dwell time td of 10-7 sec, the number of
electrons imparted per beam diameter is:
Nb = n td = 6.7 x 105 electrons (9)
With a gain of 200 in the charge multiplication layer, the
number of electrons available to charge (or recharge) the
interface is:
Nb = 6.7 X 105 X 200 = 13.4 X 107 = 1.34 X 108
g electrons/beam/diameter (10)
It is essential in reducing lag that there be many
more readout electrons available then the numher of elec-
trons required to replace those electrons which combinewith holes at the interface, i.e. electrons lost in the
charge storage surface. An estimate of the charge ratio
can be obtained from the beam diameter, the dwell time of
the beam, exposure, and gain. By way of example, the
specific example is characterized by a beam diameter Ab of
.016 mm; an input flux of radiation directed toward the
38 2Q~
.,
first electrode of 2 x 105 phontons/mm2; a sensor detec-
tive quantum efficiency (DQ~) of 50%; and a pixel area
of .02 mm2. The number of holes generated per pixel to
remove stored electrons is calculated from the expression:
5 DQE X Input Flux X Pixel area = 2 X 103 holes/pixel (11)
Thus, for perfect discharge of a pixel subject to
maximum radiation exposure, 2 X 103 electrons are
required.
one flux 2 X 105 photons/;nm2, it is noted,
10 corresponds to the flux available in a typical low light
level condition. Also, this flux corresponds to the X-ray
flux available in diagnostic radiology in an exposure of 1
mR of 50-60 KeV photons.
:; The number of electrons per pixel available for
- 15 charging Nq is then determined from:
I, .
N = N Ap = 1.34 x 108 x .02
q bg A 2X10-4
1.34 x 101 electrons/pixel (12)
The ratio of charging (or recharging) electrons to
the nu~nber of holes generated in the above exposure is
- 20 then:
1.34 x lolO = 67 x 107 (13)
2 x 103
As noted hereinbefore, when the radiation input is
X-radiation rather than light, the scanner image tube
includes a light scintillator in the form of a cesium
iodide layer. Accordingly, the effect of the cesium
iodide layer must be accounted for in the above calcula-
tions. The cesium iodide layer is known to generate on
the order of 1000 light photons for each absorbed X-ray
photon. These light photons are absorbed by the sensor,
e.g., the first layer 1006 of Figure 10, whereupon
electron-hole pairs are generated. At a quantum effici-
; 39
- ency of 100%, 1000 absorbed electron-hole pairs are gene-
; rated in response to the 1000 absorbed light protons.
The number of electrons that combine with the holes con-
veyed to t'ne interface 1012 through the first layer 1006
5 thus increases by a factor of 103. Thus, for a pixel, the
number of electrons t'nat can combine with holes equals 2 X
106. The charge:di~scharge ratio, i.e. the ratio of (a)
electrons directed to a pixel from scanning versus (b)
electrons lost through the recombination, reduces to 6.7 x
10 103.
The high charge:discharge ratio result in rapid
and total readout and recharge. It is further noted that
the ratio can be further increased by increasing the
charge density of the beam or increasing the gain in the
15 second layer. In this regard, silicon has been reported
with gains as high as 3000.
As an alternative in the X-ray application, a
suitable high resistivity photoconductive first layer 1006
- that is responsivQ directly to the X-rays may be
20 employed--thus obviating the need for the additional
cesium iodide layer 1000.
A consequence of the high charge:recharge ratio is
that the leading edge of the scanning beam is able to
effectively cause discharge. The full beam diameter i5
: 25 not necessary for recharge, and the effective MTF of the
beam's contribution to spatial resolution is improved.
3. Operation with DELST and an Optical Scanning
Beam
Figure 11 shows a further embodiment of the
30 invention. A laser scanner image DELST 1102 includes the
first layer 1106 sandwiched between the first electrode
1104 and the second layer 110~. T'ne second layer 1108 is
sandwiched between the first layer 1106 and thy second
electrode 1110. Between the first layer 1106 and the
35 second layer 1108 i5 an interface region 1112 whereat an
electronic image is formed.
i
3~7
Unlike the electron beam scanner image tubes, a
laser beam 1150 scans the second electrode 1110 which is
transparent to laser radiation.
The first layer 1106 and the second layer 1108 are
5 each pnotoconductive layers of high resistivity. The
first layer 1106 is responsive to incident radiation,
e.g., X-radiation, and does not respond to laser radia-
tion. Similarly, the second layer 1108 is responsive to
the laser input but is insensitive to the incident radia-
10 tion. Alternatively, photoconductive layers 1106 and 1108
can be selected and their thicknesses adjusted to absorb
^~ the radiation passing through their adjacent electrodes
1104 and 1110, respectively. Such layers ensure that
radiation entering a layer through an electrode blocks any
15 transmission to the opposite layer and thus each layer i
exposed to only the radiation intended. Finally, it is
also conceivable that a light blocking layer can be inter-
posed at surface 612, wile maintaining the prerequisite
- electronic properties prescribed or DELST operation.
As in the electron beam scanner image tubes, the
interface 1112 represents an electron layer displaced from
- the second electrode 1110. Electrons generated in the
second layer 608 are blocked at the interface 1112. Elec-
trons and holes are generated in the first layer 1106, the
25 electrons moving to the first electrode 1104 with the
; holes combining with electrons at the interface region
; 1112 to define a distribution of charges along the elec-
tron layer. The second layer 110~3 may provide gain, if
desired. However, the number of electrons generated by
30 the laser scanner 1150 may create a large enough
charge:discharge ratio to enable readout and recharge
following electron-hole combining at the interface 1112.
The incident radiation may be light, X-rays, or
other radiation. In each case, the first layer 1106 is
35 selected to be responsive thereto--such materials being
known in the art.
A sample laser scanner image DRLST operates as
follows with specified parameters. A lOmw laser beam
provides 105/h~ photons/sec. For = 6000 Angstroms,
40 this equals 3X1016 photons/sec. The number of photons
37
41
available at a pixel with dwell time t of 10 7 sec is:
~
nt = 3 X 1016 X 10 7 = 3 X 109 photons (14)
The number of photoelectrons created by photoconduction in
the second layer 1008 is given by :
.
ne nt X I,
where n is quantum efficiency of the layer.
For n = loo, ne = 3 X 109 electrons.
For n = 10%, ne = 3 x 10 electrons.
O
The number of electrons, ne, is compared to the
10 photon irradiation effects at the input to the imaging
sensor. A typical low light level condition is irradia-
: -tion with 105 photons/mm2. A pixel of .15mm by .lSmm
I- would be exposed to about 103 photons/mm2. assuming 100%
: quantum efficiency for the Eirst layer 1106, the ratio of
(a) electrons generated by the laser scanner 1150 to (b)
the input generated charge for combining iS:3xlo9 t3X1o6
; For 10% quantum efficiency the ratio becomes 103
3 x 105. In both cases, we have assumed conservatively
that the laser beam and pixel diameters are essentially
the same.
As an alternative, the laser scanner imager 1102
may include a cesium iodide layer (see Figure 10) to
convert X-radiation into radiation matched to the first
layer (if the first layer is not responsive to X radia-
tion). In this cue, the CsI light photon gain provides afactor of 1000 which reduces the ratio to 3 x 103 or 3 x
102 for 100% and 10~ efficiency, respectively.
By way of comparison, a conventional low velocity
electron beam operating with light and a similar pixel
size and dwell time, and a beam current of 2 micoamps
provides a charge-discharge ratio of 103. This ratio, it
is observed, is quite small compared to present invention
embodiments having light radiation input and is, in fact,
comparable to the ratio for X-ray input.
3~'
;
, 42
. .
s The charge:discharge ratio may be enhanced as
s desired by increasing the laser power, increasing the
dwell time, and/or providing gain in the second layer 1108
as in the electron beam embodiments.
The term photoconductive layer, it is noted, is
used in a generic sense to include one or more layers of
material having structures such as intrinsic, p-n, and/or
" p-i-n layers to provide a photoconductive effect with
associated properties such as high impedance, good spatial
10 resolution, and good speed of response. It applies also
j to heterojunctions from dissimiliar materials.
It is additionally noted that by employing elec-
tron beam or laser beam raster scanning, the present
invention achieves high spatial resolution and an enhanced
15 MTF. Moreover, the mixed potential electrodes combine
with the high charging electron numbers to greatly reduce
electron beam shot noise of prior art devices, 5~ nce
depleted charges are completely replaced.
~^ Note that i the bias i5 reversed, and the laser
20 scanner is trained on an n type low hole mobility semi-
is conducting target, electrons will flow away rom the
j' interface through electrode 1110 leaving the target with a
residual postive charge. When the charge is stored in the
i; target, the sensor need not face space charge limited
25 performance and can provide gain as described earlier for
I' the "high velocity" electron beam scanned DELST.
I,
4. Large Area DELST Video Imagers
The large area DELST application is typified by
diagnostic radiology, and in particular to X-ray imaging
30 of the adult chest or abdomen. The conventional X-ray
film size for such applications is 14" x 17," or 238
square inches. This large area provides a formidable
problem for imaging with a low or a high velocity image
tube. In the low velocity tube, it is necessary to over-
35 come large target and distributed capacities. In a con-
ventional high velocity tube where bias is in the order of
a hundred volts, distributed capacity is as much a prob-
lem, and in addition charge redistribution effects must be
. I,.
I:
., .
~Lf~ 2~3~
43
minimized if not eliminated. The structure of DELST is
designed to eliminate the effects of lag from target
i capacity and low sensitivity from secondary electron
redistribution.
The effect of distributed capacity can be
minimized when using "super high velocity" type electron
beam scanning (kilovolt range), which makes use of display
electron-optics. The field and suppressor meshes of low
velocity tubes, which are close to the target surface and
10 are principal sources of interelectrode capacity and thus
are a source of preamplifier noise, are eliminated. The
remaining distributed capacity related to electrodes are
associated with electrodes on the inner wall of the tube,
and as such is small. Furthermore, its increase with tube
- 15 dimensions approaches linearity with increasing sensor-
target diameter, rather than its square when proportional
to area.
; In the case of laser scanner operation, the
electrodes associated with electron optics do not exist
20 and correspondingly that source of distributed capacity
; disappears.
The use of stripe electrodes was implemented as a
solution to the problem of distributed capacity in dealing
with conventional "low" and "high" velocity tubes. This
25 approach can also be applied to the DELST structure to the
extent that capacity remains a problem in any particular
application.
Consider the large area of 16" x 16" ~400 X 400,
mm2) intended for diagnostic radiology. There are two
30 ways to provide a successful X-ray DELST imager. The
first is to generate sufficient signal gain and the second
is to minimize any DELST distributed capacity (and its
associated preamplifier noise). If, for example, the gain
can be increased by a factor of 200 or more, the signal
35 grows by a factor equivalent to the increase of noise due
to distributed capacity in going form 5 to 1000 pica-
farads. Correspondingly, the signal to the associated
preamplifier noise ratio remains the same, which is a
- design goal to provide a successful operational device.
40 Decreasing capacity can be managed to some extent by the
44 37
two layer DELST, which offers some opportunity to control
the capacity between electrodes 604 and 610 by selectiny
materials for layers 606 and 608 with optimized dielectric
constants and layer thicknesses.
,:
5. DELST with Increased Gain
There are two principal ways in which to provide
increased signal gain. The first is to use a photoconduc-
tive 606, as in Figure 6 responsive to the incident
photons or as layer 1006 in Figure 10 responsive to the
10 light from the X-ray sensor 1000, to provide substantial
photoconductive gain. The second way to provide gain is
to use an intensifier placed before the DELST structure.
For example, in the high or low velocity electron beam
types of video tubes, forefront intensifiers have been
15 incorporated in the image orthicon, image isocon, SEC and
- SIT tubes, with target gains ranging from as little as 2
or 3, to as much as 3000. The use of a DELST structure to
replace these targets requires that layer 606 in Figure 6
be responsive to imaging energetic electrons, such as
20 materials used for EBIC and SEC targets. However, fore-
front intensifiers can also include Generation I, II and
III intensifiers optically or fiber optically coupled to
the DELST target, wherein the layer 606 would be
photoconductive.
Channel multipliers incorporated into proximity
- focussed intensifiers are also applicable when coupled to
the input layer 606. Channel multipliers have also been
made to be directly responsive to incident radiation and
can avoid in some cases the need for a photoemissive
30 photosensor arranged with proximity focussing before the
input surface of the multiplier. The output surface would
traditionally be placed for proximity focussing of exiting
electrons to impinge on layer 606 through the electrode
604, where layer 606 would comprise an EBIC or an SEC
35 material.
For relatively low gain requirements, the channel
multiplier can be eliminated and the simplier, less expen-
sive one or two stage proximity focussed intensifier can
-' 'l.q~ 3~
suffice. A simple one stage device would consist of a
photoemitter placed before and in close proximity to elec-
trode 604, which would be selected to permit electrons to
pass through into ther EBIC or SEC layer 606.
The geometry of proximity focussed intensifier
r devices shown in Figure 12 is particularly advantageous
; for application to large area devices, such as required
< for diagnostic radiology. There a typical sensor would be
- a layer of CsI, on which was deposited a photoemitter such
I, 10 as CsSb. The emergent photoelectrons could then impact
electrode 1204 directly or through a channel multiplier.
- The total structure Gould comprise a metal cap 1220
I; transparent to incident X-rays 1230, whose inner surface
would support the CsI and CsSb layers 1222 and 1224
15 respectively and be proximity focussed through the vacuum
space 1226 to the DELST structure as shown in Figure 12 --
an electrode 1225 is located between the CsI and CsSb
; layers and is held at a negative potential relative to
electrode 1204. With this approach, electrode 1204 would
20 permit photoelectrons to pass through an impact layer
1206, which could be either an EBIC or an SEC layer. The
; raster scanner 1240 could be a source of a high velocity
electron beam 1242 which would pass through electrode 1210
and impact layer 1208 functioning as an electron multi-
25 plier. The interface 1212 supports the electron layer for
image reproduction as described earlier. In this arrange-
, ment, the bias is reversed to expedite electron flow
through layer 1206 to the interface.
Alternatively, the raster scanner 1240 could be a
source of laser radiation 1242, which would pass throughelectrode 1210 and be absorbed by a photoconductor in
layer 1208. Again an interface surface 1212 serves to
support the electronic imaging layer.
In arrangements with the channel multiplier
serving to provide gain of an imaging signal, the bias
must be arranged so that the first electrode 1204 is
biased negatively with respect to electrode 1210. This
ensures that the photo-emitted and multiplying electrons
drift toward the interface. Furthermore, holes generated
in a semiconducting solid state EBIC layer 1208 created by
,.
,~
46
the electron beam or laser beam from t'ne scanner, will
also drift toward the interface, as required to discharge
the imaging electrons in generating a video signal. The
mechanism differs with the SEC type of layer, where the
5 interaction at the interface only involves electron flow
discharging through the SEC layer.
The channel multiplier when operating with a
relatively low gain requirement offers the potential for
the very useful design advantage. For example, in appli-
10 cations where the input window can be rigid and strong,the channel multiplier need not be self supporting. The
input window can be used to support all device layers, and
thereby offers more flexibility in acquiring an appro-
priate multiplier structure. Multipliers can be designed
15 to function for example, without the difficult require-
ments imposed in large area applications for rigidity and
sturdiness. Since they would essentially lie against the
- sensor which in turn would be supported by the input
window, the multipliers need not be self supporting. They
20 could be optimized for spatial resolution and gain. This
could be done with far less concern for mechanical
properties the multiplier must possess as in proximity
focussing or whether it be made of metal, glass or any
other suitable substance. Fragility and rigidity simply
25become far less severe requirements with this design.
Examples of input windows that could apply here would be
metal discs as used in X-ray imaging or glass discs as
used in imaging with light.
The rear surface of the device could also provide
30a source oE structural support. In particular, w'nen using
an optical scanner wherein photons are used for the
scanning process, the rear surface need only be selected
- for optical transmission matched to the spectrum of the
scanning photons. Thus in a simple application, the rear
35window could be a disc of glass for transmission of the
light beam from a scanning laser. In such a case, it is
possible to design the device so that the front and rear
windows both provide structural support and in effect
permit the construction of a sturdy sandwich-like large
40area video sensor. Between the front and rear windows
47
= would be placed in effect three layers: In succession
from the inside surface of the front window to the inside
surface o f the rear window, they would comprise the sensor
it layer structure, the channel multiplier and the solid
, 5 state layer. They would all be contiguous to one another,
evacuated for the operation of the photoemitter and the
multiplier, and sealed to maintain proper operation of all
components. There are substantial advantages to this
design such as in sturdiness, the replacement of a mal-
10 functioning imager or a scanner without effecting the
p other and the potential for reusing components when the
imager fails.
A generalized arrangement that utilizes the
channel multiplier is shown in Figure 12a. This figure
15 illustrates the channel multiplier arranged to be
- separated from the scanned layer 1208' and the photo-
emitter 1224' by a vacuum space 1225'. The separation on
? ' either side, if sufficiently wide, permits incorporation
of electron optics as used in the Generation I image
20 intensifier, and if sufficiently thin, can function with
proximity focussing. Furthermore, the photoemissive layer
1224' -an be deposited on the input side of the channel
multiplier. Another option is to place the channel multi-
plier in contact with layer 1208'. Clearly the device can
25 function with any of these arrangements, i.e., with or
without separation on either side of the channel multi-
plier relative to the appropriate layer 1224' or 1208',
and thereby offers the opportunity to select the arrange-
ment that best fits any one application. For example, in
30 large diameter X-ray imaging, the opportunity to support
the channel multiplier against another surface reduces the
need to make the channel plate rigid and inflexible.
The options available for layers 1206 and 1208 are
desirable for maximum versatility, i.e., either could be a
35 photoconductor or a charge multiplying layer. Thus the
designer, given the nature of the incident radiation, the
cross-sectional area of the device and its imaging
requirements can seek to optimize system performance by
judicious choice of all components in Figure 12.
48
Gain achieved through the use of ,ohotoconductive
gain in layer 1006 of Figure 10 is the simplest and most
desirable method and shoula be used whenever possible. It
; reduces the number of components required for the device
5 and thereby its size and cost.
With reference to X-ray imaging, layer 1006 in
Figure 10 can be managed with a material such as ZnCdS or
even CdS, which have besn developed with high resistivi-
ties. A specific example of a CdS sputtered film
10 developed by Bell Laboratories, was reported to have in a
specific application, a gain of 750, a photoconductive
rise and decay time of 150 and 50 sec, respectively, with
dark resistivities in the range of 101-1011 ohm cm, and a
ratio of dark to light resistivity of 5 X 104. These
15 characteristics are important for DELST operation, since
even though reamining distributed capacity is reduced, any
residual capacity induced preamplifier noise can be over-
come by this magnitude of photoconductive gain. For
example, if the noise is increased by a factor of 100-200
20 and a gain of 750 were incorporated in the DELST photo-
conductor, the resultant performance would not suffer from
any distributed capacity induced preamplifier noise. For
the case of a device using display electron-optics with a
"super high velocity" scanning beam, the distributed capa-
25 city is expected to grow, at worst, approximately linearlywith diameter. Accordingly, an increase in diameter
should see a 20-fold increase in noise associated with
going from a 1 inch to a 20 inch tube. This could well be
in the order of "state of the art" preamplifier noise, so
30 that only a small gain, if any, need be included. Never-
theless, for unforseen circumstances, the order of gain
available provides solutions to any distributed capacity
problems that might arise.
In practice, gains as high as one million were
35achieved at the Bell Laboratories with CdS for resistivi-
ties in the range of 107-108 ohm cm. As stated before, in
the mode of operation for DELST where charge-storage
resides on the target, it is possible for the sensor to
have lower than customary resistivities. Since gain
40improve~ with reduced resistivities, this mode of DELST
49 7
operation becomes particularly attractive for possible low
light level applications, and for reduced lighting
requirements in conventional TV applications.
Other photoconductors and other gains are pos-
5 sible, so that the characteristics for layer 606 can betailored to match the multiplier or pnotoconductor in
layer 608.
Implications of a high-gain, sensor-target extend
to a very favorable high level of signal current. Com-
10 pared to a vidicon tube operating with hundreds ofnanoamperes, the DELST offers signal current in principal
up to hundreds of micoramperres and with more gain to
milliamperes. The result is that the preamplifiers
attached to stripe electrodes become much less sophisti-
15 cated and far less expensive. Furthermore, it permitsreduction in stripe numbers through the use of wider
; stripes and/or multiplexing. Finally, since there are no
secondary electron redistribution effects, it makes the
use of simultaneous multiple fan beam exposures easily
20 possible for diagnostic radiology, permits more scan lines
per stripe as well as wider stripes. Using multiple fans
results in reduction of the total time required to expose
the full surface of the object and sensor-target and
minimization of X-ray scatter. The number of fan beams is
25 thus determined by the sæacing required between fans to
ensure that scanner from one fan does not fall on the
sensor-target exposed by an adjacent fan; and also by the
length of the sensor-target being scanned. Thus, for a
length L for the sensor-target and a spacing S, the number
30 of fans become L/S.
A specific example illustrates the performance
possible from a DELST designed for diagnostic radiology.
Assume a worst case with the following conditions where
the distributed capacity is taken equal to a stripe
35 capacity:
Sensor-Target Layer 1006
Q E s = Quantum Efficiency = 100%
Gl Gain (Photoconductive) = 750
33~
E = Exposure 1 mR with
Average Photon Energy
Equal 60 keV = 3 x 105 photons/mm2
Gain in layer 1000 of Figure 10 CCsI) = 103
' 5 Photoconductive Layer 608 in Figure 6
-
Q.E.p = Quantum Efficiency = 100~
G3 = Gain - Unity
L = Laser Scanner Beam Power - 10 mW
T = Dwell Time/Pixel _ 107 sec.
10 NpET Photoelectrons/pixel
. within a dwell time T = 3 x 10
(see Equation 13)
DELST
A = Surface Area = 400 x 400 mm2
" x 16"
15 a = Stripe Area = 4 x 400 mm2
Cs = Stripe Capacity = 3000 picofarads
Q = Thickness between = microns
Electrodes
; p = Average Resi.stivity = 1011 ohm-cm
between Electrodes
VB = Bias Voltage = 35 Volts
Pr = Pixel Resolution = 0.1 x O.lmm
NRL = Number of Raster Lines = 2000 TVL
LD = Raster Line Desity = 5 lines/mm
25 LS = No. of Raster Lines = 20 lines
per Stripe Width
With the above conditions, DELST performance can be
predicted as follows:
1. Video Signal (Sv)
-
Iv = ExQEsxGlxG2xpr/T
s = 3 x 105x1750xlOOOx(O.lxO.1) / 10-7
= 3 x 7.5 x 1o5+2+3-2+7
51
= 2.25 x 10l6 electrons/sec
= 2.25 x 10l6 x 1.6 x lO-l9 = 3.6 ma
this current is unusually high and indicates the extent to
which gain selection is flexible.
2. Noise in Signal (RMS)
us ~ExQEsxpr GlG2/1
= 55 x 750 x 1000 x /10-7 electrons/sec
= 5-5 x 7.5 x 1ol+2+3+7 = 4.125 x 1014
electrons/sec
= 66 x 10-6 amps = 66~a
3. The Preamplifier Noise IN
associated with distributed capacity can be
calculated from Equation 1 and found to be
IN = 158.4 na
d.cap.
This noise is much less than the X-ray noise
and thus permits use of even larger DELST
capacity, which offers the opportunity to use
wider and fewer stripes. The ratio of
Ins/INd 400, and suggests for the selected
operating conditions that the stripes might
well be eliminated.
4. Power dissipated from bias Pt
Vt2 (35)2
Pt = pQ 10 xlO /40x40
= ~3.5)2Xlo2~11+3xl.6x103
1 96Xlo2-ll+3+3+l = 1.g6xlo-2 Watt
= .02 Watts
- 52
5. D~LST Exposure and Read-Out Time Conditions
a. Use a single translating fan beam with
thickness equal to a strip width.
b. Electrode stripes = 100
c. Stripe width = 4mm
d. Raster Lines per stripe width = 20
e. Number of raster lines = 2000
f. Raster area = 400 x 400 mm2
g. Pixels per line (digital) = 400
s 10 h. Dwell time/pixel - 10 7 sec.
i. Readout time/line - 4x10 5 sec.
j. Exposure time/stripe = 3 msec.
A. TIME T to "Expose and Readout"
a stripe
TStripe = 3x10 3 20x4x10 5
= 3xl0 3 + 8x10-4
= 3.8 msec
B. TraSter to Expose and Readout"
entire area of 400x400 mm2
Traster = lXTstripe = 0.38 secs.
*Note that flyback time is not included in
this calculation.
Note that time to readout a raster can be reduced
further my multiple fan beams, with and without parallel,
25 simultaneous readout into designated memory. Further,
sufficient design flexibility exists for a slower scan
readout. This latter readout reduces noise rapidly when
associated with distributed capacity since it is propor-
tional to the bandwidth raised to the 1.5 power,
30 i.e-, (~f)3/2
~2~
53
t
C. Time to "Expose and Readout" Using
Multiple Fan Beams and Simultaneous
Read-Out of Exposed Stripes
-
It is readily possible to use two or more X-ray
5 fan beams for simultaneous exposure of a corresponding
number of stripes. For example, two fan beams spaced 200
mm apart at the sensor surface, in simultaneous operation
can reduce the time to expose the surface by one-half.
Furthermore, if each exposed stripe can be readout in
parallel immediately after exposure, the readout time is
- 10 also reduced by one-half. Since each stripe can have its
own preamplifier or multiplexed to a pair of preampli-
; fiers, simultaneous readout is easy to accomplish. Sig-
- nal output from each preamplifier can then be fed to a
line in digital memory, wherein the memory is designed to
15 accept one or more lines in parallel and simultaneously.
With multiple fan beams and simultaneous readout
of exposed stripes, the time to expose and readout a
, raster is given by 0.38 seconds divided by the number of
fan beams (using the above conditions).
For two fan beams, TR.2 = .19 secs
,- For four fan beams, TR.4 = .095 secs
I, Correspondingly, the imaging rate is the reciprocal of the
above, and with four fan beams, approaches 10 images per
second.
Most of diagnostic radiology is carried out below
7 7.5 frames/sec. Only for studies of the heart and
coronary arteries is there a need for rear time imaging.
I'hus with only four fan beams, one exceeds the requirement
in speed for at least 90% of diagnostic radiology. In
30addition, because of fan beam projection and the operating
sequence, one can eliminate X-ray scattering radiation
from outside the fan beam. This involves erasure of
previously recorded scattered X-rays and accomplished by
scanning the lines in each stripe before its exposure to a
35~an beam. Since such a scan involves 0.8X10 3 sec per
37
54
stripe, the additional time required for 100 stripes
equal .08 secs.
Thus, with scatter rejection, the exposure and
readout times required for a full image are
One Fan Beam = .46 secs.
Two Fan Beams = .23 secs.
Four Fan Beams = .125 secs.
Clearly with four fan heams, the imaging, raster rate can
i approach 8 frames/second and include rejection of
1 10 scattered X-rays.
; 6. DELST With seduced Capacity
The layers 606 and 608 in Figure 6 can be thought
of as two capacitors in series. Thus the capacity between
electrode 604 and 610 is less than t'nat of the sensor-
target, the extent depending upon the thicXness and
dielectric constant of layer 608. Reduced capacity means
reduced noise associated with the preamplifier. The
extent to which this is possible depends upon the extent
of increase in bias 622 required to maintain the proper
20 internal electric fields in layers 606 and 608; the
ability to maintain a well defined electron beam of small
diameter in 608 to hold spatial resolution, maintaining
! sufficient charge storage for the desired dynamic range
and maintaining prerequisite performance as an electron
multiplier. For example, ZnCdS and/or CdS can serve asthe photoconductive layer 606, and have dielectric con-
stants that are in the range of 8.37-9.4. The multiplying
layer 608 used with an electron beam might comprise KCl
with a dielectric constant of 4.64. The combination with
thicknesses adjusted for p'notoconductive bias and multi-
plier requirements, offer the opportunity to reduce the
net capacity, while holding the density of charge in the
interface surface 612 to desired levels required for
imaging. In the case of the DELST used with a laser beam
scanner, an example of materials selection could be a
ZnCdS or CdS sensor-target of 1-2 microns thick used for
337
layer 606, to be used in conjunction with a porous anti-
mony trisulfide film as thick as ten microns or more in
the layer 608. An advantage to using KCl as an electron
multiplying layer or Sb2S3 as the photoconductor respon-
5 sive to the laser radiation is that they are relatively
- insensitive to X-rays. Accordingly, they would not be
activated by incident X-rays 630 as would be used in
diagnostic radiology. On the other hand, ZnCdS and CdS
have been used as X-ray sensors in the past. Accordingly
10 any X-rays penetrating the X-ray sensor layer of CsI 1000
in Figure 10, and absorbed in the sensor-target layer
; would only have a favorable effect in total photoresponse.
Since the layer 606 here, however, is only a few microns
thick, X-ray absorption will be small and any effect on
15 overall device performance will be small.
Unusually high target charge storage requirements
can be met with the ASOS photoconductors used by O.
! Schade, which were capable of 400 times the capacity used
in conventional vidicon type sensor-targets.
The effect of capacity reduction by adjusting the
individual capacities of layers 606 and 608, and using
electrode stripes combine to reduce any potential capa-
city problem. Nevertheless, it is conceivable for some
applications that reduced photoconductive gain in layer
25 606 and reduced signal current are desirable. This could
result in the requirement for substantially reduced
capacity beyond that possible by adjusting the dielectric
properties of layers 606 and 608, or by further reducing
stripe widths. An alternative way to achieve capacity
30 reduction is by insertion of an intensifier such as the
channel multiplier already described as a means to achieve
gain in lieu of photoconductive gain.
Reference to Figures 13 and 14 reveal two alterna-
tive approaches, both using channel multipliers. In Figure
35 13, the signal electrode is 1304 and the ground electrode
is 1310 as described heretofore. Layer 1306 is a photo-
conductor responsive to the incident radiation 1330 pas-
sing through 1304. However, layer 1308 is now a channel
multiplier placed between layer 1306 and electrode 1310.
40 A raster scanner 1340 sends a high velocity beam of elec-
~6
trons through electrode 1310 and with proper bias voltageapplied to 1304 (not shown) causes a displaced charge to
be deposited at the interface 1312. On imaging exposure,
the scanner and channel multiplier replace the lost charge
and in the process generate a video signal which is picked
up by the preamplifier 1322 through the coupling capacitor
1320. The channel plate can now be millimeters thick
versus microns thick for designs described earlier, and
correspondingly cause a drop in DELST capacity ranging
from a few hundreds to a thousand or more. The multiplier
need not operate with dramatically high gain, and the
range of 1,000 to 10,000 should prove adequate for most
cases.
Figure 14 shows the same basic scheme making use
of the channel multiplier, but incorporating a laser
raster scanner 1440 and a scanning laser beam 1442 passing
through electrode 1414 to be absorbed by a photoemitter
1410 deposited on the channel multiplier 1408. Photo-
electrons emitted from 1410 enter the channels of 1408,
are multiplied and deposited on the interface 1412. On
exposure by imaging radiation 1430, the video signal
generated by scanning is picked up by preamplifier 1422
through the coupling capacitor 1420. The photocondutor
1406 is made of a material responsive to the spectrum of
the incident radiation 1430.
The option to use the channel multiplier on the
imaging side ox the interface 1312 and 1412 for increased
electronic image gain has been described in section 5.
When used there, the capacity is reduced in the same
manner as described above. It can be reduced further by
adding the vacuum separation 1226 with proximity focussing
for example.
DELST WITH CONTROLLABLE GAIN
In the various embodiments of the DRLST structure
described heretofore, stress was placed on the potential
need for gain particularly for large area devices such as
X-ray devices and low light level applications. This gain
was a requirement on the input, imaging side of the DELST
57
to overcome the large distributed capacity induced elec-
tronic noise in the video preamplifier. Obtaining gain
was described for example using photoconductor in one
version of DELST, and a channel multiplier in another. It
should be noted that the magnitude of the gain is depen-
dent, in either case, on a voltage drop across the layer,
i.e., the photoconductor or the channel multiplier
depending upon which is being used. Accordingly, gain can
be controlled by the magnitude of the potential difference
applied across the DELST electrodes.
Applications can exist where different gains are
required to overcome different imaging conditions. For
example, diagnostic radiology requires that for certain
procedures, fluoroscopy be carried out prior to obtaining
a diagnostic quality radiograph. The ormer is managed
with a low X-ray exposure to minimize dose to the patient.
It serves the dual purposes of providing the radiologist
wtih a preliminary viewing of the body anatomy and the
positioning of the X-ray imaging apparatus. A radiograph
is then obtained with a much larger exposure, which is
essential to obtaining a diagnostic quality image.
Accordingly, fluoroscopy requires a much higher level of
gain sufficient to ensure that the resultant image is of
sufficient quality to meet its purposes. The diagnostic
exposure, on the other hand, being much larger requires
less gain in DELST to provide a suitable radiograph.
The design of the DELST structure as shown in
Figure lO, leads to a selection of resistivities for
layers 1006 and 1008. This is to ensure the appropriate
voltage drop across 1006 during exposure and 1008 during
readout of the electronic image of the interface 1012.
In effect, the potential difference applied across elec-
trodes 1004 and 1010 combined with the resistance of
layers 1006 and 1008, plus that associated with the inter-
face layer 1012, determines the voltage drop across each
of the layers and the interface. Accordingly, the poten-
tial differences across the electrodes 1004 and 1010 are
selected to provide the necessary gains in 1006 during
fluoroscopy and diagnostic imaging, and a third potential
difference for the appropriate voltage drop across 1008
~2~33~
58
during raster scanning for readout. For example, the
layer 1008 require a much larger voltage drop to provide
an optimum scan readout, than would exist during exposure
conditions set-up to optimize gain in 1006. Thus, after
exposure and the formation of the electronic image at
1012, the electrodes 1004 and 1010 potential differencP
could be switched to a higher value to accommodate the
requirements imposed by layer 1008 for optimum readout.
A substantial benefit in using the channel multi-
plier to substantially reduce capacity is in theopportunity to use fewer and wider stripe electrodes.
This leads to more scan lines per stripe, fewer preampli-
fiers and the opportunity to readout and erase scattered
events before readout within one stripe width. These
! 15 factors lead to overall improved imaging frame rates and
operational simplicity. Even real time imaging can be
designed into a DELST system given all the options avail-
able to incorporate into a system.
In the structure described heretofore, biases have
been arranged so that the first electrode is either posi-
tive or negative with respect to the second electrode
taken as ground. This results in a potential difference
across the DELST structure such that it causes charge
carriers to move in a preferred direction compatible with
a specific DELST structure. It is clear that this
description is presented for simplicity; that, in fact,
the grounded electrode could be the first electrode and
that the second electrode could be at a postive or nega-
tive potential relative to ground; that, in fact, neither
electrode need be grounded to support a preferred poten-
tial difference, and that any bias arrangement need only
be accompanied by proper electronic circuitry to transmit
the video signal to preamplifier.
Other improvements, modifications and embodiments
will become apparent to one of ordinary skill in the art
upon review of this disclosure. Such improvements, modi-
fications, and embodiments are considered to be within the
scope of this invention as defined by the following
claims.
