MESFET Semiconductor Device and Method of Making

This invention is directed to microelectronic semiconductor devices and methods of making such devices, and more particularly, to a MESFET device and a process for making said MESFET device.

The history of integrated semiconductor circuit design has been characterized by a trend toward increasing circuit densities. Various technologies have been invented to stimulate this trend. For example, transistor-transistor logic, TTL, was standard in digital equipment for a long time but has given way in many areas to N-channel MOS logic circuits because of their superiority in speed power product, packing density and ease of device fabrication. For these reasons devices fabricated using these technologies are finding application primarily in memory and microprocessor circuits. The MESFET is another device that offers many of the advantages of the N-channel MOS technology without some of its disadvantages. Its primary application will also be in digital technology such as memory and microprocessor type circuits.

One of the problems with N-channel MOS devices is that when they are scaled down in size the gate oxide thickness must be scaled down accordingly. This creates a problem because it is very difficult to fabricate thin silicon oxides that are free from pinholes. If there is a pinhole in one of the gate oxides there will be a gate to channel short and consequently a device failure. Since there are thousands of gate oxide areas on a typical N-channel MOS memory or microprocessor this problem can be very serious.

It is a principal object of this invention to provide a MESFET device which is useful in the design of high density digital logic circuits. Another object is to provide a method for fabricating the MESFET device.

In accordance with the present invention a MESFET device is provided that has implanted channel stoppers under field oxide, implanted source and drain regions, implanted channels for depletion and enhancement mode devices, a metal-semiconductor gate and a thin film of metal for the interconnects. In making the device, a thin silicon oxide layer is grown and a silicon nitride layer deposited upon it. The areas for the channel stoppers are defined and implanted, and then a thick field oxide is grown consuming the silicon as it grows. The areas for the sources and drains are defined and followed by an implant and anneal. The thin oxide layer and the nitride are removed and the channel implant takes place. After the channel implant, enhancement mode device channel areas are patterned and the depletion mode implant follows. An anneal is followed by the deposition of a silicon nitride layer and then a silicon oxide layer. The contact and gate areas are defined in the oxide and nitride layers and platinum silicide regions are created by platinum deposition and sinter. The deposition and definition of a metal system completes the process.

The novel features believed characteristic of the invention are set forth in the appended claims. The invention itself, however, as well as other features and advantages thereof, will be best understood by reference to the detailed description which follows, read in conjunction with the accompanying drawings, wherein:

FIG. 1 is a greatly enlarged plan view of a small portion of a semiconductor chip showing the physical layout of an inverter fabricated using the invented process;

FIG. 2 is an electrical schematic diagram of the inverter of FIG. 1;

FIGS. 3a-3d are elevation views in section of the inverter of FIG. 1, taken along the lines a--a; and

FIGS. 4a-4j are elevation views in section of the semiconductor device of FIGS. 1 and 3a-3d, at successive stages in the manufacturing process.

Referring to FIG. 1, a physical layout of a MESFET inverter 10 is shown which is fabricated using the process which is the subject of the invention. The inverter is greatly enlarged over actual size. The inverter is also shown in FIG. 2 drawn as an electrical schematic diagram, with the parts numbered and laid out the same.

The inverter of FIGS. 1 and 2 having an input 14 and an output 12' consists of two MESFET transistors Q.sub.1 and Q.sub.2, Q.sub.1 having its source connected to Vss or ground line 11. Q.sub.1 operates as a driver transistor having its drain connected to a node 12 of the source of Q.sub.2. Q.sub.2, operating as a depletion mode load device, has its drain connected to Vdd or a positive supply line 13 and its gate and source connected to a node 12 at the drain of Q.sub.1.

The circuit of the inverter in FIG. 2 is the same as that used in conventional MOS or JFET integrated circuits. The MESFET closely resembles a JFET except that the PN gate junction of the JFET is replaced by a metal-semiconductor junction. This metal-semiconductor junction significantly reduces minority carrier injection and allows the MESFET to turn off more quickly than a JFET.

Section views of the inverter of FIG. 1 showing the details of construction are shown in FIGS. 3a-3d. It should be kept in mind when examining the section views that not all geometries, junction depths and layer thicknesses are necessarily to scale, some having been enlarged or reduced where necessary to reflect features of the invention. The inverter 10 is a small part of the substrate 16 of P-type silicon. The transistors Q.sub.1 and Q.sub.2 are formed by N+ diffused regions 17, 18, and 20 which create the source and drain regions. The channels 21 and 22 of the transistors are implanted to vary the threshold voltages of the transistors. Platinum silicide regions 23 and 24 form metal-semiconductor junctions which are the gates of the MESFETs. Platinum silicide regions 30, 31 and 32 form contacts to the source and drain regions. Field oxide 25 covered with a silicon nitride layer 51 and a silicon oxide layer 52 exists at all areas where N+ regions or transistors do not exist, and a P+ boron-doped channel-stop region 28 is created under all areas of the field oxide. The lines 11, 12', 13, and 14 are metal strips making contact to the platinum silicide areas.

Referring now to FIGS. 4a-4j, a process for making the MESFET integrated circuit device of FIGS. 1 and 3a-3d will be described. The starting material is a slice of P-type monocrystalline silicon, perhaps 2 inches in diameter and 16 mils thick, cut on the <100> plane, of a resistivity of about 30 to 50 ohm-cm. In FIGS. 3a or 4a, a wafer or body 16 represents a very small part of the slice, chosen as a representative sample cross section. First, after appropriate cleaning, the slice is oxidized by exposing to oxygen in a furnace at an elevated temperature of perhaps 900 degrees C. to produce an oxide layer 33 of a thickness of about 500 Angstroms. Next, a layer 34 of silicon nitride is formed by exposing the slice to an atmosphere of silane and ammonia in a furnace tube of perhaps 900 degrees C. This layer 34 is grown to a thickness of about 1500 Angstroms. A coating 35 of photoresist is aplied to the entire top surface, then exposed to ultraviolet light through a mask which defines the desired pattern, and developed. This leaves areas 36 where nitride is to be etched away. The slice is subjected to a nitride etchant, which removes the exposed part of the nitride layer 34 but does not remove the oxide layer 33 and does not react with the photoresist 35.

The slice is now subjected to an ion implant step, whereby boron atoms are implanted in the areas of silicon not covered by photoresist 35 and nitride 34. The photoresist could have been removed, but preferably is left in place as it masks the implant. Boron is an impurity which produces P-type conductivity, so a more heavily doped P+ region 37 will be produced in the surface. The oxide layer 33 is left in place during the implant because it prevents the implanted boron atoms from out-diffusing from the surface during subsequent heat treatment. The boron implant is at a dosage of about 8.times.10.sup.12 cm.sup.2 at 40 KeV.

As will be seen, the region 37 does not exist in the same form in the finished device, because some of this part of the slice will have been consumed in the oxidation procedure.

The photoresist 35 is removed and the following step in the process is formation of field oxide, which is done by subjecting the slices to steam or an oxidizing atmosphere at about 950 degrees C. for perhaps 3 hours. This causes a thick field oxide region or layer to be grown, and this region extends into the silicon surface because silicon is consumed as it oxidizes. The nitride layer 34 masks oxidation beneath it. The thickness of this oxide layer 25 is about 6000 to 7000 Angstroms, half of which is above the original surface and half below. The boron doped P+ region 37 formed by implant will be partly consumed, but will also diffuse further into the silicon ahead of the oxidation front. Thus, a P+ region 28 will result which will be a channel stopper to prevent parasitic effects between devices in separate isolated areas. FIG. 4b reflects a cross-section of the slice at this point in the fabrication process.

The slice is next subjected to an etch to remove the oxide which has formed by converting the outer part of the silicon nitride layer 34 to silicon oxide during the field oxidation step. A coating 38 of photoresist is applied to the entire top surface, then exposed to ultraviolet light through a mask which defines the desired pattern, and then developed. The slice is plasma etched and the nitride that is uncovered by photoresist 38 is removed. The slice is subjected to an oxide etchant and the oxide areas uncovered by nitride 34 are removed. This leaves windows 41, 42 and 43 where the impurity used to form the source and drain regions will be implanted.

One of the features of the invention is the use of the field oxide 25 to define three sides of the source and drain windows 41 and 43 and two sides of the common source/drain window 42. As seen in FIG. 4c photoresist defines the other edges of these source and drain windows. This technique gives source and drain diffusion boundaries that are coincident with some edges of the field oxide 25, permitting at a subsequent step the corresponding boundaries of the source and drain contacts to be made coincident with the edges of the field oxide 25. This eliminates the need for a contact to diffusion edge space allownace and consequently increases chip packing density.

The slice is now subjected to an ion implant step where arsenic atoms are implanted in the areas of silicon not covered by photoresist 38, nitride 34 and oxide 33. Arsenic is an impurity which produces N-type conductivity in silicon. Implanted regions 44 and 45 are produced by implanting at a dose of about 5.times.10.sup.15 at 100 KeV. Although field oxide regions 25 are not covered with photoresist no arsenic is able to penetrate them and reach the channel stoppers 28 underneath. Therefore the implanted regions 44 which will become a drain 17 and a source 20 (see FIGS. 3a and 4d) are defined by thick field oxide 25 and photoresist 38. After implant the photoresist is removed.

The following step in the process is the growth of a thin oxide in regions 41, 42 and 43 by subjecting the slice to an oxidizing ambient at about 800 degrees C. for perhaps 1 hour. This oxide will protect the silicon underneath when the nitride is removed. The next step is the removal of the nitride by exposing the slices to phosphoric acid at about 160 degrees C. for perhaps 1 hour. The slice is then placed in hydrofluoric acid to remove all oxide from the slice except in regions 25 where there is thick field oxide. The slice is then annealed in an oxidizing ambient, and in an inert ambient, preferably nitrogen, at about 1025 degrees C. for perhaps 1 hour to diffuse regions 44 and 45 into the slice and form source/drain regions 17, 18 and 20. The thin oxide region grown during the previous anneal is removed by dipping the slice in hydrofluoric acid.

The next stage in the process is the creation of the N-channel regions 21 and 22. The slice is exposed to a phosphorus implant at a dose of about 1.times.10.sup.12 /cm.sup.2 at 80 KeV which creates implanted regions 46 and 47. The phosphorus is implanted in all regions of the slice however, the effect of the implant is negligible in regions 17, 18 and 20 where the N-type impurity concentration is much higher than that implanted. As mentioned previously the implant will not penetrate the field oxide regions 25. A photoresist coating 48 is applied over the entire slice and exposed to ultraviolet light though a mask which exposes only those portions 49 that cover the areas that will become the channels for the enhancement mode transistors. Upon developing, unexposed photoresist is removed from the slice leaving only the exposed regions 49. The slice is then subjected to an implant which penetrates the slice in all areas except those regions 49 where photoresist exists. This implant is phosphorus at a dose of about 1.3.times.10.sup.12 /cm.sup.2 at 80 KeV. For the reasons discussed above the implant is effective only in areas 50 which become the channels 21 of the depletion mode load transistors. The photoresist is removed and the slice is annealed in an inert atmosphere at about 700 degrees C. for perhaps 10 minutes to complete the formation of the channels 21 and 22 of the MESFET transistors Q.sub.1 and Q.sub.2. Although the N-regions 21 and 22 are being used here as channels, on other parts of the chip they may be used as high value resistors.

A layer 51 of silicon nitride is formed next by exposing the slice to an atmosphere of silane and ammonia in an rf plasma reactor as shown in FIG. 2 of U.S. Pat. No. 3,907,616. This layer is grown to a thickness of about 1500 Angstroms. Next, a layer 52 of silicon oxide is grown on top of the nitride layer 51 by using an rf plasma reactor of the same type as previously described. This oxide is grown to a thickness of about 1000 Angstroms. FIG. 4f shows the slice at this point in the process.

The next stage of the process is the formation of the MESFET gates 23 and 24 and source and drain contacts 30, 31 and 32. A photoresist coating is applied over the entire slice and exposed to ultraviolet light through a mask which exposes everything except those areas that are to become gates and contacts of the MESFET transistors. Upon developing, unexposed photoresist is removed from areas 54 where the gates will be created and areas 55 and 56 where the contacts will be created. The oxide layer 52 which is not covered by photoresist 53 is removed by etching the slice in hydrofluoric acid. The photoresist coating 53 is then removed and the slice is placed in phosphoric acid which etches away the nitride layer 51 in those areas which are not covered by the silicon oxide layer 52, and does not react with the silicon oxide. This leaves the slice as shown in FIG. 4g.

Another of the features of the invention, as shown in FIGS. 4g and 4h, is the use of the field oxide 25 to define three edges of the source and drain contacts 30 and 32 and contact windows 56 and two edges of the common source/drain contact 31 and contact window 55. The other edges of these contacts and windows are defined by the photoresist 53. Defining the contacts in this manner permits the contacts to be placed almost coincident with the corresponding edges of the sources and drains. This eliminates the need for a contact to diffusion edge space allowance and consequently increases chip packing density.

The slice is next placed in a metal sputtering system where a metal such as platinum is sputtered on the slice to a thickness of about 400 Angstroms. Upon removal from the sputtering system the slice is placed in an furnace type at about 500 degrees C. for perhaps 10 minutes. This heat treatment causes the platinum to react with the silicon to form platinum silicide gates 23 and 24 and contacts 30, 31 and 32 of about 800 Angstroms thickness. Next, the platinum on the oxide layer 52 is removed by placing the slice in aqua regia for about 15 minutes, resulting in the structure of FIG. 4h.

Where there is a platinum silicide-silicon boundary and the silicon is lightly doped N-type this structure forms a P-N junction where the P-region is platinum-silicide and the N-region is the silicon. Where the N-region is very heavily doped the boundary is not a P-N junction but is only an ohmic contact region. This is the difference between the gate 23 and 24 and contact 30, 31 and 32 regions. Although the platinum regions 23 and 24 are being used as gates of MESFETs, on other parts of the chip platinum silicide regions may be formed in N-regions to form discrete diodes of the type discussed above.

As seen in FIG. 4j the fabrication of the device is completed by connecting the metal leads to the device. The first step here involves the deposition of a layer 57 of barrier metal such as titanium-tungsten (Ti-W) on the slice by use of a metal sputtering system. The purpose of the Ti-W layer is to prevent the subsequent metal layer from spiking through the platinum silicide. The Ti-W is deposited to a thickness of about 2000 Angstroms. Next another layer 58 of metallication such as gold is deposited on the slice to a thickness of about 5000 Angstroms. After the evaporation procedures a photoresist coating 59 is applied over the entire slice, then exposed to ultraviolet light through a mask that exposes the areas that are to become the metal leads 11, 12' 13 and 14. Upon developing, the unexposed photoresist is removed leaving photoresist where leads are to be formed. At this stage of the process the slice cross-section appears as in FIG. 4i. The slice is then subjected to a gold etch that removes the gold layer 58 that is not covered by photoresist but does not remove the Ti-W and does not attack the photoresist. The photoresist is removed and the slice placed in hydrogen peroxide which etches the Ti-W that is not protected with gold but does not react with the gold. The slice is annealed in forming gas at about 400 degrees C. for perhaps 15 minutes which completes the process.

The operation of the MESFET is similar to that of a JFET, both devices having source, drain and gate regions. The major difference is the replacement of the semiconductor P-N junction gate on the JFET with a metal-semiconductor junction gate. The metal-semiconductor junctions on the MESFETs are at the junctions of the platinum silicide regions 23 and 24 and the channel regions 21 and 22 as seen in FIGS. 3a and 4j. To operate an N-channel enhancement mode MESFET such as Q.sub.1 the drain 18 is biased positive with respect to the source 20. Relatively little current will flow from drain to source with the gate 24 open since the depletion region of the junction formed by the platinum silicide region 24 and the channel 22 reaches all the way across the channel 22. As the gate 24 is biased positive with respect to the source 20 current begins to flow from drain to source because the depletion region no longer reaches completely across the channel. There is also some current flowing into the gate 24. The voltage from gate 24 to source 20 is clamped to approximately 0.45 volts by the platinum silicide-silicon diode. When the gate to source voltage is brought to zero volts current ceases to flow from drain to source because the depletion region again reaches across the channel. Since the depletion region width is inversely proportional to the impurity concentration in the channel 22, the threshold voltage of the MESFET can be varied by changing the impurity in the channel 22.

With an N-channel depletion mode MESFET such as Q.sub.2, with the gate 23 open current would flow from drain 17 to source 18 as soon as the drain is biased positive with respect to the source. In this type device the channel 21 is doped heavier than the channel 22 of the enhancement mode device Q.sub.1. Consequently, the depletion region of the platinum silicide-silicon junction does not reach all the way across the channel 21 and current may flow from drain 17 to source 18. To prevent current flow in this device the gate 23 must be biased negative to the source 18. This is not possible of course in our arrangement as seen in FIGS. 2, 3a and 4j because the gate 23 is connected to the source. This connection is made to use Q.sub.2 as a load device. However on other parts of the chip, depletion mode MESFETs that operation as transistors can exist and do not have their gates connected to their sources.