Field Effect Transistor

1. Field of the Invention

The present invention relates to an improvement of a dual gate GaAs FET.

2. Description of the Prior Art

Recently, the dual gate GaAs FETs which have a superior high frequency characteristic as well as a high gain are being developed.

The main feature of the dual gate structure FET is a superior gain control characteristics. That is, by impressing a D.C. bias voltage on a second gate, a gain control with small input and output detunings is easily obtainable. In case, the dual gate structure GaAs FET is used in a tuner or a front end stage of a television set, it is necessary to design its drain to source saturation current I.sub.DSS to be smaller than 40 mA (which corresponds to the voltage V.sub.G2SC second-gate and the source to cut off the drain current of 3.0 V). Such dual gate structure GaAs FET having the I.sub.DSS under 40 mA has had a poor cross modulation characteristic, especially at large gain reduction range, thereby forming a major cause of hindering wide actual use of such FET in the home use in television tuners.

In general, the distortion is dependent on the linearity of a curve between voltage V.sub.G1S between the second gate and the source and mutual conductance Gm (which is defined by Gm=dI.sub.DS /dV.sub.G1S, where I.sub.DS is a drain to source current), and in amplifiers, the distortion mainly consists of a third distortion which is induced by a secondary term of the curve. By expressing the drain current I.sub.D as follows:

where I.sub.DO, .alpha., .beta. and .gamma. are constants, and the amplitude Au of undesired signal by

where m.sub.u is a constant, then, intensity of the distortion m.sub.k is expressed by:

In a high frequency range, the linearity of the Gm can be regarded substantially dependent on the linearity of the S parameter .vertline.S.sub.21 .vertline. with respect to the gate voltage V.sub.G1S. FIG. 1 shows characteristic curves of relations between the S parameter .vertline.S.sub.21 .vertline. at 1 GHz and gate bias voltage V.sub.G1S first gate to source voltage) of a conventional Schottky barrier dual gate GaAs FET for ordinary communication use where I.sub.DSS is 50 to 70 mA (corresponding to V.sub.G2SC (second gate to source voltage to cut off drain current) of 3 to 5 V). In the conventional Schottky barrier dual gate GaAs FET, both the lengths of the first gate and the second gate are designed about 1 .mu.m. FIG. 2 shows cross-modulation characteristic of the abovementioned dual gate GaAs FET, wherein the ordinate is graduated with such levels of undesired signal of 30% modulation that which induces a 1% distortion to a desired signal and the abscissa is graduated by gain reduction (in dB). The curve of FIG. 2 is obtained by changing the first gate bias V.sub.G1S and the second gate bias V.sub.G2S along an oblique broken line in FIG. 1, that is by changing

V.sub.G1S . . . from -2 V to -0.5 and

V.sub.G2S . . . from +2 V to pinch off voltage (-3 to -5 V), respectively, where the undesired signal frequency or the interference signal frequency fu is 762 MHz and desired signal frequency fd is 750 MHz. In an actual amplifier, the curve of FIG. 2 should be in the area defined as above the dotted oblique broken line therein. The curve of FIG. 2 has two sags, one at the gain reduction of about -5 dB and the other at the gain reduction of about -20 dB, showing poor cross-modulations around these areas, which correspond to the parts of poor linearity of the curves in FIG. 1. Though there is some sags in the characteristic curve of FIG. 2, the curves being above the broken line fulfills actual demand for the cross modulation. However, when the drain current I.sub.DSS is set under 40 mA in order to fulfill the demand for home-use appliances, the distortion characteristics become worse, thereby hindering wide usage of the FET in actual circuits. FIG. 3 shows the cross-modulation characteristic of another dual gate GaAs FET, wherein I.sub.DSS is set to such a small value as 20 mA. As shown by the curve of FIG. 3, such dual gate GaAs FET of small I.sub.DSS has a poor cross-modulation characteristic around the gain reduction of -30 dB. The inventors confirmed through experiments that the smaller the I.sub.DSS is set, the poorer the distortion characteristic becomes.

The present invention has as its object to provide an improved dual gate GaAs FET of a small I.sub.DSS (drain saturation current), wherein the improvement is made to attain an improved distortion characteristic.

FIG. 1 is a graph of a characteristic curve showing relation between .vertline.S.sub.21 .vertline. (S parameter) and V.sub.G1S (voltage between the first gate and the source).

FIG. 2 is a graph of cross-modulation characteristic showing relation between gain reduction and level of undesired signal of a conventional dual gate GaAs FET having a large I.sub.DSS (drain to source saturation current).

FIG. 3 is a graph of cross-modulation characteristic showing relation between gain reduction and level of undesired signal of a conventional dual gate GaAs FET having a small I.sub.DSS.

FIG. 4 is a graph showing relations between I.sub.DS (drain to source current) and V.sub.G2S (voltage between the second gate and the source) of the present invention.

FIG. 5 is a graph showing relations between .vertline.S.sub.21 .vertline. and V.sub.G2S.

FIG. 6 is a graph of cross-modulation characteristic showing relation between gain reduction and level of undesired signal of a dual gate GaAs FET embodying the present invention.

FIG. 7 is a plan view of an example of an FET of the present invention.

FIG. 8 is a sectional elevation view at the sectional plan VIII--VIII of the present invention.

FIG. 9 is a graph showing a relation between V.sub.G2S and voltage drop V.sub.D2 under the second gate of the FET of FIG. 8 and FIG. 9.

FIG. 10 is a graph showing a relation between V.sub.G2S and gain reduction.

FIG. 11 is a circuit diagram of a tuner embodying the dual gate FET of the present invention.

A dual gate Schottky barrier gate GaAs FET having a value of 40 mA or smaller of a drain to source saturation current characterized in that the length of a second gate which is disposed between a first gate and a drain is 1.5 .mu.m or longer.

FIG. 4 shows characteristic curves between I.sub.DS (source to drain current) and V.sub.G2S (voltage between the second gate and source), for the examples of the gate electrode length L.sub.G2 of 1 .mu.m and 2 .mu.m. As shown in FIG. 4, the example of L.sub.G2 =2 .mu.m has such advantages over the other example of L.sub.G1 =1 .mu.m that I.sub.DS starts to decrease from the smaller .vertline.V.sub.G2S .vertline. voltage and that near the pinch off area (that is near the area of V.sub.G2S =-2.0 V) the curve smoothly bends. Those are reflected in the .vertline.S.sub.21 .vertline. vs. V.sub.G2S curve of FIG. 5, where the curve for L.sub.G2 =2 .mu.m has better linearity than the curve for L.sub.G2 =1 .mu.m. That is, the example of L.sub.G2 =2 .mu.m is superior to the example of L.sub.G2 =1 .mu.m in linearity around the gain reduction of -30 dB, and this means the superiority of the cross-modulation characteristics for the part of L.sub.G2 =2 .mu.m. That is to say, the cross-modulation is dependent on the linearity of the S parameter .vertline.S.sub.21 .vertline., and hence the distortion can be improved by improving the linearity of .vertline.S.sub.21 .vertline..

FIG. 6 is a graph of a distortion characteristic curve showing a relation between the level of undesired signal and gain reduction, for a dual gate GaAs FET having the following gate lengths

the first gate length L.sub.G1 =1 .mu.m, and

the second gate length L.sub.G2 =2 .mu.m.

The curve of FIG. 6 is obtained by plotting the operational points by changing V.sub.G2S and V.sub.G1S simultaneously as follows:

V.sub.G2S . . . from +2.0 V to -2.0 V and

V.sub.G1S . . . from -1.5 V to -0.2 V.

By comparing FIG. 6 with FIG. 3, it is obvious that the distortion is improved at the range around the gain reduction of -30 dB. As a result of such improvement of the linearity at the large gain reduction range, the dual gate GaAs FET in accordance with the present invention has enough applicability in practical use in home use television sets or the like appliances. The inventors confirmed through many experiments that the decrease of the distortion appears for the second gate length L.sub.G2 of

FIG. 7 and FIG. 8 show an actual example of dual gate GaAs FET embodying the present invention, wherein on a semi-insulating single crystalline n-GaAs wafer 1, a buffer layer 2 of single crystalline n-GaAs and an active layer 3 to become a channel layer of FET of single crystalline n-GaAs are formed by sequential epitaxial growth. The active layer 3 is then selectively mesa-etched to form a plan view pattern thereof as shown in FIG. 7. Than a source electrode 4, a drain electrode 5, a first gate electrode 6 and a second gate electrode 7 are formed in the patterns as shown in FIG. 7. The source and drain electrodes are formed by Au-Ge alloy so as to make ohmic contacts with the active layer 3 and the first and the second gate electrodes are formed by Cr or Ni so as to form Schottky contacts with the active layer 3. As shown in FIG. 7 and FIG. 8, the lengths L.sub.G1 and L.sub.G2 of the first gate and the second gate are selected as follows:

In general, when a negative bias is impressed on the second gate 7 to attain a gate controlling, a depletion layer d.sub.2 spreads underneath the second gate 7 as shown in FIG. 8. Then, when a signal is impressed to the first gate 6, a depletion layer d.sub.1 is formed under the first gate 6. The depletion layer d.sub.2 formed by the second gate works to increase a resistance R.sub.D2 and thereby increase a voltage drop V.sub.D2 under the second gate 7. Accordingly the voltage V.sub.D1S between the drain and source decreases. A dual gate FET can be regarded equivalent to a series connection of two FETs, the first one of them including the first gate 6 and the second one of them including the second gate 7 and source of the first FET and the drain of the second FET is the surface part of the active layer 3 at the part between the first gate 6 and the second gate 7. Therefore, the voltage V.sub.D1S is regarded as drain to source voltage of the first FET having the first gate 6; and when the voltage V.sub.D1S is lowered then the Gm of the first FET becomes lowered thereby its gain is suppressed. Hereupon, the voltage V.sub.D2 increases as the depletion layer d.sub.2 expands gradually nearing to the pinch off state; and when the second gate voltage V.sub.G2S nears to the pinch off voltage as shown by FIG. 9, the voltage V.sub.D2 rapidly increases, thereby rapidly lowering the V.sub.D1S and hence lowering the gain of the signal impressed on the first gate 6. The cross-modulation is induced mainly by non-linearity of the Gm on which the controlled gain substantially depends, and this is observed in FIG. 10, wherein the curve has a considerable non-linear part indicated in a dot circled part, where the gain reduction is about -30 dB. In the dual gate GaAs FET of small I.sub.DSS, depth of the active layer shown in FIG. 8 is small, and therefore, the depletion layer is supposed to expand to the pinch-off state with a smaller V.sub.G2S than the FET of larger I.sub.DSS. Therefore, an improvement of the distortion (cross-modulation) is attainable by improving the linearity indicated by the dotted circle in FIG. 10. For such improvement of the linearity, it is supposed that the curve of FIG. 9 should be preferably improved to have more smooth bending as shown by the broken line curve of FIG. 9. For such purpose, it is preferable to select the resistance R.sub.D2 of the channel part under the second gate to be larger hence to make the V.sub.D2 larger. It is supposed that such increase of the resistance R.sub.D2 is attained by the increasing of the length L.sub.G2 of the second gate. And through many experiments the inventors confirmed that the necessary condition for the FET of 40 mA or smaller I.sub.DSS is:

FIG. 11 is a circuit diagram of a UHF tuner which employs the dual gate GaAs FET Q in accordance with the present invention used as varactors in a resonator. This tuner has a satisfactory low noise and low cross-modulation characteristics. In FIG. 11, the UHF input terminals 10 receive UHF input signal, for example UHF television signal of 500 to 900 MHz from an antenna. The terminal 11 receives an AGC signal (a D.C. voltage). The numeral Q.sub.F designates a dual gate GaAs FET in accordance with the present invention, wherein marks

S designates the source,

D designates the drain,

G.sub.1 designates the first gate, and

G.sub.2 designates the second gate. A block 12 is an input tuning circuit, a block 13 is an inter-stage tuning circuit, a diode 14 is a mixer diode, block 15 is a local oscillator, a terminal 15' is an intermediate frequency (IF) signal output terminal, varactor diodes (variable capacitance diodes) 16, 16 are impressed with tuning voltage given through a terminal 17, and a power input terminal 27 receives a D.C. power from a D.C. source not shown here. An AFC terminal 18 receives an AFC signal. Q1 is a local oscillator transistor. An IF amplifier 19 has amplifier transistors Q2 and Q3 and outputs amplified IF signal from its output terminal 20.

The UHF tuner employing the dual gate GaAs FET in accordance with the present invention is very superior in distortion characteristics in the UHF band as elucidated referring to the characteristic graphs of FIG. 5 and FIG. 6, and therefore is very useful in attainment of gain controlling up to about 35 dB gain reduction without noticeable cross-modulation to induce distortion of the signal.