The x-ray image
intensifier has been used for almost fifty years to produce
sequences of x-ray images. Its origins lie in low light level
imaging, for example, night vision devices, to which an x-ray
intensifying screen has been added.
mode of operation and performance characteristics of x-ray image
intensifiers are considered on this page, under the following
Mode of Operation
The x-ray image
intensifier (XII) is generally a cylindrically-shaped device
containing a number of components housed in a vacuum. Figure 4.1
cross-section through this cylinder. X-rays emerging from the
patient enter at the input window and strike the input phosphor. The
input phosphor scintillates and light photons strike the
photocathode, which emits electrons. These electrons are accelerated
and focussed by the electron optics onto the output phosphor which
emits light. This light provides an image of the x-ray pattern that
emerged from the patient which has a substantially greater intensity
than when an intensifying screen is used on its own. This
description of its operation is summarised in figure 4.2 .
The major components
inside the XII include:
The input window in older XIIs was made from glass and their
performance suffered from x-ray scattering and absorption effects in
this material. This limitation has been overcome in modern devices
by using a relatively thin sheet (e.g. 0.25 - 0.5 mm) of aluminium
or titanium where good strength is achieved for containing the
vacuum with minimal x-ray attenuation.
The input phosphor is made from CsI, doped with Na, which is
deposited on an aluminium substrate. The CsI:Na is grown in a
structure of monocrystalline needles, each about 0.005 mm in
diameter and up to 0.5 mm long. The aluminium substrate is about 0.5
mm thick (figure 4.3 - note that dimensions in the figure are
not to scale). The input phosphor is typically 15 to 40 cm in
diameter, depending on the XII.
Both Cs and I are good absorbers at
diagnostic x-ray energies having K-edges at 36 and 33 keV,
respectively. The CsI:Na phosphor produces a blue light when x-rays
are absorbed and this light is guided along the needles in a
fibre-optic fashion (i.e. without much lateral spread) to the
A photograph of a glass-envelope XII
which has been cut in half to expose the inner side of the input
components is shown in figure 4.4.
An intermediate layer (less than 0.001 mm thick) is evaporated onto
the inner surface of the CsI:Na phosphor and a photcathode (about 2
nm thick) is deposited on this layer (figure 4.3).
The intermediate layer (e.g. indium
oxide) has a high optical transmission and is used to chemically
isolate the phosphor and photocathode materials. The photocathode
typically consists of an alloy of antimony and caesium (e.g. SbCs3).
Light photons emitted by the input
phosphor are absorbed via the phototelectric effect in the
photocathode to release photoelectrons.
The vacuum is required so that the electrons can travel unimpeded -
as in the case of the x-ray tube. A voltage of 25 to 35 kV is used
to accelerate the electrons and the electon optics is used for
focussing them onto the output phosphor. A current of about 10-8
to 10-7 A results and it is the acceleration and
focussing of these electrons which gives rise to the image
Note that a cross-over point exists so
that the image at the output phosphor is inverted relative to that
at the input phosphor. Note also that the input phosphor and
photocathode are in fact curved ( i.e. not perfectly straight as
shown in figure 4.1) so as to equalise the electron path
lengths and hence minimise image distortion.
Image magnification can be achieved by
varying the voltages on the electrodes of the electron optics, so
that a 38 cm XII can also be used to image field sizes of 26 cm and
17 cm, for instance. Three discrete field sizes are typical of many
systems although XIIs with a continuous zoom feature are also
available. Image brightness decreases as the field size is reduced
when the input exposure rate is maintained constant.
Most XIIs also feature mechanisms for
establishing and maintaining the vacuum, but this aspect of their
construction is beyond the scope of the treatment here.
The output phosphor is made from ZnCdS: Ag (e.g. a P20 phosphor)
deposited on the ouput window (figure 4.5 - note that
dimensions, once again, are not to scale).
This phosphor emits a green light when it
absorbs the accelerated electrons, and is typically about 0.005 mm
thick and 25 to 35 mm in diameter.
In addition, a thin aluminium film is
placed on the inner surface of the phosphor, which serves both as
the anode and to reflect light back towards the output window - so
as to increase the output luminance and to prevent these light
photons exciting the photocathode.
A number of designs of output window exist and include a glass
window (e.g. 15 mm thick) with external anti-reflection layers, a
tinted glass window and a fibre-optic window - the objective of
these designs being to minimise light diffusion and reflections.
The resulting image is fed to an optical
system to be viewed by a cine-camera, photographic camera, video
camera or combinations of these cameras. Orthochromatic film is
needed for the film-based cameras.
consider the fate of a 50 keV x-ray photon which is totally
absorbed in the input phosphor:
- The absorption
will result in about 2,000 light photons, and about half of
these might reach the photocathode.
- If the efficiency
of the photocathode is 15%, then about 150 electrons will be
- If the
acceleration voltage is 25 kV, the efficiency of the
electron optics is 90% and each 25 keV electron releases
2,000 light photons in the output phosphor, then about
270,000 light photons will result.
- Finally, if 70%
of these are transmitted through the output window, the
outcome is a light pulse of about 200,000 photons produced
following the absorption of one 50 keV x-ray.
The XII envelope is
made from glass or non-magnetic stainless steel, and the input
window is welded to this envelope. The assembly is housed inside a
metal container which contains lead, for radiation shielding, and mu-metal,
to shield the electron optics from external magnetic fields. The
input window is typically protected by an aluminium faceplate (e.g.
0.5 mm thick) which also serves as a safety device in case of
implosion of the XII. Many systems also feature a scatter-reduction
grid mounted at the faceplate. A 15 cm XII assembly is shown in
figure 4.6, with the faceplate at the top of the photograph, and
the optical system and a video camera towards the bottom.
The gain in image brightness results from
the combined effects of image minification and the acceleration of
This results because electons from a relatively large photocathode
are focussed down to the smaller area of the output phosphor which
gives rise to an increase in the number of electrons/mm2.
The gain is given by the ratio of the areas of the input and output
phosphors and can be expressed as:
for input phosphors with diameters
between 15 and 40 cm
and an output phosphor of 2.5 cm
the minification gain is between 36
This results from the acceleration given to the electrons as they
are attracted from the photocathode to the output phosphor. It is
dependent on the applied voltage and is typically between 50 and
The overall brightness gain is the
product of the minification gain and the flux gain, i.e.
Brightness Gain = (Minification
Gain) x (Flux Gain) .
Minification Gain = 100 and
Flux Gain = 50
then the Brightness Gain = 5,000
Brightness Gains to more than
10,000 are achievable
The brightness gain is not easily
measured and serves simply to illustrate the performance of an XII.
A more readily measured parameter is the conversion factor, which
relates what the XII delivers (i.e. luminance) relative to the input
(i.e. radiation exposure), and is useful for comparing the
performance of XIIs as well as that of a given XII over time. The
output luminance is measured using a photometer, the radiation
exposure with a ionization chamber and the conversion factor is
This factor is typically 7.5 - 15 Cd m-2/µGy
s-1 and higher.
Note that the resulting image is
relatively dim. For example the luminance of a standard domestic
light bulb is about 106 Cd m-2. The green
output image therefore needs a darkened room and dark-adapted eyes
for direct viewing, or a sensitive video camera for remote viewing.
This parameter expresses the broad-area,
high contrast performance of an XII. Various scattering effects
inside the XII result in a radio-opaque object not being completely
opaque in the image. The contrast ratio can be measured by imaging a
lead disk and expressing the luminance of its image relative to that
of an open field image. For standardization purposes, the size of
the disk is typically 10% of the field size and it is placed
centrally in the field of view. It is expressed as follows:
Typical values are 20:1 to 30:1 or
The contrast is affected by factors which
- X-ray scattering
in the input window
- X-ray scattering
in the input phosphor
- Light scattering
in the input phsophor
scattering in the electron optics
- Light scattering
in the output phosphor
- Light scattering
in the output window
and these scattering effects are
collectively referred to as
veiling glare. The output
phosphor has generally been regarded as the major source of veiling
glare although recent work has indicated that the input
components may also contribute significantly.
This parameter can be assessed using a Pb
bar test pattern by determining the highest spatial frequency - in
line pairs per mm (lp/mm) - that can be resolved. Images of such a
test pattern are shown in figure 4.7, where a 23 cm XII is
operated in a 23 cm (left), a 15 cm (middle) and an 11 cm (right)
mode. The parameter is generally expressed for the centre of the
field of view, since it decreases towards the image periphery
depending on the quality of the electron optics.
The performance is dependent on the field
size and the type of imaging camera used, and the table below shows
some typical results. Note that each of the cameras degrades
resolution to some extent and that the XII itself has a lower
resolution than screen/film devices.
Field Size (cm)
100/105 mm Camera
35 mm Camera
Conventional Video System
15 - 18
1.5 - 1.3 lp/mm
23 - 25
1.0 - 0.9 lp/mm
XII images of a uniform object are
generally brighter in the centre than in the periphery due to an
unequal brightness gain in different regions of the field of view.
This effect is also called
and is illustrated in figure 4.8.
The image is of a uniform object acquired with an XII coupled to a
video camera, and the graph on the right shows the brightness
profile for a horizontal line through the centre of the image. The
vignetting effect is quite apparent.
Note that this parameter is not widely
assessed for XII systems - in contrast to nuclear medicine where
image uniformity of gamma cameras is rigorously controlled - and a
standardized measurement technique is not in widespread use. Note
also that the image data in the figure reflects the combined effects
of the spatial uniformity of the detected radiation beam, the
coupling optics and the video camera, and not solely the XII.
The final performance characteristic to
be considered is spatial distortion. All XIIs suffer from this
effect, where images do not faithfully reproduce the spatial
relationships in an object because of unequal magnification in
different regions of the field of view.
The effect is illustrated in figure
4.9, where images of a regular wire matrix acquired with a
modern (on the right) and an older XII are shown. The distortion is
typically 'pincushion' in nature - as readily seen in the image on
the left. Notice, for example, that straight lines are reasonably
straight in the centre of this image and change to curves towards
the peripheral regions. Notice, also, that image areas measured in
the centre of the field will be less than those measured in the
One approach to assessing this
characteristic is to determine the integral distortion, which is
expressed as follows:
diagonal length of the central square in the image of the matrix
diagonal length of the largest square in the image
- n: a factor to
account for the relative sizes of these squares in the object.
The distortion expressed using this
approach is 8.5% for the image on the right in figure 4.9, and 3.5%
for the image on the right.
Last updated: 26 Jan