BACKGROUND OF THE INVENTION

This invention relates to an improved field-effect transistor (FET) and especially to improvement of FET's by tailoring the source-gate channel resistivity to be high in the upper layer and lower in the lower layer.

Transconductance, gate capacitance, source parasitics, and source-gate channel resistance are the factors universally known to affect the performance of field-effect transistors. Previous work in the field has done much to improve transconductance by improving materials' quality and materials' interfaces. Gate capacitance has been reduced by using submicrometer resolution lithography. Source parasitics have been greatly reduced by better metallization for ohmic contacts by "via" technology and by monolithic, Class B, push-pull circuit techniques. Only source-gate channel resistance has evaded a solution enabling it to be reduced without adversely affecting gate leakage characteristics.

SUMMARY OF THE INVENTION

Accordingly, it is an object of this invention to reduce the source-gate channel resistance in an FET without adversely affecting the gate leakage characteristic.

Another object is to raise the resistivity of a shallow layer of the source-gate channel in an FET, which layer is in contact with the Schottky-barrier portion of the gate, without increasing the resistivity of the remainder of the source-gate channel.

The above and other objects of the invention are accomplished in a GaAs FET by selectively bombarding the N+ doped source-gate channel region with boron ions to raise the resistivity of a shallow upper layer of the channel which is in contact with the Schottky-barrier film of the gate without increasing the resistivity of the underlying channel region. The technique makes use of the virtual non-existence of ionicity in the boron-arsenide bond to induce arsenic vacancies within the crystal which may, in part, be filled with column IV acceptors (e.g., silicon) with the result that deep-level compensating centers are formed (probably by silicon-silicon complexes) thereby transforming the material so treated into semi-insulating material.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic illustration of a prior-art FET made by an epitaxial growth technique.

FIG. 2. is a schematic illustration of a prior-art FET made by an ion implantation technique.

FIG. 3 is a schematic illustration of an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

FIGS. 1 and 2 illustrate the construction of typical priorart FET's. The FET of FIG. 1 is produced by the growth of epitaxial layers 12 and 14 on a semi-insulating GaAs substrate 10. Layer 14 has a higher concentration of donor impurity ions (approximatel 1 layer 12 and therefore has higher conductivity. The active layer 14 in FIG. 2 is produced by implantation of donor ions. The device also includes a source 16, a drain 18 and a gate 20. In both cases, the active channel, which may, for example, be about 0.5 microns deep, is uniformly doped in the horizontal direction (i.e., along the direction of charge-carrier movement).

Referring now to FIG. 3, the source-gate channel region 28, including a top portion 26 and a bottom portion 26', between source 16 and gate 20 (Schottky barrier gate) should be of much lower resistivity than a region 24 lying beneath the gate 20 and between the gate 20 and the drain 18. If the source-gate channel region 28 is heavily doped, however, the leakage of the Schottky barrier gate 20 is excessive. Techniques to create a vertical gradient of resistivity in this region have thus far been unsuccessful since there has been no ion implantation technology capable of reproducibly creating such a gradient. The ideal characteristic is one in which the electrically active impurity concentration at the top portion 26 of the source-gate channel region 28 is less than or equal to that of the region 24 directly beneath the gate 20, while the electrically active impurity concentration in the lower portion 26' of the source-gate channel region 28 is at least an order of magnitude greater than the impurity concentration in the same region 24 directly beneath the gate 20. The achievement of such an impurity gradient profile virtually eliminates the source-gate channel resistance as a significant factor adversely affecting FET performance and the gradient does not adversely affect the characteristics of the Schottky-barrier gate 20.

To achieve this optimum gradient profile in the source-gate channel region 28 without adversely changing the impurity profile elsewhere requires a new approach based on an understanding of how impurity complexes can be used to reproducibly and reliably control the properties of semiconductors. The implantation of boron is known to render GaAs semi-insulating since it compensates, or neutralizes, other impurities which render the GaAs more conductive. Thus the effect of the boron is to render the doped GaAs more insulative. Only recently has it been found that the boron implant dose need not be excessive to the extent of rendering the semiconductor amorphous but, instead, need only exceed the concentration of other impurities within the GaAs. More recently, it has been shown that boron implanted in GaAs does not diffuse within the GaAs at elevated temperatures as do most other impurities.

Still referring to FIG. 3, to the optimum gradient profile in the source-gate channel region 28 of the improved ion-implanted FET shown, an impurity ion is selected (e.g., Si, or Si and S) and selectively implanted into the source-gate channel region 28. This may be done simultaneously with the N+ selective source and drain implants in regions 22 and 30, respectively. (In this context, the term "selectively" applies to the particular region selected for any ion implantation.) As shown, the region 24 is implanted only to the N state, or a concentration of about 1 aforementioned, are implanted to the N+ state, or a concentration of about 1 shallower (i.e., done with a lower implantation voltage) implant (e.g., 200-500 Å) of boron into the top portion 26 of source-gate channel region 28. By so doing, As vacancies are created in the top portion 26 but not in the bottom portion 26' of the source-gate channel region 28. The concentration of the boron implant should exceed the concentration of the N+ impurity by a factor of 2 or more to ensure that some As vacancies remain in the top portion 26 after other As vacancies are filled by the acceptor (silican) ions previously implanted. The maximum concentration of boron should be below that which would cause the GaAs to become amorphous; thus, the concentration should not be more than about 5 /cm.sup.3.

Activation/annealing of the implanted ions may then proceed in a conventional manner chosen by the fabricator (i.e., thermal, laser and/or electron beam). The annealing ambient must be chosen so that the boron-implanted region, i.e., top portion 26, is not etched away in the process (e.g., use flowing arsine, proximity capping or a good silicon nitride encapsulant).

Obviously, many modifications and variations of the present invention are possible in light of the above teachings. It is therefore to be understood that, within the scope of the appended claims, the invention can be practiced otherwise than as specifically described.