Note: Descriptions are shown in the official language in which they were submitted.
~38572
Descr~tion
PLANAR T~ANSMISSION I,INE ATTENUATOR AND SWITCH
_
The present invention generalli~ relates to r.f.
switching devices, and more particularly to r.f. switching
devices formed by sections of transmission lines in which the
attenuation can be switched from high values, such as 60dB,
to low values, such as 5dB or OdB. The invention also
relates to the use of such transmission line switches in
multi-port networks such as microwave cross-bar switches and
~ switching d~-ices.
At the present time, microwave switching is accomplished
mostly by means of PIN diodes which function by changing
their capacitance in response to a change in bias voltage.
To act as a switch, PIN diodes must be combined with passive
elements in a circuit attached in some way to a transmission
line. The circuit must be carefully designed to provide the
desired attenuation and VSWR in the off state and in the on
state. The primary disadvantages of this type of switching
element are its relatively high complexity and narrow band-
width. The bandwidth is nherently limited because areactance change in the PIN diode is used to achieve the
switching function.
The planar transmission line attenuator and switch
according to this invention has certain similarities in con-
struction to the device disclosed in United States PatentNo. 3,975,690, issued August 17, 197~ to Paul ~. Fleming, one
of the co-inventors of the present invention. The Fleming
patent discloses a planar transmission line comprising a Gunn
effect semi-conductor on which transmission line conductors are
deposited. This device can be used to either amplify or
switch r.f. signals. When used as a switch, the
device is operated under conditions so as not to
exhibit gain. Near zero attenuation is achieved in either
of two ways. The product of carrier concentration~ N,
3~ time the electric field length, L, between the conductors
is made high enough so that at the on
state bias, the field configuration is near the threshold
113~57Z
field. This is the field at which the differential mobility,
~d~ of slope of the velocity versus field curve (see Figure 2
of the patent) is zero. A short segment of the current path
has a much higher field where ~d is also near zero. In the
short transition between these two regions, ~d is negative and
as a result the shunt conductivity per unit length, G, may be
negative for some frequencies but will not have sufficiently
large magnitude to overcome the series resistance losses in the
metallic conductors. The product N t, where t is the thick-
ness of the conductive semiconductor layer, must be kept below
about 5 1011 to prevent small signal instability leading tospontaneous oscillation. The other way of using the Gunn effect
is to allow a significant value of ~d over most of the current
path but keeping the resulting negative conductivity small by
making the thickness or carrier concentration of the conductive
semiconductor layer small. The value of conductivity, G, could
be adjusted to achieve exactly zero attenuation at a particular
operating frequency.
While the device according to the Fleming patent can pro-
vide excellent switching results in certain applications, it has
the disadvantage of high d.c. power dissipation. This becomes
a serious problem when a great many of these devices are arranged
in a matrix to form a cross-bar switch, for example.
The transmission line attenuator and switch according to
the present invention exhibits a wide range of attenuation over
a very broad bandwidth, and this is accomplished with a very
simple structure. Thus, the transmission line attenuator and
switch according to this invention has neither the complexity
or narrow bandwidth which are inherent in microwave switches us-
ing PIN diodes. ~t the same time, the transmission line attenu-
ator and switch according to the present invention is to be
distinguished from the device disposed in the Fleming patent in
that this device is characterized by making one or more of the
metallic conductors to form a Schottky barrier contact to the
semiconductor substrate. Use is made of the change in the
depletion layer width with bias voltage across the Schottky
barrier to vary the conductivity in the conductive semiconductor
layer. More specifically, the depletion layer is defined as a
layer in the semiconductor adjacent the metal contact, and this
layer contains no free electrons and is therefore said to be
depleted. If the bias voltage across the depletion layer is
-- 113~S~2
zero, then the width of the depletion layer is small. A positive
bias reduces the depletion layer width, and a negative bias in-
creases it. The carrier concentration and the length of the
shunt current path is chosen such that at zero biasl most of
the shunt current path is through the undepleted semiconductor,
while at an appropriate negative bias, most of this path is
through the depletion layer. This results in high attenuation
at a small forward bias and much lower attenuation at the
negative bias.
In one aspect the invention pertains to a planar trans-
mission line attenuator and switch including a semiconductor
substrate having a planar surface and consisting of a high
resistivity semiconductor material over which is formed a thin,
conductive semiconductor layer forming the planar surface, the
resistivity of the semiconductor material being very high com-
pared to that of the semiconductor layer so that electric cur-
rents flow only in the semiconductor layer. At least two metal-
lic conductors are deposited on the planar surface and spaced
to form a uniform gap therebetween, the gap defining a shunt
current path between the conductors through the semiconductor
layer. At least one of the metallic conductors forms a Schottky
barrier contact to the semiconductor substrate, the at least
two metallic conductors forming an r.f. transmission line having
input and output ports. Means provide for applying an r.f. sig-
nal to the input port of the transmission line to cause the r.f.
signal to propagate along the gap toward the output port of the
transmission line, and bias means is connected to the at least
one of the metallic conductors forming a Schottky barrier for
controlling the conductivity of the shunt current path.
The invention also pertains to a cross-bar switching
device of the type having a plurality of switches arranged in
a matrix such that the selective operation of the switches per-
mits the connection of any one of a plurality of input lines
to any one of a plurality of output lines, each of the switches
being of the type referred to above.
Further the invention comprehends a beta element switching
device comprising a semiconductor substrate of the above type
with five metallic conductors deposited on the planar surface
and spaced to form uniform gaps therebetween. One of the metal-
lic conductors has a generally square geometry and the otherfour of the metallic conductors has a generally trapezoidal
geometry and is symmetrically arranged about the one metallic
conductor~ The gaps between adjacent ones of the conductors
113857Z
defines shunt current paths between the adjacent conductors
through the semiconductor layer, the one metallic conductor
forming an ohmic contact to the semiconductor substrate while
the other four metallic conductors form Schottky barrier con-
tacts to the semiconductor substrate. Adjacent ends of theother four metallic conductors define input or output ports
of r.f. transmission lines formed by the five metallic con-
ductors. Means provide for applying an r.f. si~nal to at least
one input port to cause the r.f. signal to propagate along the
gap toward one of at least two ouput ports, and bias means
is connected to each of the other four of the metallic con-
ductors for selectively controlling the conductivity of the
shunt current paths.
In the drawings:
Figure 1 is a perspective view illustrating a planar
transmission line attenuator and switch according to the
present invention;
Figures 2 to 6 are schematic plan views illustrating
several alternative embodiments of the transmission line at-
tenuator and switch according to the invention; and
Figure 7 is a schematic plan view illustrating theconstruction of a ~ switching element using the transmission
line attenuator and switch according to the invention.
As shown in Figure 1, the transmission line attenuator
and switch comprises a semiconductor substrate having a planar
surface and consisting of a high resistivity semiconductor
material 10 over which is formed a thin, conductive semiconductor
layer 12. The semiconductor layer 12 has a carrier concentra-
tion, N, and mobility, ~, and a thickness t. The semiconductor
material 10 has a resistivity which is very high compared to
that of the semiconductor layer 12 so that electric currents
flow only in the semiconductor layer. Gallium arsenide is a
preferred semiconductor material, but in addition to the semi-
conductor materials mentioned in the Fleming patent, silicon
may also be used. The semiconductor layer 12 is preferably
formed by epitaxial growth on the base substrate material 10.
The transmission line illustrated in Figure 1 is a three
113857Z
conductor transmission line comprising ground planes 14 and
a center conductor 16. These metallic conductors are
deposited on the planar surface of the semiconductor sub-
strate so as to form a uniform gap 18 between the ground
plane conductors 14 and the center conductor 16. The gap
18 defines an electric field length L between the center
conductor 16 and the ground plane conductors 14. The semi-
conductor 12 has ~ shunt conductivity per unit length, G,
which is proportional to the sheet conductance, ~et,
divided by the distance, L (across which drift current flows
between the conductors). This shunt conductivity causes an
attenuation per unit length which can be large compared to
that due to the series resistance of the metallic conductors.
In order to bring the value of G to zero or close to
zero and thus greatly reduce the attenuation per unit length,
one or more of the metallic conductors 14 and 16 are made to
form a Schottky barrier contact to the semiconductor sub-
strate. As shown in Figure 2, the center conductor 16 can
be made to form the Schottky barrier, while the ground plane
conductors 14 may be made to make an ohmic contact with the
semiconductor substrate. Alternatively, the center conductor
16 may make an ohmic contact and the ground plane conductors
made to form Schottky barriers as illustrated in Figure 3. It
is also possible to make all three conductors form a Schottky
barrier contact as shown in Figure 4. Moreover, the invention
is not limited to three conductor transmission lines but may
also be realized with a slot line configuration. Such a
configuration is schematically illustrated in Figures 5 and
6. In these figures, metallic conductors 20 and 22 are
deposited on the semiconductor substrate to define a single
g~p 18. In Figure 5, the metallic conductor 20 is made to
form a Schottkv barrier, while the metallic ccnductoi- 22 foLm~
an ohmic contact with the semiconductor. On the other hand in
Figure 6, both of the metallic conductors 20 and 22 are made
to form Schottky barriers with the semiconductor substrate.
The mechanism by which the prescnt invention operates is
the change in the depletion layer width with bias volta~e
across the Schottky barrier. The depletion layer is a
layer in the semiconductor adjacent the metal contact, and
113857Z
this layer contains no free carriers and is therefore said
to be depleted. Because the depletion layer contains no free
carriers, an increase in the size of the depletion layer
results in a decrease in attenuation of an r.f. signal
propagated by the transmission line. In other words, minimum
attenuation occurs when a negative bias is applied to the
Schottky barrier contact, because there are no free carriers
to absorb the r.f. signal being propagated by the transmission
line. In contrast, when the bias voltage is slightly positive,
the depletion layer no longer exists resulting in high atten-
uation because of the r.f. signal absorption due to the high
shunt conductivity in the semiconductor layer 12.
In an experimental model with L = 5~m and Q = 1550~m,
the attenuation was switched from a high value, such as 18dB,
at a forward bias voltage of 0.6 to a low value, such as 4dB,
for a reverse bias voltage of -20 volts. Because the width
of the depletion layer can be controlled by the application
of the bias voltage to the Schottky barrier contact, the
transmission line according to this invention allows a
control of the attenuation of an r.f. signal over a very
wide range.
The lower power consumption of the transmission line
switches according to the present invention make them partic-
ularly attractive when combined into multi-port networks.
One important application is a microwave cross-bar switch
which can be used, for example, onboard a communications
satellite. Such a cross-bar switch is made by providing a
matrix of transmission line switches of the type shown in
Figure 1. Each of these transmission line switches would be
switchable by selective application of bias voltages to con-
nect corresponding row and column lines of the cross-bar
switch.
Another multi-port network which can be realized using
the advantage of the present invention is a ~ switching
element. Such an element is generally illustrated in
Figure 7 and comprises five metallic conductors deposited on
a semiconductor substrate. The cer.tral conductor 24 has a
generally s~uare geometry and forms an ohmic contact with the
substrate. Typically, this conductor is ~Jrounded. The other
113~57Z
four conductors 26, 28, 30 and 32 are symmetrically arranged
about the central conductor 24. Each of these conductors
has a generally trapezoidal shape and are sp~ced from one
another and the central conductor 24 to provide a uniform gap
width. Each of the conductor~ 26, 28, 30 and 32 is made
to form a Schottky barrier contact with the semiconductor
substrate. The gaps between adjacent ends of the conductors
26, 28, 30 and 32 define input or output ports of the r.f.
transmission lines formed by the five metallic conductors.
More specifically, the gap 34 between conductors 26 and 32
can be defined as an input port as can the gap 36 between the
conductors 28 and 30. The output ports may be defined as the
gaps 38 and 40 between the conductors 30, 32 and 26, 28,
respectively. Negative bias voltages are selectively applied
to the conductors 26, 28, 30 and 32 to control the coupling of
of input r.f. signals at input ports 34 and 36 to the output
ports 38 and 40. For example, if negative bias voltages are
applied to the conductors 26 and 30 while a slight forward
bias is applied to conductors 28 and 32, an r.f. signal
coupled to input port 34 will propagate to output port 40,
while an r.f. signal coupled to input port 36 will propagate
to output port 3~. On the other hand, if negative bias
voltages are applied to conductors 28 and 32 while a slight
forward bias voltage is applied to conductors 26 and 30, an
r.f. signal coupled to input port 34 will propagate to output
port 38, while an r.f. signal coupled to input port 36 will
propagate to output port 40.
