Method of preserving the top oxide of an ONO dielectric layer via use of a capping material

[0002] (1) Field of the Invention

[0003] The present invention relates to methods used to fabricate semiconductor devices and more specifically to a method used to protect a dielectric layer, used for a non-volatile memory, (particularly in EEPROM) application, during fabrication of peripheral embedded memory devices.

[0004] (2) Description of Prior Art

[0005] Non-volatile memory applications are being fabricated featuring an oxidized nitride on silicon oxide (ONO), dielectric layer formed on an underlying floating gate structure. The data retention characteristics of the non-volatile memory device is partially influenced by integrity of top portion of the ONO dielectric layer, the thin top oxide component obtained via the oxidation of a top portion of a silicon nitride layer. The thin oxide component controls the barrier height between the nitride component of the ONO dielectric and an overlying control gate structure, minimizing the charge loss from the underlying floating gate. However when peripheral high and low voltage devices are integrated with the non-volatile memory device, the critical top oxide component of the ONO dielectric layer can be exposed to processing sequences which can result in unwanted removal of portions of the top oxide layer, resulting in data retention degradation.

[0006] This invention will describe process sequences featuring materials which allow capping and protection of the top oxide component of the ONO dielectric layer to be achieved, therefore avoiding oxide loss during specific processes used in the integration of peripheral memory devices with the non-volatile memory device. In addition the thickness of the top oxide can be increased, as depicted in a second iteration of this invention. Prior art such as Lai et al in U.S. Pat. No. 6,448,126, as well as Camerlenghi in U.S. Pat. No. 6,420,223, describe methods of protecting ONO layers during some subsequent processing sequences. However these prior art do not describe the novel process sequence offered in the present invention wherein disposable capping materials are employed to protect the thin top oxide component of an ONO dielectric layer present on the non-volatile device, during specific fabrication sequences used for embedded peripheral memory devices.

[0007] It is an object of this invention to integrate the fabrication of a non-volatile memory device with the fabrication of low voltage and high voltage complimentary metal oxide semiconductor (CMOS), devices.

[0008] It is another object of this invention to protect the dielectric layer located on the floating gate structure of the non-volatile memory device during hydrofluoric acid type procedures used with the fabrication sequence for the low and high voltage CMOS devices.

[0009] It is still another object of this invention to employ a disposable silicon nitride layer, or a disposable polysilicon layer, on the floating gate structure to protect the dielectric layer on the floating gate structure so that hydrofluoric acid type cleaning procedures can be used with the fabrication of the low and high CMOS devices, resulting in better gate oxide integrity (GOI).

[0010] In accordance with the present invention a method of integrating the fabrication of a non-volatile memory device with the fabrication of low voltage and high voltage complimentary metal oxide semiconductor (CMOS), devices, featuring the use of a capping layer used to protect the dielectric layer located on the floating gate structure of the non-volatile memory device, during hydrofluoric acid type procedures, either dilute hydrofluoric (DHF), or buffered hydrofluoric (BHF), performed during the fabrication sequence of the low and high voltage CMOS devices, is described. A first iteration of this invention entails the formation of the floating gate structure of the non-volatile device in a first region of a semiconductor substrate, followed by formation of an overlying oxidized silicon nitride on silicon oxide (ONO), dielectric layer. After deposition of a silicon nitride capping layer definition of the silicon nitride layer and of the ONO dielectric layer is only on the floating gate structure in the first region of the semiconductor substrate. After a hydrofluoric acid type pre-clean procedure a first gate insulator layer is thermally grown on exposed semiconductor portions in second and third regions, followed by removal of the first gate insulator layer from the second region of semiconductor substrate. The silicon nitride cap layer is then disposed, removed by dipping into phosphoric acid (H₃PO₄). After removal of the capping silicon nitride layer, located overlying the ONO dielectric layer located on the floating gate structure, another clean procedure, without hydrofluoric acid, is performed followed by the growth of a second gate insulator layer on the exposed portions of the second region of semiconductor substrate, with the second gate insulator layer thinner than the first gate insulator layer. Completion of integrated fabrication procedure is accomplished via simultaneous formation of a polysilicon control gate structure on the ONO dielectric layer of the non-volatile memory device located in the first region of the semiconductor substrate, and of polysilicon gate structures in the second and third regions of the low and high voltage CMOS devices respectfully.

[0011] A second iteration of this invention features definition of a polysilicon capping layer on an underlying ONO dielectric layer, located on a floating gate structure of a non-volatile memory device in a first region of a semiconductor substrate. After a hydrofluoric acid type pre-clean procedure a first oxidation procedure is performed resulting in the formation of a first gate insulator layer on second and third regions of the semiconductor substrate, while partial oxidation of a top portion of the polysilicon capping layer occurs. After removal of the first gate insulator layer from the second region of the semiconductor substrate via a buffered hydrofluoric (BHF) acid type procedure, a second oxidation procedure is performed resulting in the formation of a second gate insulator layer in the second region of the semiconductor substrate, with the second gate insulator layer thinner than the first gate insulator layer, and with the second oxidation procedure resulting in total oxidation of the polysilicon capping layer located on the ONO dielectric layer. This results in self removal of the disposable polysilicon cap layer. Again completion of the integrated fabrication procedure is accomplished via simultaneous formation of a polysilicon control gate structure on the oxidized polysilicon capping layer, located overlying the ONO dielectric layer of the non-volatile memory device, and of polysilicon gate structures on the second and third regions of the low and high voltage CMOS devices respectfully.

[0012] The object and other advantages of this invention are best described in the preferred embodiments with reference to the attached drawings that include:

[0013] FIGS. 1-10, which schematically, in cross-sectional style, illustrate a first iteration of this invention, wherein a silicon nitride capping layer is used to protect the dielectric layer of a non-volatile memory device during hydrofluoric acid type procedures performed during fabrication of embedded low and high voltage CMOS devices.

[0014] FIGS. 11-16, which schematically, in cross-sectional style, illustrate a second iteration of this invention, wherein a self removing polysilicon capping layer is used to protect the dielectric layer of a non-volatile memory device during hydrofluoric acid type procedures performed during fabrication of embedded low and high voltage CMOS devices.

[0015] Methods of integrating the fabrication of a non-volatile memory device with the fabrication of embedded low voltage and high voltage complimentary metal oxide semiconductor (CMOS), devices, featuring the use of a capping layer used to protect the dielectric layer located on the floating gate structure of the non-volatile memory device during hydrofluoric acid type procedures, such as dilute hydrofluoric (DHF), performed during the fabrication sequence of the low and high voltage CMOS devices, will now be described in detail. Semiconductor substrate 4, comprised of single crystalline silicon and featuring a <100> crystallographic orientation, is used and schematically shown in FIG. 1. Region 1, of semiconductor substrate 4, will be used to accommodate a non-volatile memory device, while regions 2 and 3, will be used to accommodate a low voltage CMOS device as a high voltage CMOS device, respectfully. Isolation between devices is provided by regions 5, comprised of either thermally grown field oxide (FOX), regions, or insulator filled shallow trench isolation (STI), regions. Referring to FIG. 1, the floating gate structure 7, and an underlying, thin silicon dioxide layer 6, are formed in non-volatile memory region 1. Silicon dioxide layer 6, to be used as the tunnelling insulator layer of the non-volatile memory device is obtained at a thickness between about 75 to 95 Angstroms, via thermal oxidation procedures. A polysilicon layer is next deposited at a thickness between about 2000 to 4000 Angstroms via low pressure chemical vapor deposition (LPCVD), procedures. The polysilicon layer can be doped in situ during deposition via the addition of either arsine, phosphine, or diborane, to a silane or disilane ambient. If desired the polysilicon layer can be deposited intrinsically then doped via implantation of arsenic, phosphorous, or boron ions. A photoresist shape, not shown in the drawings, is used as an etch mask to allow polysilicon floating gate structure 7, to be defined on silicon dioxide layer 6, in region 1, via anisotropic reactive ion etching (RIE), procedures using Cl₂or SF₆as an etchant for polysilicon. Removal of the photoresist shape is accomplished via plasma oxygen ashing and wet clean procedures, with a buffered hydrofluoric acid component of the wet clean procedure removing the portions of silicon dioxide layer 6, not covered by floating gate structure 7. The result of these procedures is schematically shown in FIG. 1.

[0016] Formation of ONO layer 11, is next addressed and schematically shown using FIGS. 2-3. Silicon dioxide layer 8, is first obtained at a thickness between about 100 to 200, on floating gate structure 7, via thermal oxidation procedures. The portion of silicon dioxide layer 8, on the surfaces of floating gate structure 7, is obtained via thermal oxidation of polysilicon, while other portions of silicon dioxide layer 8, are obtained via thermal oxidation of semiconductor substrate 4. Silicon nitride layer 9a, shown schematically in FIG. 2, is next deposited via LPCVD procedures to a thickness between about 100 to 200 Angstroms. A thermal oxidation procedure is next performed at a temperature between about 900 to 1000° C., in an oxygen-steam or in an oxygen ambient, forming oxidized silicon nitride (ON), or silicon oxide layer 10, at a thickness between about 30 to 50 Angstroms, on the remaining bottom portion of silicon nitride, silicon nitride layer 9b, now at a reduced thickness between about 60 to 150 Angstroms. The result of the above procedures result in the formation of ONO, (oxidized silicon nitride on silicon oxide), layer 11, on floating gate structure 7, as well as on other surfaces of regions 1, 2, and 3. This is schematically shown in FIG. 3.

[0017] Silicon oxide layer 10, of ONO dielectric layer 11, is critical for the data retention requirement of a non-volatile memory device, and therefore the ability to maintain the thickness of silicon dioxide layer 10, during subsequent processing sequences used for fabrication of the low and high voltage CMOS devices, is paramount. If left exposed, or not protected during specific subsequent processing steps, unwanted thinning of silicon oxide layer 10, would occur, deleteriously influencing the reliability and yield of the non-volatile memory device. Therefore a first iteration of this invention will describe a process sequence in which a silicon nitride capping layer is used to protect ONO layer 11, during subsequent processing procedures. Silicon nitride capping layer 12, shown schematically in FIG. 4, is deposited at a thickness between about 100 to 200 Angstroms via LPCVD procedures. Photoresist shape 13, shown schematically in FIG. 5, is used as an etch mask to allow removal of exposed portions of silicon nitride capping layer 12, and of ONO layer 11. This is accomplished via anisotropic RIE procedures, using CF₄or Cl₂as an etchant for silicon nitride capping layer 12, and for silicon nitride layer 9b, while CHF₃is used as an etchant for oxidized silicon nitride layer 10, and for silicon oxide layer 8. The result of this procedure results in ONO layer 11, overlaid by silicon nitride capped layer 12, on floating gate structure 7, in non-volatile memory region 1. This is schematically shown in FIG. 5. Removal of photoresist shape 13, is accomplished via plasma oxygen ashing procedures.

[0018] Preparation for growth of silicon dioxide layer 14, to be used a gate insulator layer for a high voltage CMOS device, initiates with a pre-clean procedure performed using an RCA clean with a final cycle using a dilute hydrofluoric (DHF), solution, employed to remove native oxide from the surface of semiconductor substrate 4, in regions 2 and 3. If left unremoved the native oxide would be incorporated into a subsequently grown silicon dioxide gate insulator, compromising the thickness and quality of the silicon dioxide gate insulator layer. The presence of silicon nitride capping layer 12, on ONO layer 11, protects the thin, oxidized silicon nitride component 10, during the DHF pre-clean procedure. After the DHF pre-clean procedure silicon dioxide layer 14, is thermally grown in an oxygen-steam ambient, to a thickness between about 150 to 250 Angstroms. Silicon dioxide layer 14, shown schematically in FIG. 6, will be used as the gate insulator layer for a high voltage CMOS device, to be formed in region 3, of semiconductor substrate 4.

[0019] Photoresist shape 20, is next formed and used as an mask to allow silicon dioxide layer 14, to be removed only from the surface of semiconductor substrate 4, in region 2, the region in which a low voltage CMOS device will be fabricated. Removal of this portion of silicon dioxide layer 14, is accomplished via wet etch procedures using a buffered hydrofluoric (BHF), solution as an etchant for silicon dioxide layer 14. This is schematically shown in FIG. 7. After removal of photoresist shape 20, via plasma oxygen ashing procedures, silicon nitride capping layer 12, is removed from the surface of ONO layer 11, via use of a hot phosphoric acid solution. This is schematically shown in FIG. 8. After a RCA clean procedure, with or without exposure to a final short DHF procedure, resulting in minimal thickness loss of oxidized silicon nitride layer 10, silicon dioxide gate insulator layer 15, is grown. Silicon dioxide layer 15, is thermally grown in an oxygen-steam ambient, at a thickness between about 50 to 80 Angstroms, on the surface of semiconductor substrate 4, in region 2. Silicon dioxide layer 15, shown schematically in FIG. 9, will be used as the gate insulator layer for a low voltage CMOS device. Any thickness loss of silicon dioxide layer 14, in region 3, occurring during the pre-clean procedure employed prior to growth of silicon dioxide layer 15, could be regrown during the thermal oxidation procedure used for attainment of silicon dioxide layer 15.

[0020] Completion of the non-volatile memory device in region 1, of the low voltage CMOS device in region 2, and of the high voltage CMOS device in region 3, is next addressed and schematically shown in FIG. 10. A polysilicon layer is deposited to a thickness between about 2000 to 4000 Angstroms, via LPCVD procedures, with the polysilicon layer again being doped either in situ during deposition, or via ion implantation procedures after an intrinsic deposition. A photoresist shape, not shown in the drawings, is used as an etch mask to allow an anisotropic RIE procedure to define polysilicon control gate structure 22a, on ONO layer 11, for the non-volatile memory device, in addition to simultaneous definition of polysilicon gate structure 22b, on silicon dioxide gate insulator layer 15, for the low voltage CMOS device, and of polysilicon gate structure 22c, on silicon dioxide gate insulator layer 14, for the high voltage CMOS device. The anisotropic RIE procedure is performed using Cl₂or SF₆as a selective etchant for polysilicon. Removal of the photoresist shape used to define the polysilicon gate structures is accomplished via plasma oxygen ashing procedures. The formation of other elements of these devices, such as insulator spacers on the sides of the polysilicon gate structures, as well as source/drain regions in areas of semiconductor substrate 4, not covered by the gate structures, will not be described.

[0021] A second iteration of this invention, described using FIGS. 11-16, features the use of a polysilicon capping layer for protection of the oxidized silicon nitride component of an ONO layer during subsequent hydrofluoric acid procedures. After formation of ONO layer 11, on floating gate structure 7, in non-volatile memory region 1, as well as on all areas of regions 2 and 3, used for the low and high CMOS devices, polysilicon layer 16a, is deposited via LPCVD procedures to a thickness between about 100 to 300 Angstroms. Polysilicon layer 16a, shown schematically in FIG. 11, to be used as a capping layer for ONO layer 11, is deposited intrinsically. A photoresist shape, not shown in the drawings, is used as an etch mask, allowing an anisotropic RIE procedure to define a polysilicon capped ONO layer on the floating gate structure in non-volatile memory region 1. The anisotropic RIE procedure used to remove the portions of polysilicon layer 16a, and the portions of ONO layer 11, not covered by the photoresist shape in regions 2 and 3, is performed using Cl₂as an etchant for polysilicon layer 16a, Cl₂or CF₄as an etchant for the unoxidized portion of silicon nitride in the ONO layer, and CHF₃as an etchant for the silicon oxide component and for the oxidized portion of the silicon nitride of the same ONO layer. The photoresist shape used for definition of the polysilicon capping layer and for the ONO layer is removed via plasma oxygen ashing procedures. The result of these procedures is schematically shown in FIG. 12.

[0022] Preparation for the growth of silicon dioxide layer 14, to be used a gate insulator layer for a high voltage CMOS device, again as was the case in the first iteration of this invention initiates with a pre-clean procedure performed using a RCA clean procedure with exposure to a final dilute hydrofluoric (DHF), solution, employed to remove native oxide from the surface of semiconductor substrate 4, in regions 2 and 3. If left unremoved the native oxide would be incorporated into a subsequently grown silicon dioxide gate insulator, compromising the thickness and quality of the silicon dioxide gate insulator layer. The presence of polysilicon capping 16a, on ONO layer 11, protects the thin, oxidized silicon nitride component 10, during the DHF pre-clean procedure. After the DHF pre-clean procedure silicon dioxide layer 14, is thermally grown in an oxygen-steam ambient, to a thickness between about 150 to 250 Angstroms. Silicon dioxide layer 14, shown schematically in FIG. 13, will be used as the gate insulator layer for a high voltage CMOS device, to be formed in region 3, of semiconductor substrate 4. Exposure of polysilicon capping layer 16a, to the thermal oxidation procedure used to grow silicon dioxide layer 14, also results in oxidation of a top portion of the polysilicon capping layer. Silicon oxide layer 17a, at a thickness between about 100 to 200 Angstroms is now located overlying a thinned polysilicon capping layer 16b, now at a thickness between about 50 to 100 Angstroms. This is schematically shown in FIG. 13.

[0023] Photoresist shape 18, is next formed and used as an mask to allow silicon dioxide layer 14, to be completely removed from the surface of semiconductor substrate 4, in region 2, the region in which a low voltage CMOS device will be fabricated. Removal of this portion of silicon dioxide layer 14, is accomplished via wet etch procedures using either a BHF, solution as an etchant for silicon dioxide layer 14. If desired photoresist shape 18, can also cover non-volatile memory region 1, during the DHF or BHF procedure. This is schematically shown in FIG. 14. However if left uncovered silicon oxide layer 17a, would be removed but thinned polysilicon capping layer 16b, would still be present to protect the oxidized silicon nitride component of the ONO layer from the etch step.

[0024] After removal of photoresist shape 18, via plasma oxygen ashing procedures an RCA clean procedure with exposure to a final DHF preclean cycle can be employed to prepare the semiconductor surface in region 2, for growth of silicon dioxide layer 15. The DHF procedure results in either partial or total removal of silicon oxide layer 17a, the layer overlying thinned polysilicon capping layer 16b. Silicon dioxide layer 15, is then thermally grown in an oxygen-steam ambient, to a thickness between about 50 to 80 Angstroms, on the surface of semiconductor substrate 4, in region 2. Silicon dioxide layer 15, shown schematically in FIG. 15, will be used as the gate insulator layer for a low voltage memory device. The above oxidation procedure should result in complete oxidation of the thinned polysilicon capping layer, resulting in a silicon oxide layer 17b, shown schematically in FIG. 15, now at a thickness between about 50 to 100 Angstroms, overlying ONO layer 11. This additional oxide adds on to the top oxide of the ONO layer and thus helps to improve the data retention (otherwise not possible as oxidation of nitride can only produce very thin top oxide (about 50 Angstroms).

[0025] Completion of the non-volatile memory device in region 1, of the low voltage CMOS device in region 2, and of the high voltage CMOS device in region 3, is accomplished using identical processing sequences used in the first iteration of this invention. A polysilicon layer is deposited to a thickness between about 2000 to 4000 Angstroms, via LPCVD procedures, with the polysilicon layer being doped either in situ during deposition, or via ion implantation procedures after an intrinsic deposition. A photoresist shape, not shown in the drawings, is used as an etch mask to allow an anisotropic RIE procedure to define polysilicon control gate structure 22a, on ONO layer 11, (overlaid with thicker silicon oxide layer 17b), for the non-volatile memory device, in addition to simultaneously defining polysilicon gate structure 22b, on silicon dioxide gate insulator layer 15, for the low voltage CMOS device, and polysilicon gate structure 22c, on silicon dioxide gate insulator layer 14, for the high voltage CMOS device. The anisotropic RIE procedure is performed using CHF₄as an etchant for silicon oxide layer 17b, Cl₂or SF₆as a selective etchant for polysilicon. Removal of the photoresist shape used to define the polysilicon gate structures is accomplished via plasma oxygen ashing procedures. The formation of other elements of these devices, such as insulator spacers on the sides of the polysilicon gate structures, as well as source/drain regions in areas of semiconductor substrate 4, not covered by the gate structures, will not be described.

[0026] While this invention has been particularly shown and described with reference to the preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the spirit and scope of this invention.