Method of fabricating variable length vertical transistors

(1) Field of the Invention

The present invention relates to methods!used to fabricate semiconductor devices, and more specifically to a method used to fabricate vertical transistors featuring variable channel lengths.

(2) Description of Prior Art

The advent of micro-miniaturization, or the ability to fabricate semiconductor devices with sub-micron, or deep sub-micron features, has allowed a greater number of smaller semiconductor chips to be obtained from a specific size starting semiconductor substrate, resulting in a reduction of processing costs for a specific semiconductor chip. The smaller semiconductor chips still offer device densities equal to, or greater than, counterpart semiconductor chips, comprised with larger features. Advances in specific semiconductor disciplines, such as photolithography, have allowed the attainment of sub-micron features to be routinely achieved. The use of more sophisticated exposure cameras, as well as the development of more sensitive photoresist materials, have allowed images of less than 0.25 um to be resolved for photoresist shapes and for features defined using these deep sub-micron photoresist shapes as an etch mask. Still the limiting resolution provided by present photolithographic procedures, restrict the level of miniaturization available for horizontal type, semiconductor devices.

This invention will describe the fabrication of vertical type semiconductor devices, in which the channel length of a metal oxide semiconductor field effect (MOSFET) device is not defined by limiting photolithographic procedures, but by the thickness of deposited materials. This invention will also describe a unique fabrication process for vertical type, complimentary metal oxide semiconductor (CMOS), featuring both P channel (PMOS), and N channel (NMOS), devices, with the ability to form variable length channel lengths. Prior art, such as Hause et al, in U.S. Pat. No. 6,069,384, describe a method of fabricating a vertical transistor structure, however that prior art does not offer the process described in this present invention in which both NMOS and PMOS vertical devices are formed with variable channel lengths.

It is an object of this invention to fabricate vertical, metal oxide semiconductor field effect transistors (MOSFET), devices, on a semiconductor substrate.

It is another object of this invention to form vertical devices featuring variable channel lengths, specific channel lengths for specific devices, with the vertical channel length determined by the thickness of a deposited insulator, or deposited composite insulator layer.

It is still another object of this invention to form CMOS devices, P channel (PMOS), and N channel (NMOS), devices, comprised with variable channel length devices, on the same semiconductor chip.

In accordance with the present invention a method of fabricating vertical CMOS devices on a semiconductor chip, featuring devices comprised with variable channel lengths, is described. A heavily doped N type region, to be used as a drain region for a subsequent vertical NMOS device, is formed in a first region of a semiconductor substrate, followed by formation of a heavily doped P type region, to be used as a drain region for a subsequent vertical PMOS device, in a second region of the semiconductor substrate. After deposition of a first silicon oxide layer and of a first silicon nitride layer, patterning procedures are employed to remove the first silicon nitride layer from the first silicon oxide layer in the PMOS region. Deposition of a second silicon nitride layer and of an overlying second silicon oxide layer results in a first composite insulator layer in the NMOS region comprised of second silicon oxide layer, second silicon nitride layer, first silicon nitride layer, and first silicon oxide layer, while a thinner, second composite insulator layer, comprised of second silicon oxide layer, second silicon nitride layer, and first silicon oxide layer resides in the PMOS region. Channel openings are next formed in the first composite insulator layer exposing a portion of the top surface of the N type drain region, and in the thinner, second composite insulator layer exposing a portion of the top surface of the P type drain region. An epitaxial silicon layer is then selectively grown filling the channel openings. A thin polysilicon layer is then deposited followed by photolithographic masking procedures allowing the portion of the thin polysilicon layer located in the NMOS region to be doped N type via ion implantation procedures, then allowing the portion of thin polysilicon layer, located in the PMOS region, to be doped P type. After deposition of a third silicon oxide layer photolithographic and selective dry etch procedures are used to remove exposed portions of third silicon oxide layer, of N type and P type, thin polysilicon layer, and of second silicon oxide layer, with the selective dry etch procedure terminating at the appearance of the second silicon nitride layer, creating thin polysilicon shapes, to be used as source regions, and creating underlying second silicon oxide shapes to be used as insulator spacers for the subsequent CMOS devices. After removal of the photoresist mask used for definition of the above shapes a wet etch procedure is employed to remove second and first silicon nitride layers in the NMOS region, exposing a first length of selective epitaxial silicon to be used for the NMOS channel region, and to remove second silicon nitride layer in the PMOS region, exposing a shorter length of selective epitaxial silicon to be used for the PMOS channel region. A silicon dioxide gate insulator layer is formed on the channel regions, as well as on the sides of the N type, and P type polysilicon shapes. A thick polysilicon layer is deposited and again doped N type in the NMOS region, and doped P type in the PMOS region. A selective, anisotropic dry etch procedure removes regions of thick polysilicon layer not covered by the overlying third silicon oxide—thin polysilicon—second silicon oxide shapes, resulting in: a vertical NMOS device, comprised of a thick, N type polysilicon shape, surrounding the silicon dioxide gate insulator layer in a region in which the gate insulator layer overlays a channel region comprised of the first length of selective epitaxial silicon; and a vertical PMOS device, comprised of a thick, P type polysilicon shape, surrounding the silicon dioxide gate insulator layer in a region in which the gate insulator layer overlays a channel region comprised of a shorter, second length of selective epitaxial silicon. Each channel region is located between an overlying thin, doped polysilicon source and silicon oxide spacers, and underlying heavily doped drain regions.

The object and other advantages of this invention are best described in the preferred embodiment, with reference to the attached drawings that include:

FIGS. 1-14, which schematically in cross-sectional style show key process stages used to fabricate vertical CMOS devices, featuring devices comprised with variable channel lengths, on the same semiconductor chip.

The method of forming vertical CMOS devices, with the ability to form devices with variable channel lengths on the same semiconductor chip, will now be described in detail. Semiconductor substrate 1 , comprised of P type single crystalline silicon with a < 100 > crystallographic orientation, is used and schematically shown in FIG. 1 . Shallow trench isolation (STI), region 2 , is used to isolate region 50 , of semiconductor substrate 1 , to be used for subsequent N channel (NMOS), devices, from region 60 , of semiconductor substrate 1 , to subsequently used to accommodate P channel (PMOS) devices. STI region 2 , is formed by initially defining a shallow trench shape in semiconductor substrate 1 , via photolithographic and dry etching procedures, followed by filling of the shallow trench shape with a chemically vapor deposited (CVD), insulator layer such as silicon oxide. Chemical mechanical polishing (CMP), or selective, reactive ion etching (RIE), procedures are then used to remove unwanted portions of the filling insulator layer from the top surface of semiconductor substrate 1 , resulting in formation of STI region 2 . A photoresist shape, not shown in drawings, is used as a mask to allow an implantation procedure, using boron ions at an energy between about 2 to 10 KeV, at a dose between about 1E15 to 6E15 atoms/cm 2 , to form heavily doped P type region 3 , in region 60 , of semiconductor substrate 1 , to be used for PMOS devices. Heavily doped P type region 3 , will be used as the heavily doped P type drain for the PMOS devices. After removal of the masking photoresist shape via plasma oxygen ashing, another photoresist shape, again not shown in the drawings, is used as a mask to protect PMOS region 60 , from an implantation procedure used to from heavily doped N type drain region 4 , in region 50 , of semiconductor substrate 1 , where region 50 , is to be used to accommodate NMOS devices. This implantation procedure is performed using arsenic or phosphorous ions, at an energy between about 40 to 90 KeV, at an implant dose between about 1E15 to 8E15 atoms/cm 2 . The result of these procedures is schematically shown in FIG. 1 . Silicon oxide layer 5 , is next deposited at a thickness between about 50 to 200 Angstroms, via low pressure chemical vapor deposition (LPCVD), or via plasma enhanced chemical vapor deposition (PECVD) procedures, followed by the deposition of silicon nitride layer 6 , at a thickness between about 300 to 5000 Angstroms, again via LPCVD or PECVD procedures. The result of these depositions is again schematically shown in FIG. 1 .

Photoresist shape 7 , is next defined and used as an mask to protect NMOS region 50 , from a RIE procedure used to selective remove silicon nitride layer 6 , from the top surface of silicon oxide layer 5 , in PMOS region 60 . The RIE procedure is performed using Cl 2 as an etchant for silicon nitride, with a high etch rate ratio of silicon nitride to silicon oxide in a Cl 2 ambient, allowing the RIE procedure to selectively terminate at the appearance of silicon oxide layer 5 . The result of this procedure is schematically shown in FIG. 2 . After removal of photoresist shape 7 , via plasma oxygen ashing procedures, silicon nitride layer 8 , is conformally deposited at a thickness between about 50 to 500 Angstroms, via LPCVD or PECVD procedures, followed by a conformal deposition of silicon oxide layer 9 , at a thickness between about 50 to 200 Angstroms, again via LPCVD or PECVD procedures. The result of these depositions are schematically illustrated in FIG. 3 . The difference in thickness of the insulator stacks in NMOS. region 50 , and in PMOS region 60 , resulting from the presence of silicon nitride layer 6 , in only NMOS region 50 , will subsequently determine the difference in channel lengths for both device types.

Photoresist shape 10 , is now used as a mask to allow channel opening 11 , in NMOS region 50 , and channel opening 12 , located in PMOS region 60 , to be defined. An anisotropic RIE procedure is employed, first using CHF 3 as an etchant for silicon oxide layer 9 , terminating at the appearance of silicon nitride layer 8 . At this stage of the anisotropic RIE procedure Cl 2 is used to selectively remove silicon nitride layer 8 , and silicon nitride layer 6 , for channel opening 11 , and to remove only silicon nitride layer 8 , for channel opening 12 . This selective dry etch procedure using Cl 2 as the etchant terminates at the top surface of silicon oxide layer 5 , therefore the additional time needed to remove both silicon nitride layers in channel opening 11 , did not result in additional etching in channel opening 12 . The anisotropic RIE procedure is then continued using CHF 3 as an etchant for silicon oxide layer 5 , resulting in NMOS channel opening 11 , exposing a portion of the top surface of heavily doped N type drain region 4 , and resulting in PMOS channel opening 12 , exposing a portion of the top surface of heavily doped P type source region 3 . The diameter of channel opening 11 , and of channel opening 12 , is between about 0.5 to 2 um. This is schematically shown in FIG. 4 . Photoresist shape 10 , is then removed via plasma oxygen ashing procedures.

A buffered hydrofluoric (BHF), or dilute hydrofluoric (DHF), pre-clean procedure is next performed prior to selective growth of intrinsic silicon, in NMOS channel opening 11 , and in PMOS channel opening 12 . Single crystalline silicon is epitaxially grown on the underlying drain regions, at a temperature between about 800 to 1200° C., at a pressure between about 10 to 50 torr, using silane or dichlorosilane as a source. The silicon epitaxial growth, selectively grown only on exposed silicon regions, specifically the drain regions, completely fills the channel openings without forming on the top surface of silicon nitride layer 8 , resulting in silicon cylindrical shape 14 , in NMOS opening 11 , and silicon cylindrical shape 15 , in PMOS channel opening 12 . This is schematically shown in FIG. 5 . The variable length channel regions will be obtained as a result of the difference in depth of the channel openings. The epitaxially grown, undoped cylindrical shapes grow laterally after filling openings 11 and 12 , resulting in a portion of cylindrical shapes 14 and 15 , overlying portions of the top surface of silicon oxide layer 9 . The lateral over growth portion of the cylindrical shapes, initiating after complete filling of openings 11 and 12 , allow the heights of cylindrical shapes 14 and 15 , be determined by the depth of openings 11 and 12 , respectfully. This is schematically shown in FIG. 5 .

Intrinsic polysilicon layer 16 a , shown schematically in FIG. 6, is next deposited via LPCVD procedures to a thickness between about 400 to 1000 Angstroms. A photoresist shape, not shown in the drawings, is next used to block NMOS region 50 , from implantation of boron ions at an energy between about 2 to 10 KeV, and at a dose between about 1E15 to 5E15 atoms/cm 2 , converting exposed portions of intrinsic polysilicon layer 16 a , to P type polysilicon layer 16 c . After removal of the NMOS photoresist block out shape another photoresist shape is used to protect PMOS region 60 , from implantation of arsenic or phosphorous ions, at an energy between about 40 to 70 KeV, and at a dose between about 1E15 to 6E15 atoms/cm 2 , resulting in N type polysilicon layer 16 b , in unprotected portions of intrinsic polysilicon layer 16 a . The result of the procedures used to dope intrinsic polysilicon layer 16 a , is schematically shown in FIG. 7 . Removal of the PMOS photoresist block out shape is again accomplished via plasma oxygen ashing procedures.

Silicon oxide layer 17 , to be used as a hard mask for subsequent patterning procedures, is next deposited via LPCVD or PECVD procedures at a thickness between about 200 to 500 Angstroms. Photoresist shapes 18 , are then formed and used as a mask to allow an anisotropic RIE procedure, using CHF 3 as an etchant, to define the desired source shape in silicon oxide layer 17 . This is schematically shown in FIG. 8 . The anisotropic RIE procedure is then continued using Cl 2 as an etchant to define the N type polysilicon source shape 16 d , and to define P type polysilicon source shape 16 e . The final stage of the anisotropic RIE procedure features definition of silicon oxide spacers 29 , via removal of exposed regions of silicon oxide layer 9 , using CHF 3 as an etchant. This is schematically shown in FIG. 9 .

After removal of photoresist shapes 18 , via plasma oxygen ashing procedures, a hot phosphoric acid solution is used to selectively remove silicon nitride layer 8 , and silicon nitride layer 6 , resulting in exposure of a portion silicon cylindrical shape region 14 , regarded as the NMOS channel region, and of a portion of silicon cylindrical shape 15 , to be used as the PMOS channel region. This is schematically shown in FIG. 10. A thermal oxidation procedure, performed at a temperature between about 800 to 1200° C., in an oxygen—steam ambient, is next used to form silicon dioxide gate insulator layer 19 a , on the NMOS channel region, and to form silicon dioxide gate insulator layer 19 b , on the shorter, or narrower PMOS channel region, both at a thickness between about 10 to 100 Angstroms. The thermal oxidation procedure also results in the growth of silicon oxide layers 19 c , on the exposed sides of N type polysilicon source shape 16 d , and on the exposed sides of P type polysilicon source shape 16 e . This is schematically illustrated in FIG. 11 . If desired a high k dielectric layer such as hafnium oxide (HfO 2 ), or zirconium oxide (ZrO 2 ) at a thickness between about 10 to 100 Angstroms, can be formed via atomic layer deposition (ALD), and used as a gate insulator layer.

Intrinsic polysilicon layer 20 a , is next deposited via LPCVD procedures to a thickness between about 1000 to 2000 Angstroms. A selective RIE procedure, using Cl 2 or SF 6 as an etchant for polysilicon, is used to remove the top portion of intrinsic polysilicon layer 20 a , to a point at which the top surface of silicon oxide shape 17 , located in PMOS region 60 , is exposed, while the top surface o the remaining portion of intrinsic polysilicon layer 20 a , in NMOS region 50 , is at a level adjacent to silicon oxide spacers 29 . This is schematically shown in FIG. 12. A photoresist shape, not shown in the drawings, is next used to block NMOS region 50 , from implantation of boron ions performed at an energy between about 20 to 50 KeV, and at a dose between about 1E15 to 5E15 atoms/cm 2 , converting exposed portions of intrinsic polysilicon layer 20 a , to P type polysilicon layer 20 c . After removal of the NMOS photoresist block out shape another photoresist shape is used to protect PMOS region 60 , from implantation of arsenic or phosphorous ions, performed at an energy between about 50 to 100 KeV, and at a dose between about 1E15 to 6E15 atoms/cm 2 , resulting in N type polysilicon layer 20 b , in unprotected portions of intrinsic polysilicon layer 20 a . The result of the procedures used to dope intrinsic polysilicon layer 20 a , is schematically shown in FIG. 13 . The photoresist shape, used to block out PMOS region during the N type doping of intrinsic polysilicon layer 20 a , is then removed via plasma oxygen ashing procedures.

A self-aligned dry etch procedure, performed without photoresist masking, is next employed to remove portions of doped polysilicon, using silicon oxide shapes 17 , as a hard mask. The dry etch procedure is an anisotropic RIE procedure, using Cl 2 or SF 6 as an etchant for polysilicon. A high etch rate ratio of polysilicon to silicon oxide, between about 10 to 100, to 1, allows this selective dry etch procedure to terminate at the top surface of silicon oxide layer 5 . The result of this procedure, schematically shown in FIG. 14, is the formation of N type polysilicon gate structure 20 b , in NMOS region 50 , and the formation of P type polysilicon gate structure 20 c , in PMOS region 60 . The vertical CMOS devices are now comprised of channel regions 14 and 15 ; silicon dioxide gate insulator layers 19 a and 19 b ; doped polysilicon gate structures 20 b and 20 c ; heavily doped drain regions 3 and 4 , doped polysilicon source shapes 16 d and 16 e ; and silicon oxide spacers 29 . The channel length for a specific CMOS device was determined by the thickness of the composite insulator layer in which the channel opening was defined in. Therefore to obtain devices with variable channel lengths various procedures, such as deposition and selective patterning, can be employed to obtain the desired menu of channel lengths. Although this invention was described with for a PMOS device comprised with a shorter channel length than the channel length of the NMOS device, it should be understood that any combination of variable channel lengths are obtainable via use of this invention.

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 or scope of the invention.