Semiconductor Integrated Circuit Device

Abstract

Prior known static random access memory (SRAM) cells are required that a diffusion layer be bent into a key-like shape in order to make electrical contact with a substrate with a P-type well region formed therein, which would result in a decrease in asymmetry leading to occurrence of a problem as to the difficulty in micro-patterning. To avoid this problem, the P-type well region in which an inverter making up an SRAM cell is formed is subdivided into two portions, which are disposed on the opposite sides of an N-type well region NW1 and are formed so that a diffusion layer forming a transistor has no curvature while causing the layout direction to run in a direction parallel to well boundary lines and bit lines. At intermediate locations of an array, regions for use in supplying power to the substrate are formed in parallel to word lines in such a manner that one regions is provided per group of thirty two memory cell rows or sixty four cell rows.

Description

BACKGROUND OF THE INVENTION

[0001] The present invention relates generally to semiconductor integrated circuit devices and, more particularly, to layout schemes of static random access memory (SRAM) cells. The invention also relates to semiconductor memory devices using such cells.

[0002] One-port SRAM cells with complementary metal oxide semiconductor (CMOS) configurations are typically designed so that each cell consists essentially of six separate transistors. An exemplary layout of such cells has been disclosed in, for example, JP-A-10-178110 (laid open on Jun. 30, 1998).

[0003] In the prior known SRAM cell layout, a semiconductive well region of P type conductivity with inverters formed therein is subdivided into two subregions, which are disposed on the opposite sides of an N-type well region while permitting a well boundary line to extend in a direction parallel to bit lines.

[0004] The quest for higher integration and ultra-fine patterning techniques in modern memory devices requires optical exposure apparatus or equipment to decrease in wave length of beams used therein. To this end, the equipment is designed to employ exposure beams of shorter wavelength, which have advanced from G line to I line, and further to excimer laser. Unfortunately the requirements for micro-patterning architectures grows more rapidly than technological advance in trend of shortening wavelengths in such equipment. In recent years, it is strictly required that micropatterning is done with the minimum device-feature length that shrinks to less than or equal to the wavelength of an exposure beam used. This minimum feature length shrinkage would result in the layout of IC components--here, memory cells--becoming more complicated in planar shape, which must require the use of irregular polygonal layout patterns including key-shaped components, in order to achieve the intended configuration of on-chip circuitry with enhanced accuracy. This makes it impossible or at least very difficult to microfabricate ultrafine layout patterns while disadvantageously serving as the cause of destruction of the symmetry of memory cells.

[0005] Regrettably the prior art approach is associated with a need to curve or bend a diffusion layer into a complicated key-like shape for the purpose of making electrical contact with a substrate of the P-type well region. Thus, the prior art suffers from a problem as to degradation of the symmetrization of cell layout pattern, making difficult successful achievement of microfabrication architectures for higher integration densities.

SUMMARY OF THE INVENTION

[0006] In accordance with one aspect of the present invention, a semiconductor device is provided which comprises a first inverter including a first N-channel metal oxide semiconductor (MOS) transistor and a first P-channel MOS transistor, a second inverter including a second N-channel MOS transistor and a second P-channel MOS transistor with an input terminal being connected to an output terminal of the first inverter and with an output terminal being connected to an input terminal of said first inverter, a third N-channel MOS transistor having a source connected to the output terminal of said first inverter and a drain connected to a first bit line and also a gate connected to a word line, and a fourth N-channel MOS transistor having a source connected to the output terminal of said second inverter and a drain connected to a second bit line plus a gate connected to a word line, wherein the first and third N-channel MOS transistors are formed in a first P-type well region, wherein the diffusion layer has no curved or bent portions while letting the direction of layout be parallel to the boundary with respect to the first N-well region with the first and second P-channel MOS transistors formed therein, and wherein said second and fourth N-channel MOS transistors are formed in the second P-type well region whose diffusion layer has no bent portions while letting the layout direction be parallel to the boundary with respect to the first N-well region with the first and second P-channel MOS transistors formed therein.

[0007] The diffusion layer is arranged to have its outer shape that mainly consists of straight line segments including the longest straight line portion which lies parallel to the boundary with respect to the first n-well region with the first and second P-channel MOS transistors formed therein, and simultaneously in the case of defining a straight line acting as the center line extending parallel to such boundary, the longest line portion is in linear symmetry with such center line; the second and fourth N-channel MOS transistors are formed in the second P-well region whose diffusion layer is mainly arranged by straight line segments including its longest straight line portion that is parallel to the boundary with respect to the first n-well region with the first and second P-channel MOS transistors formed therein while allowing, when defining a straight line for use as the center line extending parallel to such boundary, the line portion to be linearly symmetrical with the center line. At this time, in the case of employing the linear symmetrization scheme, complete linear symmetry will not always be required; alternatively, slight nonsymmetry may also be permissible on a case-by-case basis, which nonsymmetry results from modifying the diffusion layer to have a shape with its portions on the right and left sides of the center line being substantially the same in area as each other by way of example.

[0008] In accordance with another aspect of this invention, a first polycrystalline silicon lead layer for use as the gate of said third N-channel MOS transistor and a second polycrystalline silicon lead layer for use as the gate of said first P-channel MOS transistor and also as the gate of said first N-channel MOS transistor are disposed in parallel to each other, wherein a third polycrystalline silicon lead layer for use as the gate of said fourth N-channel MOS transistor and a fourth polycrystalline silicon lead layer for use as the gate of said second N-channel MOS transistor and also as the gate of said second P-channel MOS transistor are disposed in parallel to each other, and wherein the first and third polycrystalline silicon lead layers are connected via a contact to a second layer of metal lead layer constituting word lines.

[0009] In accordance with a further aspect of the invention, the input terminal of said first inverter and the output terminal of said second inverter may be electrically connected together at a contact whereas the input terminal of said second inverter and the output terminal of said first inverter are electrically connected together at a contact.

[0010] In accordance with yet another further aspect of the invention, a power supply line connected to the first and second bit lines and the sources of said first and second P-channel MOS transistors and a ground line connected to the sources of said first and second N-channel MOS transistors may be formed of a third layer of metal lead layer lying parallel to a diffusion layer.

[0011] In accordance with a still another aspect of the invention, the first bit line formed of said third layer of metal lead layer may be arranged so that it is between a power supply line formed of said third layer of metal lead layer and a ground line as connected to the source of said first N-channel MOS transistor formed of said third layer of metal lead layer whereas the second bit line formed of said third layer of metal lead layer is between a power supply line formed of said third layer of metal lead layer and a ground line as connected to the source of said second N-channel MOS transistor formed of said third layer of metal lead layer.

[0012] In accordance with another further aspect of the invention, the first and second bit lines and a power supply line connected to the sources of said first and second P-channel MOS transistors may be formed of a second layer of metal lead layer, wherein word lines are formed of a third layer of metal lead layer, and wherein a ground line connected to the sources of said first and second N-channel MOS transistors is formed of the third layer and second layer of metal lead layer.

[0013] In accordance with a still another further aspect of the invention, memory cells are laid out into the form of an array, wherein contacts to a substrate of P-type well region and a contact to a substrate of N-type well region are linearly disposed within the array and at upper and lower portions of the array in a direction parallel to the word lines. Although the above is an example which causes two separate P-well to be disposed on the opposite sides of an N-well region, two N-well regions may be disposed on the opposite sides of a p-well region when the need arises.

[0014] In accordance with yet another further aspect of the invention, a semiconductor device is provided which comprises a plurality of memory arrays each including an array of memory cells each having at least a pair of N-type well region and P-type well region, and at least one intermediate region between the memory arrays, wherein the N-type well region and P-type well region defines therebetween a boundary with at least one straight line portion, and wherein a diffusion layer is formed in each of the P-type well region and P-type well region to have a planar shape of either (1) a shape of rectangle having long sides extending parallel to said straight line portion or (2) a shape resulting from letting a plurality of rectangles having long sides extending parallel to the straight line portion be combined together via respective short sides thereof; or alternatively,

[0015] (1) a shape of rectangle having long sides parallel to said straight line portion or (2) a shape resulting from letting a plurality of rectangles having long sides parallel to said straight line portion be combined together causing them to extend in the direction of the straight line.

[0016] At least in memory array regions, bit lines are laid out in a direction parallel to the straight line portion whereas word lines are disposed in a direction perpendicular to the straight portion. Preferably, in the intermediate region, at least one type of electrical lead is railed in a direction at right angles to the straight portion, and a lead (e.g. contact) is also formed which is for making electrical contact between a power supply voltage lead and the diffusion layer as formed in the N-well region or P-well region. This lead may include a power supply lead, ground lead, or other potential leads.

[0017] The invention is particularly useful for those semiconductor memory devices having static RAM memory cells each consisting essentially of six separate transistors.

BRIEF DESCRIPTION OF THE DRAWINGS

[0018] FIG. 1 is a diagram showing an SRAM cell in accordance with Embodiment 1 of the present invention, for explanation of a layout pattern of those contacts for connection between MOS transistors and those for connecting between MOS transistors and metal lead layers.

[0019] FIG. 2 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 1 of this invention.

[0020] FIG. 3 is a diagram showing a layout of memory cells and their associated peripheral circuitry in accordance with Embodiment 2 of the invention.

[0021] FIG. 4 is a diagram showing an SRAM cell in accordance with Embodiment 3 of the invention, for explanation of a layout of those contacts for connection between MOS transistors and those for connection between MOS transistor and metal lead layers.

[0022] FIG. 5 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 3 of the invention.

[0023] FIG. 6 is a diagram showing an SRAM cell in accordance with Embodiment 4 of the invention, for explanation of a layout of those contacts for connection between MOS transistors and those for connection between MOS transistors and metal lead layers.

[0024] FIG. 7 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 3 of the invention.

[0025] FIG. 8 is a diagram showing an SRAM cell in accordance with Embodiment 5 of the invention, for explanation of a layout of those contacts for connection between MOS transistors and those for connection between MOS transistors and metal lead layers.

[0026] FIG. 9 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 5 of the invention.

[0027] FIG. 10 is a diagram showing an SRAM cell in accordance with Embodiment 6 of the invention, for explanation of a layout of those contacts for for connection between MOS transistors and those for connection between MOS transistors and metal lead layers.

[0028] FIG. 11 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 6 of the invention.

[0029] FIGS. 12a to 12f are diagrams illustrating in cross-section some of major process steps in the manufacture of the semiconductor device in accordance with Embodiment 6 of the invention.

[0030] FIG. 13 is a diagram showing an SRAM cell in accordance with Embodiment 7 of the invention, for explanation of a layout of those contacts for connection between MOS transistors and those for connection between MOS transistors and metal lead layers.

[0031] FIG. 14 is a diagram showing a layout of via holes of SRAM cells for use in connecting between multilayered metal leads in accordance with Embodiment 7 of the invention.

[0032] FIG. 15 is a diagram showing an SRAM cell in accordance with Embodiment 8 of the invention, for explanation of a layout of those contacts for connection between MOS transistors and those for connection between MOS transistors and metal lead layers.

[0033] FIG. 16 is a diagram showing a layout of via holes of SRAM cells for connection between multilayered metal leads in accordance with Embodiment 8 of the invention.

[0034] FIG. 17 is a sectional view of a semiconductor device in accordance with Embodiment 8 of the invention.

[0035] FIGS. 18a to 18f are diagrams illustrating in cross-section some of major process steps in the manufacture of a semiconductor device in accordance with Embodiment 9 of the invention.

[0036] FIGS. 19a to 19g are diagrams illustrating in cross-section some of major process steps in the manufacture of a semiconductor device in accordance with Embodiment 10 of the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

[0037] Several preferred embodiments of the semiconductor memory device in accordance with the present invention will be explained with reference to the accompanying drawings below.

Embodiment 1

[0038] Referring to FIGS. 1 and 2, there is shown an SRAM cell layout MC embodying the invention. FIG. 1 illustrates well regions and diffusion layers plus polycrystalline silicon interconnect lead layer as well as contacts, all of which are formed in or over a semiconductor substrate whereas FIG. 2 depicts a first layer of metal lead layer, via holes 1, second layer of metal lead layer, via holes 2, and a third layer of metal lead layer. Symbols used in FIGS. 1 and 2 are indicated at lower part of FIG. 2.

[0039] An N-channel type MOS transistor TN1 formed in a P-type semiconductive well region PW1 and a P-channel type MOS transistor TP1 formed in an N-type well region NW1 constitute an inverter INV1. In addition, an N-channel MOS transistor TN2 formed in P-type well region PW2 and a P-channel MOS transistor TP2 formed in N-type well region NW1 constitute an inverter INV2.

[0040] An output node of the inverter INV1 is electrically connected by a contact SC1 to an input node of the inverter INV2. An output of the inverter INV2 is electrically connected via a contact SC2 to an input of the inverter INV1.

[0041] An N-channel MOS transistor TN3 has a drain electrode connected to a bit line BL1, a source electrode connected to a drain of the N-channel MOS transistor TN1, and a gate electrode connected to a word line WD. Similarly an N-channel MOS transistor TN4 has a drain electrode connected to a bit line BL2, a source electrode connected to a drain of the N-channel MOS transistor TN2, and a gate electrode connected to word line WD.

[0042] The N-channel MOS transistor TN1 and N-channel MOS transistor TN3 are formed over a diffusion layer LN1 whereas the N-channel MOS transistor TN2 and N-channel MOS transistor TN4 are formed over a diffusion layer LN2. The P-channel MOS transistor TP1 is formed over a diffusion layer LP1 whereas the P-channel MOS transistor TP2 is formed over a diffusion layer LP2.

[0043] As the diffusion layers (LN1, LN2, LP1, LP2) are straight lines with no curved portions, any pattern correction at folded portions is no longer necessary, resulting in the balance between nodes being improved. In case memory cells are laid out into the form of an array, the diffusion layers become four separate straight lines extending parallel to the bit lines (BL1, BL2).

[0044] In addition, a polycrystalline silicon interconnect lead layer FG3 for use as the gate electrode of the N-channel MOS transistor TN3 and a polycrystalline silicon lead layer FG4 for use as the gate electrode of N-channel MOS transistor TN4 are connected to word lines WL which are formed of the second metal lead layer in a vertical direction to the bit lines (BL1, BL2). A polycrystalline silicon interconnect lead layer FG1 for use as the gate electrodes of the N-channel MOS transistor TN1 and P-channel MOS transistor TP1 and a polycrystalline silicon interconnect lead layer FG2 for use as the gate electrode of the N-channel MOS transistor TN2 and P-channel MOS transistor TP2 plus the polycrystalline silicon lead layers (FG3, FG4) are disposed in parallel to the word lines.

[0045] The N-channel MOS transistor TN1 has its source electrode connected to a ground potential line Vss1 that is formed of the third layer of metal lead layer whereas a source electrode of the N-channel MOS transistor TN2 is connected to a ground line Vss2 as formed of the third layer of metal lead layer. In addition, source electrodes of the P-channel MOS transistors (TP1, TP2) are connected to a power supply voltage line Vcc1 which is formed of the third layer of metal lead layer.

[0046] The bit line BL1 is located midway between the power supply voltage line Vcc1 and ground line Vss1 whereas bit line BL2 is between the supply voltage line Vcc1 and ground line Vss2. This structure makes it possible to reduce cross-couple noises occurring between bit lines, which advantageously lowers voltages while increasing operation speeds.

[0047] In addition, it is considered that, in case a contact is formed on an n.sup.- layer through partial cutaway of side spacers during etching of contact holes, a leakage current from the contact via the n.sup.- layer to the substrate may be produced. When a contact is formed for connection between a polycrystalline silicon lead layer and a diffusion layer, a distance between the diffusion layer LP2 and polycrystalline silicon lead layer FG1 should be greater than the length of a side spacer to thereby eliminate formation an n.sup.- layer on the polycrystalline silicon lead layer FG1 side of the diffusion layer LP2, which in turn makes it possible to prevent a flow of leakage current.

Embodiment 2

[0048] Turning to FIG. 3, there is shown an exemplary case where the memory cells MC of Embodiment 1 are laid out into the form of an array. Symbols used herein are the same as those indicated at lower part of FIG. 2.

[0049] The memory cells MC are organized into an array of 256 rows and 128 columns, by way of example. In view of the fact that these memory cells in Embodiment 1 are less in length in the longitudinal direction of bit lines, a total length of such 256 rows of memory cells along the bit lines is shorter than that of prior art devices, thus increasing resultant operation speeds. Neighboring memory cells MC are disposed in linear symmetry with respect to a "y" axis whereas upper and lower adjacent memory cells MC are in linear symmetry with an "x" axis. In addition, specified regions ST for use in supplying more than one power supply voltage to the substrate are formed at intermediate part of the array in such a manner that the regions ST extend parallel to word lines WD. One example is that the regions ST are laid out in units of 32-row groups. Another example is that regions ST are disposed in units of 64-row groups.

[0050] An electrical lead Vbn for supplying a voltage potential to the P-well regions (PW1, PW2) and a lead Vbp for supplying a voltage to the N-well region NW1 are formed to lie parallel to word lines. The lead Vbn may be coupled to ground potential Vss or, alternatively, any voltage may be applied thereto which is potentially different from ground Vss. The lead Vbp may be coupled to the power supply voltage Vcc or, alternatively, any voltages potentially different from Vcc may be applied thereto.

[0051] Note that in each region ST, a power supply voltage line Vcc for potentially "reinforcing" a power supply voltage line Vcc1 is formed in parallel to word lines while letting a ground potential line Vss for potentially reinforcing ground potentials (Vss1, Vss2) is formed in parallel to the word lines.

[0052] Also note that the ground lines (Vss1, Vss2) are disposed in a direction perpendicular to the word lines WD whereby upon selecting of a single word line a voltage potential is supplied from the pair of ground lines to a respective one of those memory cells operatively associated with this selected word line so that any possible noises occurring at such voltage lines are reduced to thereby advantageously speed up an access operation while potentially reducing any voltages concerned.

[0053] Furthermore, the memory cells MC used are great in width in the word line direction so that the layout design of sense amplifiers AMP is made easier to thereby avoid a need to lay out one sense amplifier for two adjacent columns of memory cells, which in turn makes it possible to permit one sense amplifier to be laid out at each column. Additionally a word line driver circuit wddrv becomes flat in layout as compared to prior known ones.

Embodiment 3

[0054] FIGS. 4 and 5 show a SRAM cell layout MC2 in accordance with Embodiment 3. Symbols as used in FIGS. 4-5 are the same as those in FIG. 2. Memory cell MC2 of Embodiment 3 is similar to the memory cell MC of Embodiment 1, except that whereas in Embodiment 1 the diffusion layer (LN1, LN2) is formed into a "T"-like planar shape, which resembles a Japanese battledore plate called "hagoita," the diffusion layer (LN3, LN4) of Embodiment 4 is of a rectangular shape, and that the contacts (SC1, SC2) are replaced with contacts (SC3, SC4) in the first layer of metal lead layers (M11, M12).

[0055] To attain stability, memory cells are typically designed so that the gate width of N-channel MOS transistors (TN1, TN2) is one and a half times greater than that of N-channel MOS transistors (TN3, TN4). However, in this case, the shape of diffusion layers resembles a T-like planar shape as has been shown in Embodiment 1, which in turn requires extra techniques including pattern correction procedures such as optical proximity effect correction (OPC) processes. Additionally this would result in degradation of the balance between transistors. In contrast, Embodiment 3 is such that the diffusion layers (LN3, LN4) are designed to have a rectangular shape whereby the micro-patterning required becomes easier while at the same time enabling improvement in balance between transistors. Note however that the resultant gate width ratio becomes as large as 1.0 time, which in turn requires that the so-called cell ratio be increased by making different drivabilities therebetween, which is achievable by letting the N-channel MOS transistors (TN3, TN4) be greater in oxide film thickness than N-channel MOS transistors (TN1, TN2), or by increasing the gate length thereof, or alternatively by increasing the threshold value, or still alternatively by lowering the impurity concentration of lightly-doped drain regions for relaxation of electric fields.

[0056] In addition, Embodiment 3 is arranged to employ a contact SC3 and a first layer of metal lead layer M11 in place of the contact SC1 used in Embodiment 1 for connection between the output of inverter INV1 and the input of inverter INV2. With such an arrangement, any curved or bent contacts are no longer necessary, thereby avoiding the need for pattern correction (OPC) or the like.

Embodiment 4

[0057] FIGS. 6 and 7 show an SRAM cell layout MC3 in accordance with an embodiment 4. Symbols as used in FIGS. 6-7 are the same as those in FIG. 2. Memory cell MC3 of Embodiment 4 is similar to the memory cell MC2 of Embodiment 3 except that polycrystalline silicon lead layers (FG5, FG6, FG7, FG8) are designed to have a rectangular planar shape. With this cell, any bent/folded portions are absent thus removing the need for any additional pattern correction procedures including OPC processes, which in turn improves the balance between transistors.

Embodiment 5

[0058] FIGS. 8 and 9 show an SRAM cell layout MC4 in accordance with Embodiment 5. An explanation on those symbols used in FIGS. 8 and 9 is given at lower part of FIG. 8. Memory cell MC4 of Embodiment 5 is different in lead structure from the memory cell MC of Embodiment 1.

[0059] Bit lines (BL3, BL4) and power supply line Vcc2 are formed by use of a second layer of metal lead layer. A word line WD1 and ground lines (Vss5, Vss6) are formed using a third layer of metal lead layer in a perpendicular direction to the bit lines. Ground lines (Vss3, Vss4) are formed using a fourth layer of metal lead layer in a direction parallel to the bit lines.

[0060] A global bit line GB is the electrical interconnect lead that is used in case bit lines are of a hierarchical configuration. The global bit line GB and bit lines (BL3, BL4) are shielded by the third layer of metal lead layer, thus enabling prevention of any possible cross-couple noises. Additionally the use of ground lines (Vss3, Vss4) makes it possible to prevent occurrence of cross-couple noises between global bit lines GB.

Embodiment 6

[0061] FIGS. 10 and 11 show an SRAM cell layout MC5 in accordance with an embodiment 6. An explanation as to those symbols used in FIGS. 10-11 is given at lower part of FIG. 11. Memory cell MC5 of Embodiment 6 is different from the memory cell MC of Embodiment 1 in structure of the so-called three-layered contacts, each of which is for connection between a gate electrode and its associated diffusion layer.

[0062] Although in Embodiment 1 a gate electrode is connected to a diffusion layer via "L"-like contacts SC1, SC2, Embodiment 6 is arranged so that the gate electrode is connected to the diffusion layer via silicide in connect regions SS1, SS2. This makes it unnecessary to bend or curve the individual contact into the L-like shape in order to connect the gate electrode to the diffusion layer, which in turn makes it possible to provide "I"-like rectangular contacts SC5, SC6. No folded/bent portions are present in the contacts used, which eliminates the need for pattern correction (OPC).

[0063] One practically implementable flow of some major process steps in the manufacture of a device structure employing the connect regions SS1, SS2 each for connection between a gate electrode and a diffusion layer associated therewith via silicide is shown in FIGS. 12a through 12f. Note here that FIGS. 12a-12f are cross-sectional views each indicating a profile as taken along line A-A' in FIG. 10, with its right side corresponding to the side "A" and with left side corresponding to "A'."

[0064] Fabricate a gate electrode FG made of a chosen polycrystalline silicon material (see FIG. 12a).

[0065] Form a heavily-doped diffusion layer PM of a specified conductivity type--here, P type (FIG. 12b).

[0066] Form side spacers made of silicon nitride (SiN) by chemical vapor deposition (CVD) techniques, on side-walls of the resultant gate electrode FG (FIG. 12c).

[0067] Make use of a resist RG to etch away only one of the SiN side spacers which resides on an active region side under a prespecified condition that enables etching treatment of a silicon nitride film and oxide film at increased selectivity (FIG. 12d).

[0068] Fabricate a heavily-doped P (P+) type diffusion layer P+.

[0069] After having removed through etching certain part of the oxide film SiO that lies in the active region that is not covered by any overlying gate electrode FG, deposit a high-melting-point metal such as refractory metal including, but not limited to, cobalt (Co); then, anneal the resultant structure to thereby selectively form silicide on the poly-silicon gate electrode and diffusion layer (FIG. 12f). At this time the gate electrode's side-wall and diffusion layer are connected together by such silicide.

Embodiment 7

[0070] FIGS. 13 and 14 show an SRAM cell layout MC6 in accordance with Embodiment 7. An explanation of those symbols used in FIGS. 13-14 is the same as that given at lower part of FIG. 11. Memory cell MC6 of Embodiment 7 is similar to the memory cell MC5 of Embodiment 6 with the contacts (SC5, SC6) being replaced with contacts (SC7, SC8) in the first layer of metal lead layers (M11, M12).

[0071] With Embodiment 7, all of the contacts used therein are capable of being designed to have a square planar shape, thus avoiding the need for pattern correction (OPC).

Embodiment 8

[0072] FIGS. 15 and 16 show an SRAM cell layout MC7 in accordance with Embodiment 8. An explanation of those symbols used in FIGS. 15-16 is given at lower part of FIG. 16. Memory cell MC7 of Embodiment 8 is similar to the memory cell MC of Embodiment 1 with the contacts (SC1, SC2) being replaced with local interconnect nodes (LI1, LI2) and also with the word lines being modified in such a manner that these are formed in the first layer of metal lead layer rather than in the second layer of metal lead layer while also modifying the bit lines and power supply and ground lines from the third layer of metal lead layer to the second layer of metal lead layer. FIG. 17 depicts a sectional view taken along line A-B of FIGS. 15-16.

[0073] Embodiment 1 suffers from limitations as to an inability to dispose the first layer of metal leads over the contacts SC1, SC2 due to the fact that these contacts SC1, SC2 are formed of the same layer as the remaining contacts used. In contrast, Embodiment 8 is specifically arranged to employ the local interconnect nodes LI1, LI2 formed in a separate layer from the contacts, thus making it possible to dispose the first layer of metal lead layer at upper part, which in turn makes it possible to reduce by one the requisite number of metal lead layers when compared to Embodiment 1.

Embodiment 9

[0074] A process flow of major steps in the manufacture of a three-layer contact section of Embodiment 9 is shown in FIGS. 18a-18f. This embodiment 9 is an example of the process for fabrication of the three-layer contact section as used in Embodiments 1, 3-5 and 8.

[0075] Modern LSIs in recent years are typically designed so that micropatterning is done to form contact holes by high-selectivity etching techniques with a silicon nitride film or else used as a stopper to ensure that any unwanted over-etching occurs at filed oxide films even when contacts are offset in position from diffusion layers and/or gate electrodes due to the presence of possible alignment errors during photolithographical patterning processes. In cases gate electrodes are formed to have reduced electrical resistivities by use of the so-called salicide processes, it is possible to obtain the intended electrical conduction between a contact lying over a gate electrode and a contact overlying a diffusion layer even when the both contacts are fabricated at a time because of the fact that the contact holes required are fabricated after completion of a procedure having the steps of forming silicide through exposure of selected portions overlying diffusion layers and gate electrodes after having formed such diffusion layers, depositing thereover a silicon nitride film for use as an etching stopper, and then further depositing thereover an interlayer dielectric film. On the contrary, in the case of either traditionally widely employed polycide gate electrodes or polymetal gate electrodes that have been developed and announced recently, residual portions of an insulative film such as oxide film can overlie gate electrodes thereby preventing exposure of these gate electrodes prior to deposition of a silicon nitride film acting as the etch stopper; accordingly, whenever an attempt is made to form the intended contacts through deposition of a silicon nitride film thereover, the oxide film behaves to partly reside at the bottom of a respective one of those contacts overlying the gate electrodes, which makes it impossible or at least greatly difficult to provide electrical conduction required. Embodiment 9 is for enabling achievement of electrical conduction of such contacts overlying gate electrodes by previous removal of any silicon nitride film portions overlying gate electrodes at specified part whereat contact holes will be defined.

[0076] An explanation will now be given of the process flow in the manufacture of Embodiment 9 with reference to FIGS. 18a-18f below.

[0077] After having fabricated a gate electrode and a diffusion layer P+, deposit a silicon nitride film SiN for use as an etch stopper (FIG. 18a). The gate electrode is a lamination of polycrystalline silicon PolySi and tungsten W, with an oxide film SiO being further multilayered thereon as a protective film.

[0078] Remove by dry etching techniques specified part of the silicon nitride film at locations for definition of a contact hole overlying the gate electrode (FIG. 18b).

[0079] Deposit a TEOS film and others by plasma CVD methods to thereby form an interlayer dielectric film (FIG. 18c).

[0080] Let a selected portion of the oxide film at a contact opening portion be etched away by high selective dry etching techniques (FIG. 18d). Owing to such high selective etching, the silicon nitride film remains free from etch treatment and thus acts as a stopper. Since there is no stopper at the portion overlying the gate electrode from which the silicon nitride film has been removed away in advance, such portion will be fully etched to the upper part of the gate electrode. This permits electrical conduction on the gate electrode also.

[0081] Remove the silicon nitride film by dry etching techniques (FIG. 18e).

Deposit a chosen metal such as tungsten in the resulting contact hole, thereby forming a buried plug (FIG. 18f).

Embodiment 10

[0082] Turning to FIGS. 19a-19g, there is shown a process flow in the manufacture of the three-layer contact section of Embodiment 10. Embodiment 10 is one example of the process for forming the three-layer contact section of Embodiments 1, 3-5 and 8.

[0083] A difference of the process flow of Embodiment 10 from that of Embodiment 9 is that more than one portion of the oxide film at a specified location whereat a contact hole is to be opened over the gate electrode has been removed in advance prior to deposition of a silicon nitride film for use as the etch stopper.

[0084] The fabrication process flow of Embodiment 10 will be explained with reference to FIGS. 19a-19g below.

[0085] Fabricate a gate electrode and a diffusion layer P+ (FIG. 19a). The gate electrode is a lamination of polycrystalline silicon PolySi and tungsten W, with an oxide film SiO further stacked thereon as a protective film.

[0086] Remove by dry etching techniques a specified part of the silicon nitride film at the location for definition of a contact hole overlying the gate electrode, thus letting the gate electrode be exposed at its upper part (FIG. 19b).

[0087] Deposit a silicon nitride film SiN as an etch stopper (FIG. 19c).

[0088] Deposit a TEOS film or else by plasma CVD methods to thereby form an interlayer dielectric film (FIG. 19d).

[0089] Let a portion of the oxide film at contact opening portion be etched away by high selective dry etching techniques (FIG. 19e). Due to such high selective etching, the silicon nitride film remains free from etching treatment and thus acts as the stopper.

Remove the silicon nitride film by dry etching techniques (FIG. 19f). A certain portion from which the oxide film overlying the gate electrode has been removed prior to deposition of the silicon nitride film is thus exposed at this time, which permits electrical conduction on the gate electrode also.

[0090] Deposit a chosen metal such as tungsten in the resultant contact hole, thereby forming a buried plug (FIG. 19g).

[0091] In accordance with the embodiments stated above, any diffusion layers used therein are specifically designed to have a simplified planar shape excluding unnecessarily complicated shapes, which may in turn facilitate micro-patterning processes.

Claims

1-29. (canceled)

30. A semiconductor memory device comprising: first and second bit lines extending in a first direction; a plurality of word lines extending in a second direction crossing the first direction; a memory array having a plurality of memory cells, each of the memory cells having a first inverter including a first N-channel MOS transistor and a first P-channel MOS transistor, a second inverter including a second N-channel MOS transistor and a second P-channel MOS transistor with an input terminal of the second inverter being coupled to an output terminal of the first inverter and with an output terminal of the second inverter being coupled to an input terminal of the first inverter, a third N-channel MOS transistor having a source/drain path coupled between the output terminal of the first inverter and the first bit line, and a fourth N-channel MOS transistor having a source/drain path coupled between the output terminal of the second inverter and the second bit line, a gate of each of the third and fourth N-channel MOS transistors being connected to one of the plurality of word lines; a first P-type well region in which the first and third N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the first P-type well region extending in the first direction; a second P-type well region in which the second and fourth N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the second P-type well region extending in the first direction; and an N-type well region in which the first and second P-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the N-type well region lying between the first and second P-type well regions; wherein with respect to a plan view of a principal plane of the semiconductor memory device, (i) each of the plurality of memory cells is formed in a rectangular region which includes the first to fourth N-channel MOS transistors, the first and second P-channel MOS transistors, a first contact coupled between the third N-channel MOS transistor and the first bit line, a second contact coupled between the fourth N-channel MOS transistor and the second bit line, a third contact coupled between one of the plurality of word lines and the gate of the third N-channel MOS transistor, and a fourth contact coupled between the one of the plurality of word lines and the gate of the fourth N-channel MOS transistor, (ii) the memory array has a first area which consists of plural rectangular regions arranged in the first direction, (iii) a second area in which a first well contact to the first P-type well region is arranged without being overlapped with the first area, (iv) a third area in which a second well contact to the first P-type well region is arranged without being overlapped with the first area, and wherein the first P-type well region in the first area is supplied with a first voltage from the second area and the third area, and the first area has no contact to supply the first voltage to the first P-type well region, wherein, in the plan view, no memory cell is formed in the second area and no memory cell is formed in the third area, and wherein a first well contact arranged in the second area and a second well contact arranged in the third area are disposed at upper and lower portions of the memory array, respectively.

31. The semiconductor memory device according to claim 30, wherein the memory array has (v) a fourth area in which a third well contact to the second P-type well region is arranged without being overlapped with the first area, and (vi) a fifth area in which a fourth well contact to the second P-type well region is arranged without being overlapped with the first area, wherein the second P-type well region in the first area is supplied with the first voltage from the fourth area and the fifth area, and the first area has no contact to supply the first voltage to the second P-type well region, and wherein, in the plan view, no memory cell is formed in the fourth area and no memory cell is formed in the fifth area.

32. The semiconductor memory device according to claim 31, further comprising: a first electrical lead to supply the first voltage to the first and second P-type well regions via the first well contact and the third well contact, respectively; and a second electrical lead to supply the first voltage to the first and second P-type well regions via the second well contact and the well fourth contact, respectively, wherein the first and second electrical leads extend in the second direction.

33. The semiconductor memory device according to claim 32, wherein the first and third well contacts are linearly disposed in the second direction, and wherein the second and fourth well contacts are linearly disposed in the second direction.

34. The semiconductor memory device according to claim 31, wherein the memory array has (vii) a sixth area in which a fifth well contact to the N-type well region is arranged without being overlapped with the first area, and (viii) a seventh area in which a sixth well contact to the N-type well region is arranged without being overlapped with the first area, and wherein the N-type well region in the first area is supplied with a second voltage from the sixth area and the seventh area, and the first area has no contact to supply the second voltage to the N-type well region.

35. The semiconductor memory device according to claim 31, further comprising a sense amplifier, wherein the third and the fifth areas are arranged between the first area and the sense amplifier.

36. A semiconductor memory device comprising: first and second bit lines extending in a first direction; a plurality of word lines extending in a second direction crossing the first direction; a memory array having a plurality of memory cells, each of the memory cells having a first inverter including a first N-channel MOS transistor and a first P-channel MOS transistor, a second inverter including a second N-channel MOS transistor and a second P-channel MOS transistor with an input terminal of the second inverter being coupled to an output terminal of the first inverter and with an output terminal of the second inverter being coupled to an input terminal of the first inverter, a third N-channel MOS transistor having a source/drain path coupled between the output terminal of the first inverter and the first bit line, and a fourth N-channel MOS transistor having a source/drain path coupled between the output terminal of the second inverter and the second bit line, a gate of each of the third and fourth N-channel MOS transistors being connected to one of the plurality of word lines; a first P-type well region in which the first and third N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the first P-type well region extending in the first direction; a second P-type well region in which the second and fourth N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the second P-type well region extending in the first direction; an N-type well region in which the first and second P-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the N-type well region lying between the first and second P-type well regions; and wherein with respect to a plan view of a principal plane of the semiconductor memory device, (i) each of the plurality of memory cells is formed in a rectangular region which includes the first to fourth N-channel MOS transistors, the first and second P-channel MOS transistors, a first contact coupled between the third N-channel MOS transistor and the first bit line, a second contact coupled between the fourth N-channel MOS transistor and the second bit line, a third contact coupled between one of the plurality of word lines and the gate of the third N-channel MOS transistor, and a fourth contact coupled between the one of the plurality of word lines and the gate of the fourth N-channel MOS transistor, (ii) the memory array has a first area which consists of plural rectangular regions arranged in a first direction, (iii) a second area in which a first well contact to the first P-type well region is arranged without being overlapped with the first area, (iv) a third area in which a second well contact to the second P-type well region is arranged without being overlapped with the first area, and wherein the first P-type well region in the first area is supplied with a first voltage from the second area, the second P-type well region in the first area is supplied with the first voltage from the third area, and the first area has no contact to supply the first voltage to the first P-type well region and has no contact to supply the first voltage to the second P-type well region, wherein, in the plan view, no memory cell is formed in the second area and no memory cell is formed in the third area, and wherein a first well contact arranged in the second area and a second well contact arranged in the third area are linearly disposed at one of an upper portion and a lower portion of the memory array in the second direction.

37. The semiconductor memory device according to claim 36, further comprising: a first electrical lead to supply the first voltage to the first and second P-type well regions via the first contact and the second contact, respectively; and wherein the first electrical lead extends in the second direction.

38. The semiconductor memory device according to claim 37, wherein the first and second well contacts are linearly disposed in the second direction.

39. The semiconductor memory device according to claim 36, further comprising a sense amplifier, wherein the second and third areas are arranged between the first area and the sense amplifier.

40. A semiconductor memory device comprising: a pair of bit lines extending in a first direction; a plurality of word lines extending in a second direction crossing the first direction; a first memory array and a second memory array each having a plurality of memory cells, each of the memory cells having a first inverter including a first N-channel MOS transistor and a first P-channel MOS transistor, a second inverter including a second N-channel MOS transistor and a second P-channel MOS transistor with an input terminal of the second inverter being coupled to an output terminal of the first inverter and with an output terminal of the second inverter being coupled to an input terminal of the first inverter, a third N-channel MOS transistor having a source/drain path coupled between the output terminal of the first inverter and one of the pair of bit lines, and a fourth N-channel MOS transistor having a source/drain path coupled between the output terminal of the second inverter and the other of the pair of bit lines, a gate of each of the third and fourth N-channel MOS transistors being connected to one of the plurality of word lines; a first P-type well region in which the first and third N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the first P-type well region extending in the first direction; a second P-type well region in which the second and fourth N-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the second P-type well region extending in the first direction; an N-type well region in which the first and second P-channel MOS transistors are formed and which is formed commonly throughout the plurality of memory cells, the N-type well region lying between the first and second P-type well regions; a first electrical lead to supply a first voltage to the first and second P-type well regions; and a second electrical lead to supply a second voltage to the N-type well region, wherein the first memory array includes a first memory cell and the second memory array includes a second memory cell, wherein the plurality of word lines includes a first and word line and a second word line, wherein with respect to a plan view of a principal plane of the semiconductor memory device, (i) the first memory cell is formed in a first rectangular region defined by a first and a second side extending in the second direction and a fifth and a sixth side extending in the first direction, (ii) the second memory cell is formed in a second rectangular region defined by a third and a fourth side extending in the second direction and a seventh and an eighth side extending in the first direction, (iii) a third rectangular region bounded by the second and third sides is formed between the first and second memory arrays, the first electrical lead being electrically connected with the first and second P-type well regions in the third rectangular region, and the second electrical lead being electrically connected with the N-type well region in the third rectangular region, (iv) a first contact coupled between the third N-channel MOS transistor of the first memory cell and the one of the pair of bit lines is arranged on the first side, a second contact coupled between the fourth N-channel MOS transistor of the first memory cell and the other of the pair of bit lines is arranged on the second side, a third contact coupled between a first of the word lines and the gate of the third N-channel MOS transistor of the first memory cell is arranged on the fifth side, and a fourth contact coupled between the first word line and the gate of the fourth N-channel MOS transistor of the first memory cell is arranged on the sixth side, (v) a fifth contact coupled between the third N-channel MOS transistor of the second memory cell and the one of the pair of bit lines is arranged on the fourth side, a sixth contact coupled between the fourth N-channel MOS transistor of the second memory cell and the other of the pair of bit lines is arranged on the third side, a seventh contact coupled between a second of the word lines and the gate of the third N-channel MOS transistor of the second memory cell is arranged on the seventh side, and an eighth contact coupled between the second word line and the gate of the fourth N-channel MOS transistor of the second memory cell is arranged on the eighth side, wherein the first P-type well region in the first and second memory arrays is supplied with the first voltage from the third rectangular region, and each of the first and second memory arrays has no contact to supply the first voltage to the first P-type well region, wherein, in the plan view, no memory cell is formed in the third rectangular region, wherein the second P-type well region in the first and second memory arrays is supplied with the first voltage from the third rectangular region, and each of the first and second memory arrays has no contact to supply the first voltage to the second P-type well region, and wherein the N-type well region in the first and second memory arrays is supplied with the second voltage from the third rectangular region, and each of the first and second memory arrays has no contact to supply the second voltage to the N-type well region.

41. The semiconductor memory device according to claim 40, wherein the pair of bit lines includes a first bit line and a second bit line, wherein the source/drain path of the third N-channel MOS transistor is coupled between the output terminal of the first inverter and the first bit line, wherein the source/drain path of the fourth N-channel MOS transistor is coupled between the output terminal of the second inverter and the second bit line, wherein the first contact is coupled between the third N-channel MOS transistor of the first memory cell and the first bit line, wherein the second contact is coupled between the fourth N-channel MOS transistor of the first memory cell and the second bit line, wherein the fifth contact is coupled between the third N-channel MOS transistor of the second memory cell and first bit line, and wherein the sixth contact is coupled between the fourth N-channel MOS transistor of the second memory cell and second bit line.

42. The semiconductor memory device according to claim 41, wherein the first and second electrical leads are parallel to the plurality of word lines.

43. The semiconductor memory device according to claim 42, wherein the first and second electrical leads and the plurality of word lines are formed in a same layer.

44. The semiconductor memory device according to claim 42, further comprising a power supply line and a ground line, wherein the power supply line is parallel to the first and second bit lines, and wherein the ground line is parallel to the plurality of word lines.

45. The semiconductor memory device according to claim 42, wherein the first and second electrical leads are formed in another layer than a layer in which the first and second bit lines are formed.

46. The semiconductor memory device according to claim 40, wherein the first voltage is a ground potential and the second voltage is a power supply voltage.

47. The semiconductor memory device according to claim 40, further comprising a third memory cell arranged adjacent to the first memory cell in the second direction, a first diffusion layer forms a source of the first and third N-channel MOS transistors of the first memory cell, a second diffusion layer forms a source of a third and a fourth N-channel MOS transistor of the third memory cell, wherein the first and second diffusion layers are formed in the first P-type well region; wherein the first diffusion layer has no contact part to the second diffusion layer.

48. The semiconductor memory device according to claim 40, wherein the third rectangular region is bounded by ninth and tenth sides joining the second and third sides, wherein the ninth side is arranged in line with the fifth and seventh sides, and wherein the tenth side is arranged in line with the sixth and eighth sides.

49. The semiconductor memory device according to claim 40, wherein each of the first and the second memory arrays includes greater than or equal to 32 memory cells arranged in parallel to the first bit line in plan view.