The present invention relates to a multi-finger (comb-form configuration) transistor, and in particular, relates to a multi-finger transistor to be used in a microwave bandwidth.
The plurality of source electrodes 6 and the plurality of drain electrodes 7 are alternatively disposed, one by one. Also, one gate finger 5 is disposed between a source electrode 6 and a drain electrode 7 which are next to each other.
In the gate section, the gate electrode pad 1 is connected to the plurality of gate fingers 5 via the gate bus bar 4.
In the source section, the source electrode pad 2 is connected to the plurality of via holes 3 via the air bridge.
The plurality of via holes 3 are grounded. The plurality of via holes 3 are connected to the plurality of source electrodes 6, respectively.
In the drain section, the drain electrode pad 8 is connected to the plurality of drain electrodes 7.
In the multi-finger FET which is configured as above, an associated capacity and a resistor in series per unit length of each gate finger 5 are symbolized by C and R, respectively. Also, a finger-length of the gate finger 5 is symbolized by Lw.
In
In this coordinate system, a voltage equation of a distance x of any gate finger 11 from the gate bus bar 10, that is the coordinate x, can be shown, with distribution constants, as below.
∂V2(x)/∂x=CR∂V(x)/∂t
A current component I(x) and a voltage component V(x) can be obtained by using an input voltage V0 inputted to the gate in a boundary condition of the gate finger 11. The boundary condition in
V(0)=V0
I(Lw)=0
In the multi-finger FET of the related art, by using a distribution constant expression, a current and a voltage on an arbitrary point of the gate finger are not uniform if an end of the finger is open. This means that device characteristics of the gate finger vary by position and that device characteristics of the multi-finger FET easily fluctuate.
In the configuration of the multi-finger FET of the related art, by using the distribution constant expression, the current and the voltage in the gate finger are not uniform. Therefore, a source inductance seen from the gate finger varies by from which position of the gate finger it is seen. As the result, a device gain characteristic is influenced; it is a subject existing in the multi-finger FET of the related art.
Also, a decrease of the FET gain is contributed by the source inductance. However, in the multi-finger FET of the related art, the via hole is connected to only one side of the source electrode. Thus, the source inductance seen from the gate finger varies in accordance with from which position of the gate finger from it is seen. In particular, an increase of the source inductance seen from an end of the gate finger causes a significant deterioration of the device gain characteristic. This is also a subject existing in the multi-finger FET of the related art.
Furthermore, if a starting point and an end point of the multi-finger are connected to one end of the gate power feeding line, the gate power feeding line becomes a closed circuit. In this case, if conditions are met, a loop oscillation occurs and the multi-finger FET may become unstable. This is also a subject existing in the multi-finger FET of the related art.
In relation with above, a description about a semiconductor device is disclosed in a first patent literature (Japanese Laid-Open Application 2000-138236). This semiconductor device is using a field effect transistor in which each of a plurality of source electrodes is disposed on a same axis and connected via a conductor; this field effect transistor has a gate electrode and a drain electrode, both of which are configured in a comb-form. This semiconductor device has via holes, each of which is configured to ground each ground electrode, respectively and correspondingly; each ground electrode is connected to a corresponding source electrode disposed on both ends of each of source electrodes. Each via hole has an elliptic hole shape.
PTL 1: Japanese Laid-Open Application 2000-138236
A subject of the present invention is to provide a multi-finger FET in which a source inductance seen from each point of a gate finger is uniform and stable.
The semiconductor device of the present invention includes a source electrode, a drain electrode, a gate electrode and a gate power feeding line. Here, the gate electrode is disposed between the source electrode and the drain electrode. The gate power feeding line is connected to both ends of the gate electrode.
In the semiconductor device of the present invention, a source inductance seen from each point of the gate finger is uniform and stable. Therefore, a higher gain of a FET is realized in a bandwidth like microwave or millimeter-wave.
The subject, the effect and the characteristics of the above invention are more clarified by exemplary embodiments in cooperation with attached drawings.
Hereinafter, exemplary embodiments of a semiconductor device of the present invention will be described with reference to attached drawings.
Here, the gate fingers 16 work as gate electrodes. The gate bus bar 15 works as gate power feeding line. Via holes 14 are grounded and work as ground sections.
The plurality of source electrodes 17 and the plurality of drain electrodes 18 are alternatively disposed, one by one. Also, one gate finger 16 is disposed between a source electrode 17 and a drain electrode 18 which are next to each other.
In the source section, two source electrode pads 13 are connected to the plurality of via holes 14. Each of the plurality of via holes 14 is grounded. In each of the plurality of source electrodes, one end is connected to one of the source electrode pads 13 and another end is connected to the other one of the source electrode pads 13.
In the gate section, the gate electrode pad 12 is connected to a middle section of the gate bus bar 15. A part of the gate bus bar 15 from the position where the gate electrode pad 12 is connected to one end will be called one end section of the gate bus bar 15. Similarly, another part of the gate bus bar 15 from the position where the gate electrode pad 12 is connected to another end will be called other end section of the gate bus bar 15. Each of the two end sections of the gate bus bar 15 are disposed along an aligned set of the plurality of source electrode 17, the plurality of drain electrode 18 and the plurality of gate fingers 16. Both ends of the plurality of gate fingers 16 are connected to the one end section and the other end section of the gate bus bar 15. Therefore, the whole area of every gate fingers 16 has a same voltage.
In the drain section, drain electrode pad 19 is connected to a first drain electrode 18. The first drain electrode 18 is connected to a first air bridge 57. The first air bridge 57 is connected to a second drain electrode 18. Here, the first air bridge 57 crosses two gate fingers 16 and one source electrode 17 which are disposed between the first drain electrode 18 and the second drain electrode 18. Similarly, the plurality of drain electrodes 18 and the plurality of air bridges 57 are alternatively connected, one by one, and, each air bridge 57 crosses two gate fingers 16 and one source electrode 17 which are disposed between two drain electrodes 18 connected to both ends of the air bridge 57.
In
In this coordinate system, a voltage equation in a distance x from the one end section of any gate finger 22, that is a coordinate x, can be shown with distribution constants, as below.
∂V2(x)/∂x=CR∂V(x)/∂t
Here, C and R show a parasitic capacitance and serial resistance by a unit length of any gate finger 5, respectively.
Also, a boundary condition can be shown as below.
V(0)=V(Lw)=V0
An input impedance of the gate finger 22 can be shown as below.
Z
in(x)=V(x)/I(x)
Its real resistance component, which is
Re[Z
in(x=0)]
shows a gate resistance. Therefore, the gate resistance can be obtained by above boundary condition.
By connecting both ends of the gate finger 22 to the gate bus bar 21, a voltage becomes uniform in a whole area of the gate finger 22; in such case, the gate resistance will be as below.
Re[Z
in(0)]=( 1/12)RLw
Incidentally, the gate resistance in a case where the one end section of the gate finger is connected to the gate bus bar 21 and the other end section is open, as same as the multi-finger FET presented as a related art, is as below.
Re[Zin(0)]=(⅓)RLw
This result shows that, in the case where both ends of the gate finger 22 are connected to the gate bus bar so that the voltage becomes uniform in a whole area of the gate finger, the gate resistance can be decreased to ¼ of the related art configuration.
Therefore, by decreasing the gate resistance, the multi-finger FET of the present invention can be obtain a higher gain.
Also, since both ends of the gate finger are connected to the gate bus bar in the multi-finger FET configuration of the present invention, the voltage is uniform in whole area of the gate finger and an influence of device characteristics variability is small.
This semiconductor device further includes two source electrode pads 25, two via holes 26, a gate electrode pad 23, a gate bus bar 24, a drain electrode pad 30 and an air bridge 58. The two source electrode pads 25, the two via holes 26, the gate electrode pad 23, the gate bus bar 24 and the drain electrode pad 30 of
As shown in
A value of the source inductance of the semiconductor device of the related art was 0.08 nH. It can be understood from the graph of
Thus, by connecting both ends of the gate finger 27 to the gate bus bar 24 and grounding the source electrode 28 via the via holes 26 which are disposed on both ends of the source electrode 28, the multi-finger FET of the present invention can obtain, in a high frequency, a higher gain characteristic than in the relate art.
The multi-finger FET of
This ladder circuit is configured by connecting the resistor 31 and the capacitor 32 with a via hole in series. This ladder circuit is connected with the gate finger 53 in series to suppress or avoid a loop oscillation.
Here, if the values of the resistor 31 and the capacitor 32 with a via hole are shown by R and C, respectively, a resonant frequency f of the ladder circuit as a parallel resonance circuit is given as below.
f=½πRC
A loop oscillation is likely to occur when the phase difference is near 180 degrees. This phase difference is determined by a combination of the lengths of the gate bus bar and the gate finger on a layout.
A loop oscillation frequency bandwidth, which is determined by a combination of lengths of the gate bus bar and the gate finger on a layout, can be avoided by providing a parallel resonance circuit. Therefore, a loop oscillation condition can be avoided in a desired operation frequency bandwidth by using this resonant frequency. Thus, a stable operation becomes possible in a closed circuit network of a gate finger of which both a starting point and an end point are connected to the gate bus bar.
In this MMIC, the bias circuit 39 is connected to the gate electrode pad 23 in the multi-finger FET of the second exemplary embodiment of the present invention. The drain electrode pad 30 of the multi-finger FET of the second exemplary embodiment of the present embodiment is connected to a first capacitor 42a and the inter-stage signal circuit 40. The inter-stage signal circuit 40 is connected to the gate electrode pad 12 in the multi-finger FET of the first exemplary embodiment of the present invention via a second capacitor 42b. The drain electrode pad 19 of the multi-finger FET of the first exemplary embodiment of the present invention is connected to a third capacitor 42c. The drain electrode pad 19 of the multi-finger FET of the first exemplary embodiment of the present invention is connected to a fourth capacitor 42d and the output matching circuit 41.
In this example, a high gain characteristic can be reached in a frequency bandwidth from the microwave band to the millimeter wave band.
In same time, an expansion to a high gain MMIC becomes possible.
The multi-finger configuration of the present invention can be expanded to a high gain FET and MMIC using a compound semiconductor used in a high frequency FET, like GaAs (Gallium Arsenide), InP (Indium Phosphide), GaN (Gallium Nitride), SiC (Silicon Carbide) and ZnO (Zinc Oxide), and a Si (silicon) based semiconductor, like CMOS (Complementary Metal Oxide Semiconductor) and SiGe (Silicon Germanium).
The present invention has been described above by referring to exemplary embodiments (and example); however, the present invention is not supposed to be limited by above exemplary embodiments (and example). The configurations and the details of the present invention can be given various changes in a scope of the present invention that skilled person may understand.
This application claims a priority based on Japanese Laid-Open Application 2009-83063 filed on Mar. 30, 2009 of which all the disclosures are incorporated in this application.
Number | Date | Country | Kind |
---|---|---|---|
2009 083063 | Mar 2010 | JP | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/JP2010/055321 | 3/26/2010 | WO | 00 | 9/28/2011 |