Method of forming a nand-type flash memory device having a non-stacked gate transistor structure

[0002] This application is a divisional of Ser. No. 08/993,368 filed Dec. 18, 1997, which is entitled NOVEL NAND TYPE CORE CELL STRUCTURE AND CORRESPONDING PROCESS FOR HIGH DENSITY AND HIGH PERFORMANCE FLASH MEMORY DEVICE.

[0003] The present invention generally relates to integrated circuits and, in particular, to a flash memory structure and a method of flash memory fabrication wherein a new core cell structure eliminates a stacked gate structure for the select gate transistors while eliminating a core dual oxide manufacturing step. The elimination of the core dual oxide step substantially simplifies the process, eliminates associated tunnel oxide reliability concerns and shrinks the size of the select gate transistor by eliminating the need for a poly1 contact.

[0004] Semiconductor devices typically include multiple individual components formed on or within a substrate. Such devices often comprise a high density section and a low density section. For example, as illustrated in prior art FIG. 1a, a memory device such as a flash memory 10 comprises one or more high density core regions 11 and a low density peripheral portion 12 on a single substrate 13. The high density core regions 11 typically consist of at least one M×N array of individually addressable, substantially identical floating-gate type memory cells and the low density peripheral portion 12 typically includes input/output (I/O) circuitry and circuitry for selectively addressing the individual cells (such as decoders for connecting the source, gate and drain of selected cells to predetermined voltages or impedances to effect designated operations of the cell such as programming, reading or erasing).

[0005] The memory cells within the core portion 11 are coupled together in a NAND-type circuit configuration, such as, for example, the configuration illustrated in prior art FIG. 1b. Each memory cell 14 has a drain 14a, a source 14band a stacked gate 14c. A plurality of memory cells 14 connected together in series with a drain select transistor at one end and a source select transistor at the other end to form a NAND string as illustrated in prior art FIG. 1b. Each stacked gate 14cis coupled to a word line (WL0, WL1, . . . , WLn) while each drain of the drain select transistors are coupled to a bit line (BL0, BL1, . . . , BLn). Lastly, each source of the source select transistors are coupled to a common source line Vss. Using peripheral decoder and control circuitry, each memory cell 14 can be addressed for programming, reading or erasing functions.

[0006] Prior art FIG. 1crepresents a fragmentary cross section diagram of a typical memory cell 14 in the core region 11 of prior art FIGS. 1aand 1b. Such a cell 14 typically includes the source 14b, the drain 14aand a channel 15 in a substrate or P-well 16; and the stacked gate structure 14coverlying the channel 15. The stacked gate 14cfurther includes a thin gate dielectric layer 17a(commonly referred to as the tunnel oxide) formed on the surface of the P-well 16. The stacked gate 14calso includes a polysilicon floating gate 17bwhich overlies the tunnel oxide 17aand an interpoly dielectric layer 17coverlies the floating gate 17b. The interpoly dielectric layer 17cis often a multilayer insulator such as an oxide-nitride-oxide (ONO) layer having two oxide layers sandwiching a nitride layer. Lastly, a polysilicon control gate 17doverlies the interpoly dielectric layer 17c. The control gates 17dof the respective cells 14 that are formed in a lateral row share a common word line (WL) associated with the row of cells (see, e.g., prior art FIG. 1b). In addition, as highlighted above, the drain regions 14aof the respective cells in a vertical column are connected together by a conductive bit line (BL). The channel 15 of the cell 14 conducts current between the source 14band the drain 14ain accordance with an electric field developed in the channel 15 by the stacked gate structure 14c.

[0007] According to conventional operation, the flash memory cell 14 operates in the following manner. The cell 14 is programmed by applying a relatively high voltage V_G(e.g., approximately 18 volts) to the control gate 17dand connecting the drain, source and P-well to ground. A resulting high electric field across the tunnel oxide 17aleads to a phenomena called “Fowler-Nordheim” tunneling. During this process, electrons in the core cell channel region tunnel through the gate oxide into the floating gate 17band become trapped in the floating gate 17bsince the floating gate 17bis surrounded by insulators (the interpoly dielectric 17cand the tunnel oxide 17a). As a result of the trapped electrons, the threshold voltage of the cell 14 increases by about 3 to 5 volts. This change in the threshold voltage (and thereby the channel conductance) of the cell 14 created by the trapped electrons is what causes the cell to be programmed.

[0008] To read the memory cell 14, a predetermined voltage V_Gthat is greater than the threshold voltage of an erased cell, but less than the threshold voltage of a programmed cell, is applied to the control gate 17dwith a voltage applied between the source 14band the drain 14a. If the cell 14 conducts, then the cell 14 has not been programmed (the cell 14 is therefore at a first logic state, e.g., a zero “0”). Likewise, if the cell 14 does not conduct, then the cell 14 has been programmed (the cell 14 is therefore at a second logic state, e.g., a one “1”). Consequently, one can read each cell 14 to determine whether it has been programmed (and therefore identify its logic state).

[0009] In order to erase the flash memory cell 14, a relatively high voltage V_S(e.g., approximately 20 volts) is applied to the P-well 16 and the control gate 17dis held at a ground potential (V_G=0), while the drain 14aand the source 14bare allowed to float. Under these conditions, a strong electric field is developed across the tunnel oxide 17abetween the floating gate 17band the P-well 16. The electrons that are trapped in the floating gate 17bflow toward and cluster at the portion of the floating gate 17boverlying the source region 14band are extracted from the floating gate 17band into the source region 14bby way of Fowler-Nordheim tunneling through the tunnel oxide 17a. Consequently, as the electrons are removed from the floating gate 17b, the cell 14 is erased.

[0010] There is a strong need in the art for a flash memory device structure and process for manufacture that improves the performance and reliability of the device while simplifying its method of manufacture.

[0011] The present invention relates to flash memory device structure and a method for its manufacture. In a core portion of a NAND-type flash memory cell, a select gate transistor has a structure which includes a stacked ONO dielectric layer for a gate oxide. The select gate transistor structure allows the device to be easily fabricated by eliminating the dual core oxide formation process. Elimination of the dual core oxide process steps reduces the number of process steps and prevents a potential source of contamination prior to the formation of the core tunnel oxide, thereby improving the reliability of the device.

[0012] According to another aspect of the present invention, the formation of a select gate transistor structure utilizing the stacked ONO dielectric layer for a gate oxide eliminates the stacked gate structure for the select gate transistors and the need for a poly1 contact to short out the poly1 and poly2 layers, thereby advantageously reducing the die area. In addition, eliminating the stacked gate structure for the select gate transistor allows for a channel stop implant to be performed in the select gate transistor region, thereby improving the bit line isolation and decreasing the potential for bit line to bit line punch through.

[0013] According to yet another aspect of the present invention, the select gate transistor uses a stacked ONO dielectric layer which improves the resistance of the select gate transistor to high field stress induced dielectric damage. In addition, the ONO layer is substantially thinner than the other select gate transistor oxide, thereby improving the drive current and performance of the select gate transistor.

[0014] To the accomplishment of the foregoing and related ends, the invention comprises the features hereinafter fully described and particularly pointed out in the claims. The following description and the annexed drawings set forth in detail certain illustrative embodiments of the invention. These embodiments are indicative, however, of but a few of the various ways in which the principles of the invention may be employed. Other objects, advantages and novel features of the invention will become apparent from the following detailed description of the invention when considered in conjunction with the drawings.

[0015] FIG. 1ais a plan view illustrating a prior art layout of a flash memory chip;

[0016] FIG. 1bis a schematic diagram illustrating a prior art NAND-type flash memory circuit configuration;

[0017] FIG. 1cis a fragmentary cross section illustrating a prior art stacked gate flash memory cell;

[0018] FIG. 2ais a schematic diagram illustrating a core portion of a NAND-type flash memory device;

[0019] FIG. 2bis a plan layout view of the core portion of the NAND-type flash memory device of FIG. 2a;

[0020] FIG. 3ais a cross section of a stacked gate flash memory cell;

[0021] FIG. 3bis a cross section of a select gate transistor having a stacked gate structure, wherein the poly1 and poly2 layers are shorted together;

[0022] FIG. 3cis a cross section of a periphery low voltage transistor;

[0023] FIG. 3dis a cross section of a periphery high voltage transistor;.

[0024] FIG. 4 is a flow chart diagram illustrating a semiconductor manufacturing process flow for forming the NAND-type flash memory device of FIGS. 2aand 2b;

[0025] FIGS. 5a-5lare cross section diagrams illustrating the various steps in the manufacturing process of FIG. 4;

[0026] FIG. 6ais a schematic diagram illustrating a core portion of a NAND-type flash memory device according to the present invention;

[0027] FIG. 6bis a plan layout view of the NAND-type flash memory device of FIG. 6aaccording to the present invention;

[0028] FIG. 7ais a cross section of a stacked gate flash memory cell in the core region according to the present invention;

[0029] FIG. 7bis a cross section of a select gate transistor in the core region according to the present invention;

[0030] FIG. 7cis a cross section of a periphery low voltage transistor according to the present invention;

[0031] FIG. 7dis a cross section of a periphery high voltage transistor according to the present invention;

[0032] FIG. 8 is a flow chart diagram illustrating a semiconductor manufacturing process flow for forming the NAND-type flash memory device of FIGS. 6aand 6baccording to the present invention; and

[0033] FIGS. 9a-9iare cross section diagrams illustrating the various steps in the manufacturing process of FIG. 8 according to the present invention.

[0034] The present invention will now be described with reference to the drawings wherein like reference numerals are used to refer to like elements throughout. The present invention relates to a flash memory device and a method for its manufacture that, according to one aspect of the present invention, eliminates the dual core oxide processing steps by replacing the stacked gate transistor structure with a transistor structure that includes the stacked gate ONO layer as a gate oxide. Consequently, the steps used in forming the stacked gate transistor structures may be used to construct the core select gate transistors, thereby substantially reducing the number of processing steps needed to construct the flash memory device.

[0035] According to another aspect of the present invention, elimination of the dual core oxide processing steps advantageously improves the manufacturability of the process by eliminating a potential source of tunnel oxide contamination and thereby improving the tunnel oxide integrity in the core flash memory cell region. The present invention also provides for a reduction in the size of the core portion of the device since the new select gate transistor structure is not a stacked gate structure. Consequently, a poly1 contact is not required to short out the poly1 and poly2 layers. Eliminating the poly1 contact reduces the size of each select gate transistor. Eliminating the need for a poly1 contact also allows for the elimination of subsequent processing needed to create the poly1 contact.

[0036] According to yet another aspect of the present invention, elimination of the stacked gate structure for the select gate transistors in the core region allows for a channel stop implant to be performed in the select gate transistor regions which improves bit line isolation and substantially reduces the potential for bit line to bit line punch through, thereby improving the reliability of the flash memory device. In addition, since the select gate transistor utilizes the stacked ONO dielectric layer as a gate oxide, the select gate transistor exhibits improved performance by being more resistant to high field stress induced dielectric darage. Furthermore, the ONO gate oxide is substantially thinner than other select gate transistor gate oxides, thereby improving the drive current of the select gate transistors.

[0037] The present invention may best be understood and its advantages appreciated in conjunction with the core structure and process of FIGS. 2a-5k. A circuit schematic diagram illustrating a core portion 11 of a NAND-type flash memory device is illustrated in FIG. 2a. The core portion 11 includes a memory cell region 22 which is bounded on one side by a drain select transistor portion 24 and bounded on another side by a source select transistor portion 26. Each of the select transistor portions 24 and 26 contain select gate transistors 24a-24cand 26a-26c, respectively, which operate to selectively activate a desired bit line (e.g., BLN−1, BLN, BLN+1) by ensuring the selectivity of each bit line and preventing the cell current from conducting current through the bit line during a programming operation as is well known by those skilled in the art.

[0038] In the NAND-type flash memory process which forms the core circuit 12 of FIG. 2aand its corresponding circuit layout (of which FIG. 2bis a plan view), a dual core cell oxide process is used to form the memory cell oxides and the select gate transistor oxides, respectively (since the memory cell oxides and select gate transistor oxides have different thicknesses). In addition, a dual periphery gate oxide process for formation of the high and low voltage transistors in the periphery region 14 (see, e.g., FIG. 1) is also utilized. The process further includes a double polysilicon layer formation layer to generate the stacked gate structures in the core region 12.

[0039] The core select gate transistors 24a-24cand 26a-26care similar in structure to the stacked gate flash memory structures in the memory cell region 22 except that their gate oxide is approximately twice as thick as the cell oxide (also called the tunnel oxide) in the stacked gate structure of the memory cell (about 180 Å compared to about 95 Å). The select transistors 24a-24cutilize a gate oxide of 180 Å to improve the reliability of the transistors by reducing the vulnerability of the devices to band to band tunneling induced hot carrier stress during programming. The select gate transistors 24a-24cand 26a-26cfurther differ from the stacked gate flash memory cell structures of the region 22 because they operate as conventional MOS transistors and therefore have the first and second polysilicon layers shorted together to simulate a standard MOS transistor configuration. A plan circuit layout view of the conventional NAND-type circuit arrangement 12 is illustrated in FIG. 2b.

[0040] The dual core oxides (the tunnel oxide and the select gate oxide) are created by growing a first gate oxide of approximately 140 Å and then using a mask to open a space for etching to thereby define a tunnel oxide region. Subsequent the etching, another oxide is grown to form the tunnel oxide of about 95 Å in the previously etched region while the unetched select gate oxide region continues to grow to a thickness of about 180 Å.

[0041] Similarly, the dual periphery gate oxides are later formed in a dual oxide process that includes growing an initial gate oxide of about 285 Å and then using a mask to define for etching a thin low voltage transistor gate oxide region. Subsequent the etching, another oxide is grown to generate a 160 Å low voltage transistor oxide in the previously etched region while the unetched high voltage transistor oxide region continues to grow to a thickness of about 400 Å.

[0042] In the above process, the first polysilicon layer (poly1) is used as the floating gate region for the core memory cells and the second polysilicon layer (poly2) is used for both the core memory cell control gates and the gate regions for the periphery transistors (both high and low voltage). Consequently, the NAND-type process involves the fabrication of four types of devices, as illustrated in FIGS. 3a-3d, wherein FIG. 3ais a cross section of a stacked gate memory cell in the core memory cell region 22, FIG. 3bis a cross section of a select gate transistor in the select gate regions 24 and 26 (wherein the poly1 and poly2 layers are shorted together), FIG. 3cis a cross section of a periphery low voltage transistor and FIG. 3dis a cross section of a periphery high voltage transistor.

[0043] A semiconductor manufacturing process flow illustrating the NAND-type flash memory device process 100 is illustrated in FIG. 4 and will be discussed in detail in conjunction with FIGS. 5a-5l. The process 100 begins with a core mask and core Vt (threshold voltage) implant step 102 wherein a photoresist mask 103 is formed and patterned over a P-well which resides in an N-well which resides in a P-substrate 104 to define an opening over a core region 105. The core region 105 is then subjected to a P-type impurity dopant via, for example, ion implantation to thereby form a highly doped channel region 106 in the P-well of the core region 105, as illustrated in FIG. 5a.

[0044] The photoresist mask 103 is then removed, a select gate pre-clean is performed and an oxide layer 108 is formed over the surface of the device at step 110, as illustrated in FIG. 5b. A tunnel oxide mask (TOX) (not shown) is then formed at step 112 to define an area 114 within the core region 105 in which the tunnel oxide for the core memory cells will be formed. An etch step follows to remove the oxide layer 108 in the region 114 exposed by the TOX mask but not in the peripheral regions 116 and 118, respectively, as illustrated in FIG. 5c. Step 112 continues with a photoresist clean to remove the TOX mask.

[0045] Once the cleaning is complete, a tunnel oxide 119 having a thickness of about 95 Å is grown in the region 114 at step 120. At the same time, the second oxidation step causes the oxide layer 108 to increase to about 180 Å. Therefore steps 110, 112 and 120 constitute the core dual oxide processing steps to form a memory cell tunnel oxide of about 95 Å and a select gate transistor gate oxide of about 180 Å in the core region 105.

[0046] Step 120 also includes the deposition of a first polysilicon layer 122 over the surface of the device, as illustrated in FIG. 5d. The polysilicon layer 122 is then covered with a photoresist mask (not shown) at step 124 and etched to define a plurality of poly1 regions in the core region 105, wherein the regions comprise the floating gates for stacked gate flash memory cells and the first polysilicon layer for the source select transistor and drain select transistor, respectively, as illustrated in FIG. 5e. Note that in step 124, the polysilicon layer 122 is not etched between bit lines in the select gate transistor regions 24 and 26, as illustrated in FIG. 2b. Since FIG. 5etransversely cuts across the word lines, this lack of poly etching between the bit lines is not visible. Consequently, FIGS. 5fand 5gare provided which are cross sections of the core region 105 taken along dotted lines 5f—5fand 5g—5gin FIG. 5e, respectively.

[0047] Note that in select gate transistor regions of FIG. 5g, the poly region 122 is not etched between the bit lines 125 since this poly1 layer 122 is later used as an interconnect to short out the poly1 and poly2 layers (see also FIG. 3b).

[0048] A channel stop implant followed by a photoresist clean is then performed at step 126 to form an isolation implant doping region 129 under the field oxide 127. The channel stop implant serves to isolate the bit lines from one another in the memory cell region 22 of FIG. 2ain regions where the poly1 layer 122 was etched (that is, over the field oxide regions which are not shown). This channel stop implant, however, is not performed in the select gate regions 24 and 26 of FIG. 2a, because the poly1 layer 122 is not etched away, as illustrated in FIG. 5g. Consequently, the poly1 layer 122 shields the field oxide regions 127 from the channel stop implant which results in poor bit line to bit line isolation in the select gate transistor regions 24 and 26.

[0049] The process 100 continues at step 128 wherein an interpoly dielectric layer 130 is formed over the surface of the device and patterned to overlie the first polysilicon regions 122, as illustrated in FIG. 5h. The interpoly dielectric layer 130 often consists of an ONO layer which is an oxide/nitride/oxide layer formed via conventional processing techniques as is well known by those skilled in the art.

[0050] Following the ONO deposition and patterning at step 128, a pre-oxidation clean-up step is conducted followed by another oxidation of about 285 Å across the surface of the device at step 132. A channel threshold voltage mask is then formed at step 134 followed by an implant to adjust the threshold voltage at step 134. The peripheral gate oxide mask (GOX)(not shown) is then formed at step 136 which defines the opening for the thin gate oxide area of the low voltage peripheral transistor in the low voltage peripheral region 116. Step 136 includes the etching away of the oxide layer in the region 116 but not in the core region 105 since the oxide does not grow appreciably on the ONO layer, thereby leaving an intermediate oxide layer 137 remaining in the high voltage peripheral region 118, as illustrated in FIG. 5i. The etching is then followed by a photoresist clean to remove the GOX mask.

[0051] Once the GOX mask is removed, another oxidation step takes place at step 138, wherein a thin gate oxide 140 of about 160 Å is formed in the low voltage peripheral region 116 and a thick gate oxide 142 of about 400 Å continues to grow (from the intermediate oxide layer 137) in the high voltage peripheral region 118, as illustrated in FIG. 5j.Therefore steps 136 and 138 constitute the periphery dual oxide processing steps to form the periphery low voltage and high voltage transistor oxides, respectively. Step 138 also includes the deposition of a second polysilicon layer and a tungsten silicide layer which is collectively illustrated as a layer 144 in FIG. 5j.

[0052] The process 100 continues at step 145, wherein a photoresist mask (not shown) is formed and patterned with openings to define the low and high voltage periphery transistors in the periphery regions 116 and 118, respectively, and the control gates for the stacked gate structures in the core region 105, as illustrated in FIG. 5k. The process 100 is then completed at step 146 (at least with respect to the formation of the four separate transistor structures illustrated in FIGS. 3a-3d. Step 146 includes the formation of a self-aligned mask (not shown) which defines openings in the core region 105 and the etching of the interpoly dielectric layer 130 and first polysilicon layer 122, respectively, as illustrated in FIG. 5l. The step 146 thereby results in a low voltage peripheral transistor 150 having a gate oxide 152 of about 160 Å and a poly2 gate region 154, and a high voltage peripheral transistor 160 having a gate oxide 162 of about 400 Å and poly2 gate 164. The etching of step 146 further defines the stacked gate structures that form the select gate transistors and the stacked gate/memory cells, respectively. The select gate transistors 170 (the select source and select drain transistors) each have a select gate oxide layer 108 of about 180 Å, a poly1 layer 122aand 122d, and an interpoly dielectric or ONO layer 130 and the poly2 layer 144. At a later stage in the process, the poly1 122aand 122dand poly two 144 are shorted together to achieve the select gate transistors as illustrated in FIG. 3b. The remaining stacked gate/memory cells 180 each have a tunnel oxide layer of about 95 Å, a floating gate layer 122, an interpoly dielectric or ONO layer 130 and a control gate layer 144.

[0053] The fabrication process discussed above in conjunction with FIGS. 2a-5lhas several advantageous features, however, the process may be improved. First, the dual core oxide steps needed to produce the memory cells and select gate transistors having different oxide thicknesses in the core region is somewhat complicated and requires an additional mask, mainly the TOX mask discussed above, as well as the associated resist clean and oxide regrowth steps. In addition, poor tunnel oxide integrity may arise due to contamination during the TOX masking steps since a conventional HF dip can not be used to remove residual oxide. Furthermore, a channel stop implant can not be performed in the select gate transistor regions after the poly1 etch step as discussed above since the poly1 is used as an interconnect layer to short out the poly1 and poly2. The lack of a channel stop implant in the select gate transistor areas may result in bit line to bit line punch through. In addition, the spacing region between the first word line (WL1) and the select drain transistors 24a-24cof FIG. 2bmay turn on when WL1 equals twenty (20) volts during programming which limits the memory device operation at lower programming voltages, thereby resulting in a slower memory device. Furthermore, since the select gate structures require that the poly1 and poly2 layers be shorted together, a special poly1 contact process is later required to effectuate that short circuit. Lastly, the poly1 contact increases the size of each select gate transistor which adversely affects the die size of the flash memory device.

[0054] A circuit schematic of a NAND-type flash. memory device and a corresponding plan view layout diagram according to one embodiment of the present invention is illustrated in FIGS. 6aand 6b, respectively. A core portion 200 of the flash memory device includes a core memory cell portion 202 which is similar to the memory cell portion 22 of FIG. 2a. The core memory cell portion 202 is bounded on one side by a drain select gate transistor region 204 and on another side by a source select gate transistor region 206. The select gate transistor regions 204 and 206 differ from the select gate transistors 24 and 26 in that the select gate transistors within the regions 204 and 206 of the present invention are not stacked gate structures (which required the shorting of the poly1 and poly2 layers). Furthermore, since the select gate transistor regions 204 and 206 are not stacked gate structures, but instead are like standard transistor structures that utilize the stacked ONO dielectric layer as a gate oxide, the core dual oxide fabrication steps required to form the stacked gate structures having differing oxide thicknesses is eliminated, thereby simplifying the semiconductor processing.

[0055] In a preferred embodiment of the present invention, the tunnel oxide and poly1 layers used to form the stacked gate structure are removed from the select gate transistor regions 204 and 206 and the subsequent ONO and poly2 layer formation steps which are used to complete the stacked gate memory cell structure are also used to form the select gate transistors in the regions 204 and 206. Therefore the process steps and masks used to form the stacked gate memory cell structures may also be used to form the core select gate transistors. Therefore, instead of the process requiring two separate dual oxide fabrication process steps, the four transistor structures illustrated in FIGS. 7a-7deliminate one set of dual oxide processing steps in forming the tunnel oxide and the select gate transistor oxide in the core region, thereby substantially eliminating sensitive tunnel oxide contamination associated with the core dual oxide process.

[0056] The structural differences between the present invention and that discussed in FIGS. 2a-5lthe may further be seen in a comparison of the plan layout view of the present invention of FIG. 6bwith FIG. 2b. Note that in FIG. 6b, the first layer of polysilicon (poly1) is etched between the bit lines (BLN−1, BLN, BLN+1) in the core memory region 202 and is completely etched away in the select gate transistor regions 204 and 206. In addition, a core implant mask (which will be described in greater detail supra) prohibits a core Vt (threshold voltage) implant in the select gate transistor regions 204 and 206. In addition, the removal of the poly1 in the regions 204 and 206 allow a channel stop implant to be performed in the select gate regions. These differences provide improved bit line isolation, allow the select gate transistors to have a reasonable threshold voltage and result in improved word line field turn-on. The substantial improvement in bit line isolation allows for higher programming voltages which provide for higher speed programming functionality and the potential for use of the structure in multi-level NAND-type devices. In addition to the above features, the poly2 gate region, which may be composed of a poly/tungsten silicide combination, may be used as an interconnect for the select gate transistors, thereby further improving their performance.

[0057] A method of forming a NAND type flash memory device according to the preferred embodiment of the present invention will now be described in conjunction with a semiconductor manufacturing flow chart in FIG. 8 and FIGS. 9a-9i. The process 300 begins with a core mask and implant step 302 wherein a photoresist mask 303 is formed and patterned over a P-well which resides in an N-well in the P-substrate 304 to define an opening over a core region 305. The opening in the core region 305 is then subjected to a P-type impurity dopant via, for example, ion implantation to thereby form a highly doped core channel region 306 in the P-well of the core region 305 where the stacked gate memory cell structures will be formed, as illustrated in FIG. 9a. Note that the photoresist mask 303 shields portions of the core region 305 where the select gate transistors will be formed, as illustrated both in FIGS. 6band 9a.

[0058] The photoresist mask 303 is then removed and a tunnel oxide 308 having a thickness of about 95 Å is grown over the surface of the device at step 310. The oxide layer is grown, for example, in a dry oxidation furnace.

[0059] Step 310 also includes the deposition of a first polysilicon layer 312 over the surface of the device, as illustrated in FIG. 9b. The polysilicon layer 312 is then covered with a photoresist mask 312aat step 314 and etched to define a plurality of poly1 regions 312, wherein the regions 312 comprise the floating gates for stacked gate flash memory cells in the core region 305, as illustrated in FIG. 9c.

[0060] The etching of the poly1 layer 312 completely removes the polysilicon layer 312 and the tunnel oxide 308 in the select gate transistor areas 313 but not in the low voltage and high voltage periphery areas 314 and 315, respectively. The etching of the poly1 layer 312 in the area 317 where the stacked gate memory cells will be formed is not illustrated in FIG. 9cbecause the cross section diagram is taken parallel to the bit lines, FIG. 9d, however, illustrates the manner in which the poly1 etch step 314 forms the various bit lines (and floating gate regions) in the region 317, wherein FIG. 9dis a cross section diagram taken along dotted lines 9d—9dof FIG. 9c.FIG. 9dillustrates the polysilicon layer 312 etched to form a plurality of bit lines 318 which are separated by core field oxide regions 319.

[0061] A channel stop implant is then performed at step 318. Since the cross section diagram of FIG. 9cis oriented to cut transversely across the word lines, the channel stop implant (which helps isolate the bit lines) is not illustrated in the Figure. FIG. 9d, however, illustrates the core portion 305 of FIG. 9cso that the cross section transversely cuts across the bit lines. FIG. 9dillustrates how the absence of poly1 between the bit lines in the select gate transistor regions allows the channel stop implant in the region 319 to be formed thereby forming a plurality of channel stop regions 320 under the field oxide regions 319.

[0062] The channel stop implant uses the poly1 etch photomask 312aof step 314 to define the regions 320 that receive the implant in the region 317. Step 318 then concludes with a photoresist clean step to remove the photomask 312aused for the poly1 etch and the channel stop implant.

[0063] FIG. 9eis a cross section diagram taken along dotted lines 9e—9eof FIG. 9c. FIG. 9eillustrates the region 313 in the core region 305 that corresponds to the select gate transistor regions 204 and 206 of FIGS. 6aand 6b. Note that the polysilicon has been completely etched away in the region 313, thereby leaving only the field oxide regions 319. The channel stop implant of step 318 therefore impacts the entire region 313 of FIG. 9cto alter the dopant concentration of the substrate 304 in those regions, resulting in a region 321, as illustrated in FIG. 9e. The core implant step 318 thereby serves as the threshold voltage implant of the select gate transistors that will be later formed in the regions 313.

[0064] The present invention provides a significant advantage over the structure and method of FIGS. 2a-5lsince the poly1 layer is etched away in the select gate regions 204 and 206, as illustrated in FIGS. 6b, 9cand 9e. Etching away the polysilicon in the regions 204 and 206 allows the channel stop implant to be performed in the regions 204 and 206, which thereby decreases the potential for bit line to bit line punch through or leakage in the select gate regions 204 and 206, respectively. The channel stop implant region step 318 also provides good bit line to bit line isolation which allows the device to operate at a wider range of program voltages.

[0065] The process 300 continues at step 320 wherein an interpoly dielectric layer 322 composed of an insulating material is formed over the surface of the device and patterned to overlie the poly1 floating gate regions 322 in the memory cell region 317 and form a gate oxide layer in the select gate transistor regions 313, respectively, as illustrated in FIG. 9f. The interpoly dielectric layer 322 preferably consists of an ONO layer of about 130 Å which is an oxide/nitride/oxide layer formed via conventional processing techniques as is well known by those skilled in the art.

[0066] Following the ONO deposition and patterning at step 320, a pre-oxidation clean-up step is conducted followed by the formation of an oxide layer across the surface of the device at step 322. A periphery transistor threshold voltage mask is then formed (not shown) followed by an implant to adjust the threshold voltage at step 324. A peripheral gate oxide mask (GOX)(not shown) is then formed at step 326 which defines an opening for the thin gate oxide area for the low voltage peripheral transistor in the low voltage peripheral region 314. Step 326 includes the etching away of the oxide layer in the region 314, thereby leaving an intermediate oxide layer 330 remaining in the high voltage peripheral region 315, as illustrated in FIG. 9f. The etching is then followed by a photoresist clean to remove the GOX mask.

[0067] Once the GOX mask is removed, another oxidation step takes place at step 334, wherein a thin gate oxide 336 of about 160 Å is formed in the low voltage peripheral region 314 to form the gate oxide for the low voltage peripheral transistor while a thick gate oxide 337 of about 400 Å continues to grow (from the intermediate oxide layer 330) in the high voltage peripheral region 315, as illustrated in FIG. 9g. Therefore steps 326 and 334 constitute the periphery dual oxide processing steps to form the periphery low voltage and the high voltage transistor oxides, respectively. Step 334 also includes the deposition of a second polysilicon layer and a tungsten silicide layer which is collectively illustrated as a layer 338 in FIG. 9g.

[0068] The process 300 then continues at step 339, wherein a photoresist mask (not shown) is formed and patterned with openings to define the low and high voltage periphery transistors in the periphery regions 314 and 315, respectively, the select gate transistors and the control gate structures for the stacked gate devices in the core region 305, as illustrated in FIG. 9h.

[0069] The process 300 is then completed at step 340 (at least with respect to the formation of the four separate transistor structures illustrated in FIGS. 7a-7d). Step 340 includes the formation of a self-aligned mask (not shown) and the etching of the ONO layer 322 and poly1 layer 312 in selective regions, as illustrated in FIG. 9i. The etching step 340 thereby results in a low voltage peripheral transistor 342 having a gate oxide 336 of about 160 Å and a poly2 gate region 338, and a high voltage peripheral transistor 350 having a gate oxide 337 of about 400 Å and a poly2 gate 338. The etching of step 340 further defines the stacked gate structures that form the stacked gate/memory cells 346 and results in select gate transistors 348 having an ONO gate oxide 322 of about 130 Å and a poly2 gate region 338.

[0070] Although the invention has been shown and described with respect to a certain preferred embodiment or embodiments, it is obvious that equivalent alterations and modifications will occur to others skilled in the art upon the reading and understanding of this specification and the annexed drawings. In particular regard to the various functions performed by the above described components (assemblies, devices, circuits, etc.), the terms (including a reference to a “means”) used to describe such components are intended to correspond, unless otherwise indicated, to any component which performs the specified function of the described component (i.e., that is functionally equivalent), even though not structurally equivalent to the disclosed structure which performs the function in the herein illustrated exemplary embodiments of the invention. In addition, while a particular feature of the invention may have been disclosed with respect to only one of several embodiments, such feature may be combined with one or more other features of the other embodiments as may be desired and advantageous for any given or particular application.