Semiconductor Integrated Circuit Device

The present invention relates to a semiconductor integrated circuit device, and more particularly to a structure of a flip-flop constituting a memory cell in a memory device.

The memory cell of Static RAM type comprises a pair of driver transistors, a pair of transfer gate transistors coupled to the driver transistors and digit lines and a pair of load elements coupled to a power supply line and the driver transistors. The driver transistors constitute a flip-flop circuit. When these transistors are insulated gate type field effect transistors (hereinafter referred to IGFET), a drain region of a first driver IGFET is connected to a gate electrode of a second driver IGFET so as to obtain a first node while a drain region of the second driver IGFET is connected to a gate electrode of the first driver IGFET so as to obtain a second node. This circuit is used in a state wherein, depending on whether each of the first and second driver IGFET's is an n-channel type or a p-channel type, the common source region is connected to a low-potential side or a high-potential side and the two nodes are separately connected to a power source either on a high-potential side or on a low-potential side through load elements.

There has recently been a strong demand for the provision of a large-scale, highly-integrated type of semiconductor integrated circuit by developments in minute processing techniques. In these circumstances, however, if the flip-flop structure which relies upon the conventional structure is to be miniaturized, it involves the disadvantage of losing resistance to the alteration of stored information which is caused by natural radiation and the rays radiating from unstable radioisotopes contained in package material, a so-called soft error. Specifically, soft error is the phenomenon by which the above-mentioned radiation mainly composed of .alpha. particles causes ionization in the nodes of each flip-flop, the substrate and so forth, and excessive electrical charges are thus produced to temporarily change the potential of each node, thereby causing the state of the flip-flop to be inverted. This phenomenon frequently occurs when, as a result of miniaturization, an impurity diffused layer and so forth are made small and the electrical capacitance of each node is thus reduced since the above-mentioned electrical charges are recombined and each node potential thus easily fluctuates. In order to solve this disadvantage, it is preferable to prevent any decrease in the electrical capacitance of each node while satisfying the need for miniaturization. To this end, what might first be considered is to enhance the impurity concentration in the semiconductor substrate so as to increase the junction capacitance of the drain region of the driver transistor which forms each node. However, in this case, the transfer gate IGFET's are also formed in the same semiconductor substrate of high impurity concentration, and therefore, the electrical capacitance of source or drain region of the transfer gate IGFET and that of the digit line coupled to a plurality of the transfer gate IGFET's are enhanced. Consequently, the switching or operation speed for reading and writing is delayed. Another method which might be considered is to locate close together the diffused drain regions of the pair of the driver IGFET's constituting the nodes of the IGFET's which form flip-flops, so that their capacitance can be mutually increased. However, in a memory device having flip-flops which are arranged at a high density, a pattern providing interconnection, for example, an aluminum pattern, is normally formed over the structure with an insulating layer film interposed therebetween by a deposition process, reducing the threshold voltage of a parasitic transistor formed between the impurity diffused drain regions of the driver IGFET's as source and drain. Therefore, it is impossible to allow to come close enough to the drain regions to affect an increase in the electrical capacitance of each node.

Accordingly, it is an object of the present invention to provide a semiconductor integrated circuit device such as SRAM type memory device having an effective flip-flop structure which has nodes of a sufficiently large electrical capacitance and is free from soft errors while satisfying the two great demands of the miniaturization required for a high degree of integration and an improvement in the operating speed resulting from the miniaturization.

According to a feature of the present invention, there is provided a semiconductor integrated circuit device comprising a first IGFET having a source region, a drain region and a gate electrode, and a second IGFET having a source region, a drain region and a gate electrode, the gate electrode of the first IGFET being connected to one of the source and drain regions of the second IGFET, the gate electrode of the second IGFET being connected to one of the source and drain regions of the first IGFET, and a part of the one of the source and drain regions of the first IGFET and a part of the one of the source and drain regions of the second IGFET being overlapped with an insulating film being interposed therebetween to form a capacitor element.

The insulating film of the capacitor element, that is, the so-called MOS type capacitor may be a silicon oxide film, a silicon nitride film, a aluminum oxide film, etc. or a multi-layered film of two or more these films. In view of reliability of the device and the effect of the present invention, that is, an enhancement of the capacitance at the node, the thickness of the insulating film using as a dielectric film of the capacitor element ranges from 200 .ANG. to 500 .ANG. in case of silicon oxide and from 200 .ANG. to 1000 .ANG. in case of silicon nitride which has a higher dielectric constant (.epsilon.) than that of silicon oxide. The above-mentioned flip-flop structure may be used in a static type RAM cell in a memory device.

FIG. 1 is a circuit diagram of a conventional static RAM type memory cell;

FIG. 2A is a plan view showing a conventional flip-flop structure used in FIG. 1, and FIG. 2B is a cross-sectional view taken along B--B in FIG. 2A as viewed in the direction of arrows;

FIG. 3 is a circuit diagram of a static RAM type memory cell according to an embodiment of the present invention;

FIG. 4A is a plan view showing the flip-flop structure of an embodiment of the present invention used in FIG. 3, and FIG. 4B, FIG. 4C and FIG. 4D are cross-sectional views taken along B--B, C--C and D--D (D'--D') in FIG. 4A as viewed in the direction of arrows, respectively.

Referring to FIG. 1, the basic arrangement of the memory cell is disclosed. A drain region 8 of a first driver IGFET 13 is connected to a gate electrode 6 of a second driver IGFET 14 so as to obtain a first node 3 while a drain region 7 of the second driver IGFET is connected to a gate electrode 5 of the first driver IGFET so as to obtain a second node 4. A source region 10 of the IGFET 13 is connected to a ground line with a source region 9 of of the IGFET 14 through a wiring layer 20. When the flip-flop circuit is used in the memory cell, the nodes 3, 4 are connected to a pair of digit lines D, D through transfer gate IGFET's 15, respectively, and connected to a power supply line Vcc through load elements 16 such as load IGFET's or load resistors, respectively. The gate electrodes of the transfer gate IGFET's 15 are commonly connected to a word line W. In the memory device, a plurality of memory cells are provided and connected to respective word lines, pairs of digit lines and power supply lines.

The structure of the flip-flop circuit encircled by a phantom line 100 in FIG. 1 is shown in FIGS. 2A and 2B. Elongated active regions 1, 2, that is, IGFET forming regions are provided in a major surface of a P-type silicon substrate 11 having an impurity concentration of 10.sup.15 to 10.sup.16 atoms/cm.sup.3 and surrounded by a thick field oxide layer 12. N-type source and drain regions 7, 8, 9, 10 of the first and second driver IGFET's 13, 14 having an impurity concentration of 10.sup.20 atoms/cm.sup.3 and depth of 0.3 .mu.m are respectively formed in the first and second active regions 1, 2. The gate electrode 5 of N-type polycrystalline silicon is provided on a channel region 13' of the first IGFET 13 via a thin gate insulating film, extending on the thick field insulating layer 12 and directly connected to the drain region 7 of the second IGFET 14. Also, the gate electrode 6 of N-type polycrystalline silicon is provided on a channel region 14' of the second IGFET 14 via a thin gate insulating film, extending on the thick field insulating layer 12 and directly connected to the drain region 8 of the first IGFET 13. After forming the structure shown in FIG. 2B, a passivation film (not shown) is entirely deposited and the source regions 9, 10 of the IGFET's are commonly connected by a metallic layer 20 (FIG. 2A) through respective contact holes formed in the passivation layer. In such a prior art flip-flop structure, the nodes 3, 4 have only the PN junction capacitance formed between respective N-type drain regions 8, 7 and the P-type substrate. In this case, if the length l.sub.1 and the width l.sub.2 of each drain region are 5.5 .mu.m and 1.5 .mu.m, respectively, each PN junction capacitance C.sub.1 (FIG. 1) of the node 3, 4 becomes about 7.5 fF (7.5.times.10 farad). To protect better against the soft errors mentioned above, the capacitance of the nodes must be higher.

Referring to FIGS. 3 and 4, a new capacitor C.sub.2 of the present invention is provided. In these FIGS. 3 and 4, components like those of FIGS. 1 and 2 have been identified with like numerals or symbols.

In the embodiment, an N-type impurity region 38", that is, the buried diffused layer having an impurity concentration of 10.sup.20 atoms/cm.sup.3 is formed in a P-type silicon monocrystalline body 61 having an impurity concentration of 10.sup.15 to 10.sup.16 atoms/cm.sup.3, and a thin silicon oxide film 70 of 300 .ANG. thickness is selectively formed on a portion 68 of the impurity region 38" with a broader width than that of the portion 68 as shown in FIG. 4B by a thermal oxidation process followed by a photo lithography process. Next, a P-type silicon layer 62 having an impurity concentration of 10.sup.15 to 10.sup.16 atoms/cm.sup.3 is grown on the silicon body 61 having the impurity region 38" and the silicon oxide film 70. In this case, since the portion covered by the thin film 70 of silicon dioxide is sufficiently small, the single-crystallization of the entire layer 62 is enabled by using as a seed the surface of the single-crystal silicon body 61 whose majority portions are exposed. The silicon body 61 and the silicon layer 62 constitute a silicon substrate 60. The surface of the single-crystal silicon layer 62 is subjected to a selective oxidation process so as to obtain a thick field silicon oxide layer 12, and elongated active regions 50, 51, 52, that is, IGFET forming regions abutted against the field silicon oxide layer 12 and separated and provided in parallel with each other by the field silicon oxide layer 12. As shown in FIG. 4A, in the embodiment each active region and the buried impurity region 38" have a uniform width to form a rectangular shape in the plan view. Subsequently, thermal oxidation is effected to the substrate surface whose actice regions are exposed, thereby forming thin silicon oxide films 53 including gate insulating films 53'. After holes 69 are so formed as to provide buried contacts, gate electrodes 35 and 36 are formed by the deposition of polycrystalline silicon containing an N-type impurity, that is, of conductivity type opposite to that of the substrate and by preforming a selective etching process. An of N-type impurity is introduced into the active regions 50, 51, 52 by using the gate electrodes 35, 36 and the thick field insulating layer 12 as a mask by ion implantation, and thermal treatment is effected so as to form N-type impurity regions 40, 38', 38"', 39, and 37 having an impurity concentration of 10.sup.20 atoms/cm.sup.3. The parts of the silicon layer 62 directly contacted with the N-type silicon gate electrode in the holes 69 are converted to N-type when the N-type polycrystalline silicon layer is deposited and/or when the heat treatment is carried out after the ion implantation by diffusing N-type impurity through the polycrystalline silicon gate, and the other parts except the channel regions are converted to N-type by ion implanting an N-type ion such as arsenic through the thin insulating film 53 on which them to gate electrodes do not exist and by subjecting the heat treatment for activation of the implanted ions. According to the process steps mentioned above, the N-type impurity regions 38', 38"' are contacted at their bottoms to the N-type buried region 38", and the N-type impurity region 37 is contacted at its bottom to the upper surface of the thin silicon oxide film 70. In the embodiment, the first driver IGFET 63 comprises an N-type drain region 38 including the N-type impurity regions 38', 38", 38"', the N-type source region 40, the gate insulating film 53' on a channel region 63' in the active region 50, and the gate electrode 35. The second driver IGFET 64 comprises the N-type drain region 37, the N-type source region 39, the gate insulating film 53' on a channel region 64' in the active region 51, and the gate electrode 36. The gate electrode 35 of the first IGFET 63 extends on the field insulating layer 12 and is directly connected to the drain region 37 of the second IGFET 64, through the hole 69 in the film 53, and also, the gate electrode 36 extends on the field insulating layer 12 and is directly connected to the impurity region 38"' of the drain region 38 of the first driver IGFET 63 through the hole 69 in the film 53. In the embodiment, the impurity regions and/or the thick field insulating layer 12 may be formed into the body 61.

According to the embodiment, a portion 68 of the drain region 38" of the first driver IGFET 63, the thin silicon oxide film 70 and a portion 67 of the drain region 37 of the second driver IGFET 64 constitute a capacitor element, that is, a so-called MOS capacitor by using the region 37 as the upper electrode, the region 38" as the lower electrode and the thin film 70 as the dielectric film. When the length l.sub.1 and the width l.sub.2 of the drain region 37 of the second IGFET 64 are 5.5 .mu.m and 1.5 .mu.m, respectively, and the width l.sub.3 of the drain region 38" of the first IGFET 63 is 1.5 .mu.m, at the node 34, the PN capacitor C.sub.1 becomes about 6.3 fF, and the MOS capacitor C.sub.2 of about 2.6 fF can be added. Then, the total capacitance at the node 34 becomes 8.9 fF which is about 20% higher than that of the prior art structure shown in FIG. 2 when they are compared in the same conditions of the dimensions and the impurity concentrations. Further, the prior art structure depends only on the PN junction capacitor and the values mentioned above are for the state in which no positive potential is applied to the drain regions. In the state of the positive potential such as +5 volt being applied to the drain region, the PN junction capacitance becomes about one half of the value mentioned above. In this case, the existance of the MOS capacitor C.sub.2 of the present invention is important. Moreover, if the material of the film 70 is changed from silicon oxide to silicon nitride, the capacity of the MOS capacitor C.sub.2 is enhanced to about two times of the value of silicon oxide with the same thickness. Further, according to the present invention, to increase the capacitance of the node, only the portion 68 in the impurity region 38" and the portion 67 in the impurity region 37 which contribute to form the MOS capacitor C.sub.2 can be enlarged without enlarging other portions in the impurity regions, that is, without sacrificing the integration of the device.

It is a matter of course that this invention does not depend on whether the IGFET's are of N-channel type or P-channel type and the gate electrode is not necessarily made of polycrystalline silicon. For example, the gate electrode may be a compound of tungsten, titanium, molybdenum and silicon and may also have a double layer of polycrystalline silicon and this compound.