WO1983002197A1

WO1983002197A1 - Process for forming a cmos integrated circuit structure

Info

Publication number: WO1983002197A1
Application number: PCT/US1982/001762
Authority: WO
Inventors: Corporation Ncr; Ronald Wayne Brower; Samuel Yue Chiao; Robert Frederick Pfeifer; Roberto Romano-Moran
Original assignee: NCR Corp
Current assignee: NCR Voyix Corp
Priority date: 1981-12-18
Filing date: 1982-12-14
Publication date: 1983-06-23
Anticipated expiration: 1984-06-18
Also published as: US4422885A; EP0097703A1

Abstract

Dans un procédé de formage d'un dispositif semiconducteur complémentaire dans un substrat semiconducteur possédant une première et une deuxième régions actives, on dépose une mince couche d'oxyde de porte (27) sur le substrat (10) et on dépose une couche de polysilicium (31) sur la couche d'oxyde de porte (27). La couche de polysilicium (31) est ensuite dopée jusqu'au niveau désiré par des impuretés du type n, après quoi on forme sur la couche de polysilicium (31) une couche relativement épaisse d'oxyde de masquage (32). Des électrodes de porte en polysilicium (34, 35) et des conducteurs d'interconnexion (36) sont définis et gravés dans la couche de polysilicium dopée (31) laissant des parties (32A, 32B) de la couche d'oxyde de masquage sur les électrodes de porte (34, 35). La première et la deuxième régions actives sont alors masquées alternativement, tout en effectuant une implantation ionique de dopants pour former des régions de source et de collecteurs (38, 39 et 41, 42). Pendant l'implantation ionique les parties de masque (32A, 32B) protègent les électrodes de porte en polysilicium (34, 35) d'un dopage ultérieur.In a method of forming a complementary semiconductor device in a semiconductor substrate having first and second active regions, a thin layer of gate oxide (27) is deposited on the substrate (10) and a layer of polysilicon is deposited (31) on the door oxide layer (27). The polysilicon layer (31) is then doped to the desired level by n-type impurities, after which a relatively thick layer of masking oxide (32) is formed on the polysilicon layer (31). Polysilicon door electrodes (34, 35) and interconnecting conductors (36) are defined and etched in the doped polysilicon layer (31) leaving portions (32A, 32B) of the masking oxide layer on the door electrodes (34, 35). The first and second active regions are then alternately masked, while performing ion implantation of dopants to form source and collector regions (38, 39 and 41, 42). During ion implantation the mask parts (32A, 32B) protect the polysilicon door electrodes (34, 35) from further doping.

Description

PROCESS FOR FORMING A CMOS INTEGRATED CIRCUIT STRUCTURE

Technical Field

This invention relates to processes of the kind for forming an integrated circuit from a structure including a body of a semiconductor material having a first active region of a first conductivity type, a second active region of the second conductivity type, and a silicon dioxide layer formed on said body encom- passing respective gate regions within said active re¬ gions.

Background Art

A CMOS device consists of an n-FET and a p-FET which are coupled so that the source or drain of one device is connected to the source or drain of the other device. In operation, one of the transistors will be functioning and the other will be off. When operating conditions within the circuit dictate that the func¬ tioning transistor turn off, the transistor which was previously off will begin to function due to the inter¬ connection of sources and/or drains of the two trans¬ istors. Thus, very little power is required during the non-switching state for these transistors, thereby enabling reduced power consumption. Other advantageous features of a CMOS device include high speed and per¬ formance. Balanced against these is the complexity of the processing technology of CMOS devices, particularly in meeting the stringent requirements of high inte¬ gration (or packing) density and high performance in very large scale integrated circuit applications.

In an attempt to meet the above requirements, conventional metal gate CMOS processing techniques have been replaced by silicon gate technology since this provides an area savings due to the better tolerances arising from the self-aligned nature of such gate

OMPI structures, due to the ability to provide an additional layer of interconnect and due to the availability of buried contacts.

An example of a silicon gate CMOS process, International Patent Application Publication No. W082/

02283 (European Publication No. 0067206), discloses form¬ ing the desired pattern of gate electrode and intercon¬ necting conductors on a semiconductor substrate such that each of the gate electrodes and conductors overlies a thin layer of gate oxide and is covered by a suitable oxidation and diffusion mask. The oxidation and dif¬ fusion mask, which consists of a dual layer of silicon dioxide and silicon nitride, serves as an oxidation mask and an implant/diffusion mask for the gate electrodes. The source and drain regions of the p-FET's and n-FET's are alternatively masked and formed by diffusing or im¬ planting p-type and n-type impurities into the substrate. A layer of barrier oxide is then thermally grown on the semiconductor substrate to completely cover the source and drain regions, thereby protecting these regions from further doping during subsequent doping of the gate electrodes and conductors. Thereafter, the oxidation and diffusion mask which overlies each of the gate electrodes and conductors is selectively removed (while leaving the source and drain regions covered) , thereby exposing the underlying gate electrodes and conductors. The exposed gate electrodes and conductors are then doped using a conventional doping technique.

Using this process a CMOS device having poly- silicon gates of a single conductivity and n-type im¬ purities is provided.

Because all of the gate electrodes and con¬ ductors have the same type of impurities, the need for metal bridges between (n and p⁺) polysilicon conduc- tors, which would otherwise be required, is avoided thereby making more efficient use of chip area. Also, because all the gate electrodes are not subjected to

O PI p-type dopant (e.g., boron) the problem of dopant (particularly, boron) penetration from the gate elec¬ trode, through the thin oxide layer which separates the gate electrode from the underlying semiconductor sub- strate, and into the underlying substrate thereby chan¬ ging the threshold voltage of the p-FET or producing a short between their source and drain regions is elimi¬ nated.

The process discussed above has the disadvan- tage that the oxidation of source and drain regions may consume some of the dopant in these regions. This may result in a higher source-to-drain interconnect resis¬ tance, thereby decreasing the device speed.

Disclosure of the Invention It is an object of the present invention to provide a process of the kind specified wherein the aforementioned disadvantage is alleviated.

Therefore, according to the present invention, there is provided a process of the kind specified char- acterized by the steps of forming, over said silicon dioxide layer, a polysilicon layer doped to the desired conductivity level for gate electrodes, patterning said polysilicon layer into gate electrodes corresponding to said first and second active regions, and forming source and drain regions of a first conductivity type in said second active region and source and drain regions of- the second conductivity type in said first active region without further doping of said polysilicon gate elec¬ trodes. It will be appreciated that in a process according to the invention the step of initially doping to the desired level polysilicon utilized for gate electrodes eliminates the requirement for oxidation of the source and drain areas. In brief summary of one embodiment of the invention, first a polysilicon layer is formed over the p-FET and n-FET regions and then doped n . Thereafter, an oxide mask is formed over the n polysilicon to pro¬ tect the polysilicon from further doping during the n-FET/p-FET source and drain forming steps. N poly- silicon gates are then defined followed by forming self- aligned n-FET and p-FET source and drain regions by alternately masking and implanting n-type and p-type impurities into the substrate. Finally, without re¬ moving the oxide mask over the polysilicon gates an iso- lation oxide is formed over the entire structure. Con¬ tact vias are then simultaneously etched in the different thickness oxide over the gates and source and drain re¬ gions, followed by metallization. It will be appre¬ ciated that since the mask covering the polysilicon gates is not removed, the need, in the process discussed above, for precise control of the etch time for selective mask removal from the polysilicon to be doped, is eliminated.

Brief Description of the Drawings

One embodiment of the present invention will now be described by way of example with reference to the accompanying drawings, in which:

Figs. 1-11 are cross-sectional views illus¬ trating various process steps of the present invention for forming a CMOS device with polysilicon gates of a single conductivity and single impurity type.

Best Mode for Carrying Out the Invention

Reference is now made to Figs. 1-11 wherein the successive steps of the fabrication process of the present invention are illustrated in detail. The steps to be discussed hereafter are illustrative of one tech¬ nique for implementing the process of the subject inven¬ tion. It will be obvious to those skilled in the micro¬ electronics art that the specific processes for imple¬ menting the various steps may be carried out in a number of different ways. The drawings herein are not to scale. The scale has been changed where needed to clearly show the structure.

Referring now to Fig. 1, the process of the present invention is initiated by selecting a semicon¬ ductor substrate 10 as the starting material. In order to accommodate both p-FET's and n-FET's on one chip, it is necessary to locate either type transistor (or both) in an isolation well(s). In other words, n-FET*s should be located in p-wells formed on an n-type substrate, or p-FET's located in n-wells formed on a p-type substrate, or each type of transistor located in its corresponding well formed on an n-type or p-type substrate. The choice presented here depends on such considerations as the particular application of the CMOS device, compati¬ bility of this process with other processes on the manufacturing line, and device physics, e.g., latch-up phenomenon and transconductance of n-FET's and p-FET's. For illustrative purposes, the substrate 10 shown in Fig. 1 is n-type silicon upon which a p-well 15 is formed. To form the p-well 15, the surface of substrate 10 is first cleaned and then oxidized, for example, by placing in a high temperature (1,000°C.) steam oxidation furnace, forming a relatively thick (of the order of 6500 Angstroms) silicon oxide 11. Next, a p-well open¬ ing 12 is etched in the oxide 11 using conventional photolithography and buffered hydrofluoric acid etching. Then, a relatively thin (e.g., 900 Angstroms) layer of silicon oxide 13 is thermally grown in the opening 12. During this oxide 13 growing step, little or no oxida¬ tion takes place in the remainder of the substrate since the oxidation thereover is diffusion limited by the thick oxide 11. Also, during this process step the silicon substrate in the opening 12 is consumed to a depth equal to approximately one-half of the oxide 13 thickness (i.e., 450 Angstroms, in the above example) due to the well-accepted fact that the thickness of the oxide grown from silicon is approximately twice that of the silicon consumed. As a consequence of this latest oxidation, a topographical step 14-14 (of an approximate 450 Angstroms height) is formed in the substrate, as shown in Fig. 1. This topographical step will be useful for aligning purposes at a later stage of the present process, as explained hereinbelow.

Referring further to Fig. 1, the p-well 15 is then formed by implanting p—type ions (such as boron) into the substrate 10 through the opening 12. During this implantation, the energy of the ions is so chosen as to penetrate the thin oxide layer 13, but not the thick oxide layer 11, thereby doping only the p-well 15.

For the oxide thickness specified above, typically, boron ions of energy 60 keV and dose 4 x 10 12 ions per square centimeter is used. To achieve the proper depth for the p-well, the structure is then subjected to a long, high temperature diffusion cycle. An exemplary procedure is used to introduce the structure in a fur- nace having a gaseous nitrogen atmosphere at a temper¬ ature of 1200^βC for about 24 hours. After the comple¬ tion of this step, the boron ions diffuse laterally and downward to a distance of about 6 microns.

After forming the p-well 15, the oxide layers 11 and 13 are removed by using a conventional etching technique.

Referring to Figs. 2 and 3, the next step of the process involves formation of channel stops which define the active region of the p-FET and n-FET to be formed subsequently. Referring to Fig. 2, as a first step in forming the channel stops, an oxide layer 16, hereinafter called support oxide, is grown from the substrate by dry thermal oxidation to isolate the sub¬ strate 10 from the next to be formed silicon nitride layer 17. The support oxide 16 prevents damage to the silicon substrate 10 caused by stresses which would be created on the substrate by a silicon nitride 17-silicon

O PI 10 interface. Such stresses induce dislocations in the silicon substrate 10 which result in undesirable leakage current channels and otherwise have a deleterious effect on the electrical characteristics of the interface. A typical thickness of the support oxide is about 550

Angstroms. Thereafter, and referring further to Fig. 2, a pair of oxidation and implantation masks are formed over the support oxide 16. First, nitride layer 17 is formed on the support oxide 16, then oxide layer 18 is formed on the nitride 17. The thicknesses of the nitride 17 and oxide 18 are typically in the range of (500-600) Ang¬ stroms and (1100-1300) Angstroms, respectively. The layers 17 and 18 are then patterned in a conventional manner using a photoresist 19 and then etched to form the dual-layer masks 20 and 21 which delineate the p-FET and n-FET active regions respectively. The topographical steps 14-14 in the p-well 15 discussed in connection with Fig. 1 are used to visually align the mask used to define the photoresist 19 over the dual mask 21. As shown in Fig. 3, the active regions are covered by the masks 20 and 21, whereas the field regions are covered only by support oxide 16.

Referring further to Fig. 3, the channel stops are then formed by doping the field regions which are potential sites of parasitic devices. In the completely formed CMOS device, metallized interconnect lines are formed over the field oxide. When an exceptionally high voltage is applied to a metallized line, the underlying substrate is prone to be inverted causing unwanted current flow between unrelated sources and drains. To suppress such parasitic FET's the field regions are doped, raising the threshold voltage of these regions to a level higher than the voltage that may be applied to the metallized lines. This field doping consists, in an exemplary process, of doping the p-well field regions 22-22 with p-type impurities and the n-substrate field regions 23-23 with n-type impurities. To dope the p- well field regions 22-22, first, a photoresist layer 24 is applied over the mask 20 and the n-substrate field regions 23-23 as shown in Fig. 3. Then, p-type ions, such as boron, are implanted in the p-well field regions 22-22 through the exposed support oxide 16. During this implantation step, the photoresist 24 pro¬ tects the n-substrate field regions 23-23 and the p-FET active regions from doping and the mask 21 protects the n-FET active region from doping. The photoresist mask 24 is then removed. Next, referring to Fig. 4, the n-substrate field regions 23-23 are doped by forming a photoresist layer 25 over the mask 21 and the p-well field regions 22-22 and implanting with n-type ions, such as phos¬ phorus, in a manner analogous to the previous step. The photoresist mask 25 is then removed.

Referring to Fig. 5, next the field (isolation) oxide 26 of thickness in the range 6,000-10,000 Angstroms is grown, for example by wet oxidation at a temperature of about l,000^βC. During this high temperature oxida- tion the nitride layer 17 of masks 20 and 21 acts as a barrier to diffusion of oxygen species, thereby pre¬ venting growth of oxide in the n-FET and p-FET active regions.

As shown in Fig. 6, the oxide mask 18, the nitride mask 17 and the support oxide 16 are then re¬ moved using conventional etching techniques. During this etching process, the thick field oxide regions 26 may be thinned somewhat, but not to a degree that their dielectric function is hampered. The bare active areas are now re-oxidized to form the gate oxide 27 of thick¬ ness in the range (300-1,000) Angstroms.

It is noted that the support oxide 16 (Fig. 5) may be left in place and used as the gate oxide. How¬ ever, a preferred procedure is to remove the oxide 16 (Fig. 5) and form a new oxide 27 (Fig. 6) in its place since this procedure permits a better control of the gate oxide thickness and also eliminates any possible

OMPI contamination of the gate oxide which might take place as in the case of support oxide 16 during the deposition and subsequent etching of nitride 17, etc.

After forming the gate oxide 27, the device active regions are subjected to one or two ion implan¬ tations (of the same or opposite impurity type) to adjust the threshold voltages of p-FET's and n-FET's to the desired value. If two implantations of opposite impurity type are used, two photoresist masks may have to be used to alternately mask and implant the p-FET and n-FET active areas. In the present exemplary process two ion implantations of the same conductivity type are accomplished. The type of implantation species, their energy, and dose is a function of several factors in- eluding the resistivity or doping level of the starting material, the doping level of the p-well 15, the gate oxide 27 thickness, the type of polysilicon gate (to be formed later) and the actual threshold voltages desired. For example, when the substrate 10 is n-type having a <100> crystal orientation and bulk resistivity of about (3-6) ohm-centimeter, the p-well 15 sheet resistance is about 7,000 ohms per square, the gate oxide 27 thickness is about 600 Angstroms, the polysilicon gates (to be formed) are n -type, and the desired n-FET and p-FET threshold voltages are +1 volt and -1 volt, respectively, the two implantations can be accomplished using boron ions of different doses. First, boron of dose 1.6 x 10 11 ions per square centime^'ter and energy 40 keV is implanted into the entire structure. During this step, referring to

Fig. 6, the boron ions penetrate through the thin gate oxide 27 and are lodged in the p-FET and n-FET active regions forming the surface charge layers 28 and 29, respectively. As a result of this implantation step, the p-FET threshold voltage will be about -1 volt.

Next, referring to Fig. 7, the p-FET active region is protected using a photoresist mask 30 and the exposed n-FET active area is implanted using boron ions of a higher dose, 3.9 x 10 ions per square centimeter, and the same energy (of 40 keV) as before. As a result of this second implantation, the surface charge of layer 29 will be increased to a level that will result if a single implantation of dose (1.6 x 10 + 3.9 10 =) 5.5 x 10 ions per square centimeter was used. At any rate, this combination of implantation steps sets the n- FET threshold voltage at about +1 volt, thereby matching it with the p-FET threshold voltage. The photoresist 30 is then removed.

Once the devices thresholds have been estab¬ lished at the desired level, the polysilicon gates and interconnecting lines can be formed. As shown in Fig. 8, a polysilicon layer 31 is formed on the semiconductor structure to a typical thickness of about 5000 Angstroms by, for example, chemical vapor deposition at atmos¬ pheric or low pressure of (0.3-1) millitorr. After forming, the polysilicon 31 is doped heavily using, for example, an n-type dopant. An exemplary process of doping the polysilicon 31 utilizes a phosphorus pre- deposition and furnace diffusion steps. During this step, the back surface of the substrate 10 is also doped for insuring good back surface ohmic contact which will be beneficial during the packaging stage of the CMOS device, and also acting as a getter for metallic impur¬ ities.

After the conductivity of the polysilicon 31 is increased to the desired level, a layer of silicon oxide 32 of about 3500 Angstroms thickness is formed by chemical vapor deposition on the polysilicon 31. The layer 32 protects the underlying silicon 31 from addi- ^"" tional doping during the later p-FET and n-FET source and drain implantation steps. Then, using a photoresist mask 33, the polysilicon gates and interconnecting conductor lines are defined by sequentially etching the oxide layer 32 and polysilicon 31 using buffered hydro¬ fluoric acid and a plasma, respectively. The photoresist 33 is then removed. Two polysilicon gates 34, 35 and one polysilicon conductor line 36 formed in the manner just described are shown in Fig. 9. Gate 34 is for the p-FET and gate 35 is for the n-FET. Conductor line 36 is formed on the top of the central field isolation oxide 26. Note that the oxide layer 32A, 32B, and 32C over the gates 34 and 35 and the interconnecting line 36, respectively, is not removed.'

The next step of the process, shown in Fig. 9, is.the formation of the source and drain for the n-FET. An exemplary technique for forming this step of the process involves depositing a layer of photoresist 37 on the semiconductor structure such that it uniformly covers the previously formed gate structures and con- ductor structures, then exposing and developing the photoresist to provide the photoresist mask 37 shown in Fig. 9. Specifically, mask 37 covers the p-FET active region, gate electrode 34 and conductor line 36. The thick field oxide regions 26, regardless of whether covered by mask 36 or not, of course, preserve intact the channel stops formed thereunder. There is no need to cover the n-FET gate electrode 35 by mask 37 because of the presence of the oxide mask 32B. As shown, only the active region of the n-FET is exposed. Thereafter, the source 38 and drain 39 of the n-FET are formed by implanting n-type impurities into the substrate. In performing this implantation step, n-FET gate 35 acts as a mask which serves to align the transistor's source 38 and drain 39 with the gate electrode 35. During this implantation step, the energy of the n-type ions are so chosen as to penetrate the thin oxide layer 27 over the source 39 and drain 40, but do not penetrate the rela¬ tively thick oxide 32B over the gate 35. Also, the n- type ion dose is chosen sufficiently high to counterdope the p-type surface layer 29 in the source 38 and drain 39 areas that was introduced previously, and produce n⁺ source 38 and drain 39. An exemplary set of implan¬ tation parameters for this step is arsenic ions of dose 8 x 10 ions per square centimeter and energy 80 keV. The sheet resistance of the n source 38 and drain 39 formed in this manner will be about 30 ohms per square. The photoresist 37 is then removed. The source and drain regions of the p-FET are then formed in a manner analogous to the formation of n-FET source and drain. Referring to Fig. 10, using conventional photolithographic techniques, a photoresist mask 40 is formed over the n-FET gate 35, source 38 and drain 39, and the conductor line 36. The p-FET gate electrode 34 is not covered by mask 40 since the oxide layer 32A masks this electrode. Thus, only the p-FET active region is exposed. P-type ions are then im¬ planted to for the p source 41 and drain 42 in a self- aligned relationship with the gate 34, in a manner fully analogous to the previous step in connection with the formation of the n-FET source 38 and drain 39. An exemplary set of implantation parameters to form the source 41 and drain 42 is boron ions of dose 6 x 10 ions per square centimeter and energy 35 keV. The photoresist 40 is then removed.

After forming the n-FET and p-FET sources and drains, the semiconductor structure is subjected to a high temperature anneal step to repair any damage to the silicon substrate lattice that may have been caused by the source and drain implants and to insure that the implanted ions are electrically activated. Typically, annealing is done at about 1000^βC. in a gaseous nitrogen atmosphere. It is noted that no specific annealing was necessary after the earlier implantation steps of this process, such as the implantations to form the channel stops and adjust the devices threshold voltages because during subsequent high temperature process steps, such as oxidation steps and polysilicon 31 deposition step, impurity activation naturally takes place.

The remainder of the process is well-known. Referring to Fig. 11, which is a cross-sectional repre¬ sentation of a completed CMOS device in accordance with the process of this invention, the remaining steps of the process include: (1) formation of a thick (about 10,000 Angstroms) oxide 43 over the entire semiconductor structure to serve as an electrically insulating layer; (2) etching contact vias in oxide 43 corresponding to the sources 38 and 41, drains 39 and 42, gates 34 and 35 and interconnecting conductor 36; (3) depositing a con¬ ductive layer such as aluminum or aluminum-silicon alloy over the structure; (4) delineation of this conductive layer forming conductive contacts 44 and 45 for the polysilicon gates 34 and 35, respectively, conductive contact 46 for the interconnecting conductor 36, conductive contacts 47 and 48 for sources 38 and 41, respectively, contacts 49 and 50 for drains 39 and 42, respectively; and (5) forming a passivation layer 51 of a material such as phosphosilicate glass over the device.

OMPI WIPO

Claims

CLAIMS :

1. A process for forming an integrated cir¬ cuit from a structure including a body of a semicon¬ ductor material (10) having a first active region of a first conductivity type, a second active region of the second conductivity type, and a silicon dioxide layer (27) formed on said body encompassing respective gate regions within said active regions, characterized by the steps of forming, over said silicon dioxide layer (27), a polysilicon layer (31) doped to the desired conduc- tivity level for gate electrodes, patterning said poly¬ silicon layer (31) into gate electrodes (34, 35) corres¬ ponding to said first and second active regions, and forming source and drain regions (41, 42) of a first conductivity type in said second active region and source and drain regions (38, 39) of the second conductivity type in said first active region without further doping of said polysilicon gate electrodes.

2. A process according to claim 1, charac¬ terized in that the step of patterning the polysilicon layer (31) into gate electrodes (34, 35) includes the steps of forming an oxide mask (32A, 32B, 32C) over the polysilicon layer (31) defining gate electrodes therein and etching the polysilicon layer (31) in the presence of the oxide mask (32A, 32B, 32C) to form the gate elec¬ trodes (34, 35).

3. A process according to claim 2, charac¬ terized in that the step of forming the source and drain regions (38, 39, 41, 42) in said active regions includes the steps of doping the substrate (10) in the presence of the oxide mask (32A, 32B, 32C) to form source and drain regions (38, 39, 41, 42) in said first and second active regions such that said source and drain regions (38, 39, 41, 42) are aligned with their corresponding gate electrodes (34, 35).

4. A process according to claim 4, charac¬ terized in that said oxide mask (32A, 32B, 32C) addition¬ ally defines regions for interconnecting conductor (36) in said polysilicon layer (31).

5.. A process according to claim 3, charac¬ terized in that said step of forming said source and drain regions (38, 39, 41, 42) includes the steps of forming a first mask (37) over the second active region, implanting impurities of said second conductivity type in the source and drain regions (38, 39) in the first, active region utilizing said oxide mask (32A, 32B, 32C) as an implantation mask, removing said first mask (37), ^"forming a second mask (40) over said first active region, and implanting impurities of said first conductivity type in the source and drain regions (41, 42) in the second active region utilizing said oxide mask (32A, 32B, 32C) as an implantation mask.

6. A process according to claim 5, charac¬ terized in that said first and second masks (37, 40) are of photoresist material.

7. A process according to claim 1, charac¬ terized in that said step of forming said polysilicon layer (31) includes the steps of depositing polysilicon to a desired thickness arid doping the deposited pol - silicon with n-type impurities.

8. A process according to claim 7, charac¬ terized in that the deposited polysilicon is doped with phosphorus.

9. A process according to any one of the pre¬ ceding claims, characterized in that the thickness of said silicon dioxide layer (27) is in the range of 300- 100 Angstroms, the thickness of said polysilicon layer 9. (concluded)

(31) is about 5000 Angstroms and the thickness of said oxide mask layer (32A, 32B, 32C) is about 3500 Angstroms.