Polysilicon Interconnects With Pin Poly Diodes

This invention relates to a semiconductor device comprising a plurality of bipolar transistors.

As the packing density of a semiconductor integrated circuit is increased, the electrodes of the circuit occupy more area. To further increase the packing density, it is necessary to reduce the area occupied by the electrodes. The packing density of MOS integrated circuit is improved by using impurity-doped polysilicon films as at least gate electrodes. A polysilicon film gate, which is generally known as "silicon gate", can easily be formed in a multilayer structure, thus providing a great latitude of design.

Polysilicon films are seldom used for electrodes in a bipolar integrated circuit. In this specification electrode means also interconnection. Aluminum films are used because a bipolar device differes from a MOS device in an operation mechanism and the structure. First, electric currents flow through the emitter, base and collector of the bipolar device, whereas no gate current flows through the MOS device. Secondly, the bipolar device has diffusion layers of both conductivity types, i.e. diffusion layers of P type and diffusion layers of N type. By contrast, the MOS device has diffusion layers of only N type if it is an N-channel device or diffusion layers of only P type if it is a P-channel device.

Since a base current flows through it, the bipolar device must have low resistance base electrodes. Otherwise it would fail to operate. Therefore the electrodes are generatly aluminum films, not polysilicon films. On the other hand, since no gate current flows through it, the MOS device may have relatively high resistance gate electrodes. This is why polysilicon films which have a higher resistance than aluminum films are used for gate electrodes in an MOS device.

Since the bipolar device has diffusion layers of both conductivity types, PN junction would be formed in many cases between the P diffusion layers and the N diffusion layers if the electrodes are made of polysilicon. If such PN junctions are formed, the bipolar device will cease to function. This problem does not occur in the MOS device whose diffusion layers are of either P type or N type.

It is difficult, however, to form a multilayer structure of aluminium films. Further it is difficult to arrange aluminum films in the desired pattern. Therefore in the known bipolar device, aluminum films are so arranged as to provide so-called "single layer electrodes". The bipolar devices cannot have their packing density improved.

Integrated injection logic gates (IIL gates) are well-known bipolar devices. IIL gates are popular because they are logic gates which have a low power consumption. An IIL gate comprises an upwards vertical transistor for an inverter and a transistor for an injector. The injector transistor is complimentary to the inverter transistor and applies a load on the base of the inverter transistor. The injector transistor is a lateral transistor, whose collector and base are respectively the base and emitter of the inverter transistor. The electrodes of the IIL gates are aluminum films. Until now it has not been through possible to construct IIL gates whose electrodes are made of polysilicon. As mentioned above, as it is difficult to form aluminum films one upon another, the packing density of the IIL gates cannot be increased.

When closely studied, an IIL gate is found to possess the following characteristic features. First, its logic amplitude is small, and its operating current is extremely small. Secondly, in the electrode connecting a P layer of one IIL gate and an N layer of another IIL gate, current flows always from the P layer to the N layer. The IIL gate has such a structure as shown in FIG. 1 which is a cross sectional view.

As shown in FIG. 1, a buried N.sup.+ layer 14 is formed on a P.sup.- substrate 12. On the substrate 12 and the buried N.sup.+ layer 14 an N epitaxial layer 16 is formed. Formed in the N epitaxial layer 16 is an N.sup.+ diffusion layer 18 which contacts the busied N.sup.+ layer 14. Two P diffusion layers 20 and 22 are formed in that portion of the N epitaxial layer 16 which is laid on the N.sup.+ layer 14. In the P diffusion layer 22 two N.sup.+ diffusion layers 24 and 26 are formed. The N.sup.+ diffusion layer 18 is connected to the ground GND, and the P diffusion layer 20 is connected to a power source voltage +E. The P diffusion layer 22 is connected to an input terminal IN. The N.sup.+ type diffusion layers 24 and 26 are connected to output terminals OUT.sub.1 and OUT.sub.2, respectively.

FIG. 2 is an equivalent circuit diagram of the IIL gate shown in FIG. 1. An NPN vertical transistor Q.sub.1 for an inverter consists of the N layer 16 which serves as an emitter, the P layer 22 which serves as a base and the N.sup.+ layers 24 and 26 which serve as collectors. A PNP lateral transistor Q.sub.2 consists of the P layer 20 which serves as an emitter, the N layer 16 which serves as a base and the P layer 22 which serves as a collector. The IIL gate may be represented by such a symbol as shown in FIG. 3. Then, a T type flip-flop comprised of IIL gates will be illustrated in FIG. 4.

As shown in FIG. 4, to connect the input terminal (i.e. P layer) of, for example, an IIL gate 32 to the output terminal (i.e. N layer) of the adjacent IIL gate 34, an electrode is necessary. Now, current flow in such electrodes of an integrated circuit which comprises IIL gates will be discussed, and reference will be made to FIG. 5.

FIG. 5 is a schematical plan view of an integrated circuit comprising a plurality of IIL gates, of which only three, i.e. IIL gates 36.sub.1, 36.sub.2 and 36.sub.3, are illustrated. Electrodes connecting these IIL gates 36.sub.1, 36.sub.2 and 36.sub.3 are also shown in FIG. 5. The IIL gate 36.sub.1 has a P layer 38.sub.1, an N.sup.+ layer 40.sub.1 and a P layer 42.sub.1. Similarly, the IIL gate 36.sub.2 has a P layer 38.sub.2, an N.sup.+ layer 40.sub.2 and a P layer 42.sub.2, and the IIL gate 36.sub.3 has a P layer 38.sub.3, an N.sup.+ layer 40.sub.3 and a P layer 42.sub.3. The P layers 38.sub.1, 38.sub.2 and 38.sub.3 are each the base (i.e. signal input terminal) of an inverter transistor. The N.sup.+ layers 40.sub.1, 40.sub.2 and 40.sub.3 are each the collector (i.e. signal output terminal) of the inverter transistor. The P layers 42.sub.1, 42.sub.2 and 42.sub.3 are each the emitter of an injector transistor. They are generally called "injectors". Hereinafter they will be called "injector P layers".

An electrode film 44.sub.1 is formed and connects the N.sup.+ layer 40.sub.1 of the IIL unit 36.sub.1 to the P layer 38.sub.2 of the IIL gate 36.sub.2. Likewise, an electrode film 44.sub.2 is formed and connects the N.sup.+ layer 40.sub.2 of the IIL gate 36.sub.2 to the P layer 38.sub.3 of the IIL gate 36.sub.3. Further provided is an electrode film 46 which connects the injector P layer 42.sub.1 of the IIL gate 36.sub.1, the injector P layer 42.sub.2 of the IIL gate 36.sub.2 and the injector P layer 42.sub.3 of the IIL gate 36.sub.3.

As long as the inverter transistor of the IIL gate 36.sub.1 is conductive, its base current flows from the injector P layer 42.sub.1 and its collector current flows through the base (i.e. P layer 38.sub.2) of the inverter transistor of the next IIL gate 36.sub.2. The inverter transistor of the IIL gate 36.sub.2 still remains non-conductive. When the inverter transistor of the IIL gate 36.sub.2 is rendered conductive, its collector current flows through the base (i.e. P layer 38.sub.3) of the IIL gate 36.sub.3. That is, in the electrode film 44.sub.1 connecting the N.sup.+ layer 40.sub.1 to the P layer 38.sub.2 current flows in the direction of arrow A.sub.1, from the P layer 38.sub.2 to the N.sup.+ layer 40.sub.1. And in the electrode film 44.sub.2 connecting the N.sup.+ layer 40.sub.2 to the P layer 38.sub.3 current flows in the direction of arrow A.sub.2, from the P layer 38.sub.3 to the N.sup.+ layer 40.sub.2.

This invention is based on the fact that current flows in specific directions in a semiconductor device which includes a plurality of bipolar transistors such as IIL gates. The object of the invention is to provide a semiconductor device which comprises a plurality of bipolar transistors and whose packing density is increased by using polysilicon films as electrodes.

According to this invention a semiconductor device is provided, which comprises a semiconductor substrate, an integrated circuit of a plurality of bipolar transistors formed in the substrate, each transistor including a diffusion layer of P conductivity type and a diffusion layer of N conductivity type, and polysilicon films of P and N conductivity types formed as electrodes connected with the diffusion layers of P and N conductivity types, respectively.

Further, this invention provides a method of manufacturing a semiconductor device which comprises a step of forming a metal layer of a high melting point on the all of the semiconductor surfaces including an insulation layer and P and N polysilicon films, both the insulation layer and the polysilicon films being formed on a semiconductor substrate and the P and N polysilicon films being connected to each other, a step of heating the metal layer to form a silicide of the metal on the P and N polysilicon films, and a step of removing by an aqua regia boiling process that portion of the metal layer which is formed on the insulation layer so that only the silicide of the metal formed on the polysilicon films is left in a self-aligned process.

FIG. 1 is a cross sectional view of an IIL gate;

FIG. 2 shows an equivalent circuit of the IIL gate shown in FIG. 1;

FIG. 3 is a symbol of the IIL gate shown in FIG. 1;

FIG. 4 is a circuit diagram of a known T type flip-flop in which the symbol shown in FIG. 3 is used;

FIG. 5 is a schematical plan view of a portion of a semiconductor device including IIL gates and electrode films connecting the IIL gates;

FIG. 6 is a schematical plan view of a portion of a semiconductor device according to this invention;

FIGS. 7A and 7B illustrate the connection of P and N polysilicon films in the device shown in FIG. 6;

FIG. 8 shows another way of connecting P and N polysilicon films in the device shown in FIG. 6;

FIGS. 9A to 9G show various connections between a P polysilicon layer and an N polysilicon layer, which reduce forward voltage across the polysilicon layers;

FIGS. 10A to 10C illustrate the forming of films of a metal silicide on the P polysilicon layers and N polysilicon layers of a semiconductor device according to this invention;

FIGS. 11A to 11C are cross sectional views taken along line 11A--11A in FIG. 10A, line 11B--11B in FIG. 10B and line 11C--11C in FIG. 10C; and

FIG. 12 is an electrode pattern of a logic circuit including a plurality of IIL gates.

FIG. 6 is a schematical plan view of a portion of a semiconductor device according to this invention, which comprises a plurality of IIL gates. In FIG. 6 only two IIL gates 56.sub.1 and 56.sub.2 are shown for simplicity's sake. A P layer 58.sub.1 and an N.sup.+ layer 60.sub.1 are respectively the base and collector of the inverter transistor of the IIL gate 56.sub.1. Similarly, a P layer 58.sub.2 and an N.sup.+ layer 60.sub.2 are respectively, the base and collector of the inverter transistor of the IIL gate 56.sub.2. P layers 62.sub.1 and 62.sub.2 are respectively the emitters of the injector transistors of the IIL gates 56.sub.1 and 56.sub.2. The base layers 58.sub.1 and 58.sub.2 are signal input terminals and the collector layers 60.sub.1 and 60.sub.2 are signal output terminals. The P layers 62.sub.1 and 62.sub.2 are generally called "injectors" and will hereinafter be called "injector P layers".

Further provided are electrode films 64.sub.1 and 64.sub.2 which are made of N polysilicon and electrode films 66.sub.1 and 66.sub.2 which are made of P polysilicon. The electrode films 64.sub.1 and 64.sub.2 contact the N.sup.+ layers 60.sub.1 and 60.sub.2 through contact holes 63.sub.1 and 63.sub.2, respectively. The electrode films 66.sub.1 and 66.sub.2 contact the P layers 58.sub.1 and 58.sub.2 through contact holes 65.sub.1 and 65.sub.2, respectively. Further, another electrode film 69 made of P polysilicon is provided. The film 69 contacts the injector P layers 62.sub.1 and 62.sub.2 through contact holes 68.sub.1 and 68.sub.2, respectively.

Between every semiconductor layer and every polysilicon film an insulation film of, for example, SiO.sub.2 is formed. The insulation film is not illustrated in FIG. 6 for simplicity's sake though.

To connect the N.sup.+ layer 60.sub.1 of the IIL gate 56.sub.1 to the P layer 58.sub.2 of the IIL gate 56.sub.2, the N polysilicon film 64.sub.1 and the P polysilicon film 66.sub.2 are connected to each other. These films 64.sub.1 and 66.sub.2 therefore form a PN junction. As mentioned above, in a PN junction of two IIL gates current flows always in the forward direction, never in the reverse direction. Likewise, in the PN junction formed of the polysilicon films 64.sub.1 and 66.sub.2 current flows in the forward direction as indicated by arrow B. Therefore, if the forward voltage drop across the PN junction is smaller than the base-emitter voltage of the inverter NPN transistor, the PN junction does not adversely affect the operation of the semiconductor device. Further, as experimentally known, the forward voltage across the PN junction is smaller than the forward voltage across a PN junction formed of two single-crystalline silicon films.

The semiconductor device of FIG. 6 has a small logic amplitude and an extremely small power consumption. Therefore, its electrode films, though made of polysilicon whose specific resistance is higher than that of aluminum, very little affects the operation of the device. Since its electrode films are made of polysilicon, the semiconductor device may have a multilayer structure, may provide a great lastitude of design and may therefore has its packing density improved.

The polysilicon films 64.sub.1 and 66.sub.2 may be connected in such a way as illustrated in FIG. 7A or FIG. 7B. In FIGS. 7A and 7B an N type substrate 72 and an oxide film 74 are shown. The oxide film 74 is made of, for example, SiO.sub.2 and formed on the N type substrate 72. To form the films 64.sub.1 and 66.sub.2 as shown in FIG. 7A, an intrinsic polysilicon film is formed on the oxide film 74. Then, phosphorus or arsenic is difussed into a portion of the intrinsic polysilicon film, thus obtaining the N polysilicon film 64.sub.1. Baron is then difussed into the remaining portion of the intrinsic polysilicon film, thus forming the P polysilicon film 66.sub.2. To form the films 64.sub.1 and 66.sub.2 as shown in FIG. 7B, the N polysilicon film 64.sub.1 is first formed on the oxide film 74, and then the P type polysilicon film 66.sub.2 is formed on the oxide film 74 with its end portion overlapping the N polysilicon film 64.sub.1.

In FIG. 7B, the P type polysilicon film 66.sub.2 is formed on the oxide film 74 with its end portion overlapping the N polysilicon film 64.sub.1. Conversely, the N type polysilicon film 64.sub.1 may be formed on the oxide film 74 with its end portion overlapping the P polysilicon film 66.sub.2.

In the embodiment of FIG. 6, the N polysilicon film 64.sub.1 and the P polysilicon film 66.sub.2 have substantially the same length and extend for some distance from the N.sup.+ layer 60.sub.1 and the P layer 58.sub.2, respectively. Instead, as shown in FIG. 8, the N polysilicon film 64.sub.1 may be formed only on the N.sup.+ layer 60.sub.1, and the P polysilicon film 66.sub.2 may be made long enough to contact the N polysilicon film 64.sub.1. Conversely, the P polysilicon film 66.sub.2 may be formed only on the P layer 58.sub.2, and the N polysilicon film 64.sub.1 may be made long enough to contact the P polysilicon film 66.sub.2.

In the embodiment of FIG. 6, the film 69 which connects the injector P layers 62.sub.1 and 62.sub.2 is made of P polysilicon. Instead, it may be made of aluminum.

To make the semiconductor device stably function, it is desired that the forward voltage across the PN junction of the polysilicon films 64.sub.1 and 66.sub.2 is as small as possible. The forward voltage can be reduced by connecting the polysilicon films 64.sub.1 and 66.sub.2 as shown in FIGS. 9A to 9G. More specifically, as shown in FIG. 9A, the abutting end portions of the polysilicon films 64.sub.1 and 66.sub.2 are doped with more impurities than the remaining portions, thus forming an N.sup.+ region 75 and a P.sup.+ region 76. The N.sup.+ region 75 and the P.sup.+ region 76 thus provided constitute a tunnel diode in the PN junction. Alternatively, as shown in FIG. 9B, between the N polisilicon film 64.sub.1 and the P polysilicon film 66.sub.2 an intrisic polysilicon region 77 is interposed, whereby the films 64.sub.1 and 66.sub.2 and the region 76 constitute a PIN diode. Further, as shown in FIG. 9C, Ne ions, for example, may be implanted into the PN junction of the polysilicon films 64.sub.1 and 66.sub.2, thus forming an amorphous layer 78 and thus destroying the PN junction. As illustrated in FIG. 9D, the PN junction of the polysilicon films 64.sub.1 and 66.sub.2 may be covered with a metal silicide film 80 broader than the films 64.sub.1 and 66.sub.2, thereby providing a short circuit in the PN junction. The film 80 may be narrower than the films 64.sub.1 and 66.sub.2 as shown in FIG. 9E or as broad as the films 64.sub.1 and 66.sub.2 as shown in FIG. 9F. The metal silicide film 80 can be formed by laying a metal film having a high melting point, such as platinum film, molybdenum film, tungsten film, titanium film and tantalum film, on the PN junction and then heating the polysilicon films 64.sub.1 and 66.sub.2 and the metal film at 500.degree. to 600.degree. C. Alternatively, the metal silicide film 80 may be provided by sputtering metal silicade having a high melting point onto the PN junction. Instead of such a metal silicide film, an aluminum film or other metal films may be used to shortcircuit the PN junction. Still further, as shown in FIG. 9G, the polysilicon films 64.sub.1 and 66.sub.2 may be arranged with a gap between them, and they are connected by means of a metal silicide film 82 or an aluminum film 82. The film 82 may be broader than the films 64.sub.1 and 66.sub.2 as shown in FIG. 9G. Alternatively, it may be as broad as the films 64.sub.1 and 66.sub.2 or narrower than these film.

Such a tunnel diode as shown in FIG. 9A and such a PIN diode as shown in FIG. 9B effectively reduce the forward voltage across the PN junction particularly when the P polysilicon film 66.sub.2 overlaps the N polysilicon film 64.sub.1 as illustrated in FIG. 7B and FIG. 8.

If platinum is used in order to provide a short circuit in the PN junctions of the N polysilicon films and the P polysilicon films, only that portion of the platinum film which is laid on the oxide film 74 can be successfully removed by aqua reia boiling process. This makes it possible to form platinum silicide films exclusively on the polysilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69, without a mask-aligning process. In practice, the platinum silicide films are formed in the following way.

First, such a circuit as shown in FIG. 10A is prepared. The circuit consists of a semiconductor substrate and an oxide film 74, N polysilicon films 64.sub.1 and 64.sub.2, P polysilicon films 66.sub.1 and 66.sub.2 and a P polysilicon film 69; all formed on the substrate. Then, as shown in FIG. 10B, a platinum film 86 is formed on the entire surface of the circuit board. Thereafter, the circuit board with the film 86 formed on it is heated at a temperature of 450.degree. C. to 800.degree. C. until platinum silicide films are formed on the polysilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69. The circuit board is then put into a aqua regia bath and boiled until that portion of the platinum film which lies on the oxide film 74 is removed. As a result, as shown in FIG. 10C, platinum silicide films 88.sub.1, 88.sub.2, 88.sub.3, 88.sub.4 and 88.sub.5 are formed respectively on the polysilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69, in self-alignment process.

FIGS. 11A, 11B and 11C are cross sectional views taken along line 11A--11A in FIG. 10A, line 11B--11B in FIG. 10B and line 11C--11C in FIG. 10C, respectively. Since the platinum silicide films 88.sub.1 to 88.sub.5 are self-aligned respectively with the polysilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69 as shown in FIG. 11C. Thus, it is easy to provide a short circuit in the PN junctions of the P polysilicon films and the N polysilicon films.

Further, since the platinum silicide films 88.sub.1 to 88.sub.5 are formed on the entire surfaces of the polisilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69, the these films have the sheet resistance about ten times lower than that of an only polisilicon film.

If molybdenum, tungsten, titanium or tentalum is used instead of plutinum in order to provide a short circuit in the PN junctions of the N polysilicon films and the P polysilicon films, molybdenum silicide films are formed on the polysilicon films 64.sub.1, 64.sub.2, 66.sub.1, 66.sub.2 and 69 substantially in the same way as illustrated in FIGS. 10A, 10B and 10C and FIGS. 11A, 11B and 11C.

In the above-described embodiment, the electrode film 64.sub.1 which contacts the N layer 60.sub.1 (i.e. signal output terminal) is made of N polysilicon, and the electrode film 66.sub.2 which contacts the P layer 58.sub.2 (i.e. signal input terminal) is made of P polysilicon. Instead, both electrode films 64.sub.1 and 66.sub.2 may be made of polysilicon of the same conductivity type, either P type or N type. If both films 64.sub.1 and 66.sub.2 are made of P polysilicon, they form at the N layer 60.sub.1 a PN junction which consists of P polysilicon and N single-crystalline silicon. To reduce the forward voltage across this PN junction, it is sufficient to form in the PN junction such a tunnel diode as shown in FIG. 9A or such a PIN diode as shown in FIG. 9B. On the other hand, if both films 64.sub.1 and 66.sub.2 are made of N polysilicon film, they form at the P layer 58.sub.2 a PN junction consisting of P single-crystalline silicon and N polysilicon. FIG. 12 is an electrode pattern of a logic circuit which comprises a semiconductor substrate, a plurality of IIL gates formed in the substrate and polysilicon films formed on the substrate in multilayer process. For simplicity's sake only five IIL gates 92.sub.1, 92.sub.2, 92.sub.3, 92.sub.4 and 92.sub.5 are illustrated. P polysilicon films 94.sub.1, 94.sub.2, 94.sub.3 and 94.sub.4 are formed on the substrate. N polysilicon films 96.sub.1, 96.sub.2, 96.sub.3 and 96.sub.4 are formed also on the substrate. Further, an N polysilicon film 98 is formed connecting to the N polysilicon film 96.sub.2. Still further, aluminum films 100.sub.1, 100.sub.2 and 100.sub.3 are provided. The aluminum film 100.sub.1 connects to the N polysilicon film 96.sub.2, the aluminum film 100.sub.2 connects to the P polysilicon film 94.sub.3 and the N polysilicon film 96.sub.1. An insulation film is provided between the N polysilicon film 98 on one hand and the polysilicon films 94.sub.1 to 94.sub.4 and 96.sub.1 to 96.sub.4 on the other hand. Another insulation film is provided between the N polysilicon film 98 on one hand and the aluminum films 100.sub.1 to 100.sub.3 on the other hand. These insulation films are not shown in FIG. 12 for simplicity's sake.

The N polysilicon film 94.sub.2 overlaps the P polysilicon film 96.sub.3 to form a PN junction 102 in such a manner as shown in FIG. 8. The N polysilicon film 94.sub.2 and the P polysilicon film 96.sub.4 are connected to form a PN junction 104 in such a manner as shown in FIG. 7A or FIG. 7B. At the input terminal in the IIL gate 92.sub.1, the PN junction consists of N polysilicon film 96.sub.2 and diffused P conductivity single crystalline silicon. Through these PN junctions current flows always in the forward direction. The forward votlage across the PN junctions is small and will not much affect the operation of the logic circuit. To reduce the forward voltage across the PN junctions and thus to make the logic circuit operate stably, a tunnel diode or a PIN diode may be formed in these PN junctions.

Since most of the electrode films are made of polysilicon, they are easily formed one upon another, thus providing a multilayer structure. The electrode films therefore occupy a smaller area than otherwise, whereby the packing density of the logic circuit is increased. The multilayer structure of electrode films is advantageous because various logic circuits may be manufactured merely by arranging the electrode films in various patterns, without changing the arrangement of the IIL gates.

As mentioned above, this invention uses polysilicon to form the electrode films of a semiconductor device including a plurality of bipolar transistors such as IIL gates, wherein current flows through the electrode films always from P semiconductor layers to N semiconductor layers. Its electrode films being made of polysilicon, the semiconductor device has its packing density increased and can be designed more easily than otherwise.

All the embodiments mentioned above are semiconductor devices comprising IIL gates. This does not mean that the invention is limited to IIL gates. This invention may be applied to other types of semiconductor devices comprising bipolar transistors if the forward voltage across the PN junctions of P and N polysilicon films and the voltage drop in each polysilicon film are reduced to proper values.