Strained channel CMOS device with fully silicided gate electrode

This invention generally relates to formation of MOSFET devices in integrated circuit manufacturing processes and more particularly to a strained channel NMOS and PMOS device pair and method of forming the same including fully silicided gate electrodes to improve a drive current.

As is well known, increased device density, together with higher speed performance and lower power consumption are major driving forces in integrated circuit manufacturing. CMOS design consideration for high speed digital applications are usually determined by the pull up time and pull down time of each individual gate. Individual gates are associated with a delay time period for signal propagation in PMOS and NMOS gate electrodes. The delay time period, in turn, is inversely proportional to the drive current (I drive ). It is therefore clear that maximizing the drive current will increase the performance speed or Figure of Merit (FOM) of a CMOS device.

Mechanical stresses are known to play a role in charge carrier mobility which affects several critical parameters including Voltage threshold (V T ) shift, drive current saturation (I Dsat ), and ON/Off current. The effect of induced mechanical stresses to strain a MOSFET device channel region, and the effect on charge carrier mobility is believed to be influenced by complex physical processes related to acoustic and optical phonon scattering. Ideally, an increase in charge carrier mobility will also increase a drive current.

In addition, drive current is affected by gate sheet resistance. Thus the higher the sheet resistance of the gate electrode, the larger the delay time in signal propagation. Approaches in the prior art to reduce gate electrode sheet resistance have included forming silicides in an upper portion of a polysilicon gate electrode as well as forming the gate electrode of a conductive metal.

In addition, due to complex relationships between the thickness of a silicide, which remains about constant with device scaling, and scaling size reductions of the CMOS device, for example including a junction depth of source and drain regions, the problem of current leakage (diode leakage) becomes increasingly problematical at smaller device critical dimensions. Thus, prior art approaches of forming silicided gate electrodes and silicided source and drain regions increasingly leads to short channel effects including current leakage.

Conventional silicided gate electrodes of the prior art have the added liability of suffering poly-depletion effects. For example, when a gate bias is applied to the CMOS device the electrical field formed on the gate dielectric penetrates into the gate electrode causing charge carrier depletion at the electrode/gate interface which thereby decreases drive current and lowers CMOS speed performance.

These and other shortcomings demonstrate a need in the semiconductor device integrated circuit manufacturing art for improved CMOS devices and methods of manufacturing the same to achieve improved CMOS device speed performance including increased drive current.

It is therefore an object of the present invention to provide improved CMOS devices and methods of manufacturing the same to achieve improved CMOS device speed performance including increased drive current, while overcoming other shortcomings of the prior art.

To achieve the foregoing and other objects, and in accordance with the purposes of the present invention, as embodied and broadly described herein, the present invention provides a strained channel NMOS and PMOS device pair including fully silicided gate electrodes and method for forming the same.

In a first embodiment, the method includes A strained channel NMOS and PMOS device pair including fully silicided gate electrodes and method for forming the same, the method including providing a semiconductor substrate including NMOS and PMOS device regions including respective gate structures including polysilicon gate electrodes; forming recessed regions on either side of a channel region including at least one of the NMOS and PMOS device regions; backfilling portions of the recessed regions with a semiconducting silicon alloy to exert a strain on the channel region; forming offset spacers on either side of the gate structures; thinning the polysilicon gate electrodes to a silicidation thickness to allow full metal silicidation through the silicidation thickness; ion implanting the polysilicon gate electrodes to adjust a work function; and, forming a metal silicide through the silicidation thickness to form metal silicide gate electrodes.

These and other embodiments, aspects and features of the invention will be better understood from a detailed description of the preferred embodiments of the invention which are further described below in conjunction with the accompanying Figures.

FIGS. 1A–1H are cross sectional schematic views of exemplary portions of a CMOS device including NMOS and PMOS devices at stages of manufacture according to an embodiment of the present invention.

FIG. 2 is a process flow diagram including several embodiments of the present invention.

Although the method of the present invention is explained with reference to exemplary NMOS and PMOS MOSFET devices, it will be appreciated that aspects of the present invention may be applied to the formation of any MOSFET device including dual gate or twin gate CMOS inverters where a fully silicided gate electrode is formed to reduce a gate electrode resistance and avoid a poly-depletion effect. It will be appreciated that additional benefits are realized by forming a strained channel region to improve charge carrier mobility, both processes advantageously improving device performance including drive current (I dsat ) and device speed.

Referring to FIGS. 1A–1H in an exemplary process flow for forming the fully silicided gate electrode CMOS structures of the present invention, are shown cross-sectional schematic views of a portion of a semiconductor wafer at stages of an exemplary production process.

For example, referring to FIG. 1A , is shown a semiconductor substrate 12 , which may include silicon, strained semiconductor, compound semiconductor, multi-layered semiconductors, or combinations thereof. For example, the substrate 12 may include, but is not limited to, silicon on insulator (SOI), stacked SOI (SSOI), stacked SiGe on insulator (S—SiGeOI), SiGeOI, and GeOI, or combinations thereof. For example, the substrate may include doped well regions 12 A and 12 B making up respective NMOS and PMOS device regions formed by conventional methods, for example a masking process followed by ion implantation and activation annealing. Electrical isolation regions optionally separate the NMOS and PMOS device regions, for example, for forming single gate devices and are preferably shallow trench isolation (STI) structures e.g., 14 , formed by trench etching and backfilling with an oxide dielectric, for example TEOS silicon oxide, followed by planarization. It will be appreciated that dual or twin gate devices may be formed as well.

Still referring to FIG. 1A , gate structures are formed by conventional processes including a gate dielectric portion e.g., 16 A and 16 B and overlying polysilicon gate electrode portions e.g., NMOS device gate electrode 18 A and PMOS device gate electrode 18 B. In an important aspect of the invention the polysilicon gate electrodes are formed by first forming a gate dielectric layer followed by deposition of an undoped polysilicon layer. A hardmask layer of silicon nitride and/or silicon oxynitride is then deposited over the undoped polysilicon layer by conventional CVD processes, e.g., LPCVD or PECVD followed by photolithographic patterning and plasma assisted etching (e.g., RIE) to form the respective NMOS and PMOS gate structures with remaining overlying hardmask layer portions e.g., 20 A and 20 B.

The gate dielectric portions e.g., 16 A and 16 B may be formed of silicon oxide, silicon oxynitride, silicon nitride, nitrogen doped silicon oxide, high-K dielectrics, or combinations thereof. The high-K dielectrics may include metal oxides, metal silicates, metal nitrides, transition metal-oxides, transition metal silicates, metal aluminates, and transition metal nitrides, or combinations thereof. The gate dielectric portions e.g., 16 A and 16 B may be formed by any process known in the art, e.g., thermal oxidation, nitridation, sputter deposition, or chemical vapor deposition. The physical thickness of the gate dielectric portions e.g., 16 A and 16 B may be in the range of 5 to 100 Angstroms. When using a high permittivity (high-K) gate dielectric, the dielectric constant is preferably greater than about 3.9. The high-K dielectric may be selected from a group comprising aluminum oxide (Al 2 O 3 ), hafnium oxide (HfO 2 ), hafnium oxynitride (HfON), hafnium silicate (HfSiO 4 ), zirconium oxide (ZrO 2 ), zirconium oxynitride (ZrON), zirconium silicate (ZrSiO 2 ), yttrium oxide (Y 2 O 3 ), lanthanum oxide (La 2 O 3 ), cerium oxide (CeO 2 ), titanium oxide (TiO 2 ), tantalum oxide (Ta 2 O 5 ), or combinations thereof.

Still referring to FIG. 1A , nitride offset liners e.g., 22 A and 22 B, are formed along either side of the gate respective NMOS and PMOS gate structure sidewalls by depositing one or more layers of silicon nitride (e.g., Si 3 N 4 ) and/or silicon oxynitride (e.g., SiON), followed by wet or dry etching away portions of the one or more layers to form self-aligned offset spacers on either side of the gate structures. The term ‘nitride’ as used herein is intended to include both nitrides and oxynitrides.

Referring to FIG. 1B , in an important aspect of the invention, a protective oxide layer portion e.g., 24 is formed overlying the NMOS device region e.g., 12 A, including the gate structure. For example, a silicon oxide layer is first formed over the process surface by a conventional CVD deposition process using a furnace or rapid temperature process (RTP), also referred to as rapid process oxidation. The protective oxide layer is then patterned by a conventional lithographic process and a wet and/or dry silicon oxide etching process is then carried out to leave the protective oxide layer portion e.g., 24 overlying the NMOS device region 12 A and leave portions of the substrate 12 exposed adjacent the PMOS gate structure over the PMOS device regions 12 B.

Referring to FIG. 1C , the substrate in the PMOS region 12 B is then subjected to a conventional wet or dry etching process, preferably a dry etching process, to etch recessed areas e.g., 25 A and 25 B into exposed portions of the substrate 12 on adjacent the PMOS gate structure and on either side of the PMOS channel region 26 B. The recessed areas e.g., 25 A and 25 B subsequently form portions of source and drain (S/D) regions in a completed CMOS (PMOS) device. The depth of the recessed areas will vary depending on the desired level of stress desired to be subsequently exerted on the channel region e.g., 26 B disposed between the recessed areas as explained below, for example from about 10 Angstroms to about 800 Angstroms, more preferably from about 200 to about 400 Angstroms.

Referring to FIG. 1D , a strained silicon alloy composite e.g., 28 A including silicon and an element having a larger atomic radius is deposited to form a first strained semiconducting composite to backfill a first portion of the first recessed areas 25 A and 25 B. In a preferred embodiment, the first strained silicon alloy composite 28 A is formed of SiGe, and is grown in the recessed areas 25 A and 25 B by a conventional SiGe growth process, e.g., epitaxial. It will be appreciated that the upper level of the backfilled strained silicon alloy composite 28 A may be formed at about the same level, or slightly above or below the silicon substrate 12 level, e.g., from about 1 to 20 Angstroms, but is preferably formed slightly below the substrate level. The first strained silicon alloy composite advantageously exerts a compressive stress on the channel region 26 B disposed between the backfilled recessed regions 25 A and 25 B. Still referring to FIG. 1D , an epitaxially grown silicon layer 28 B is then deposited on the strained silicon alloy composite 28 A to complete filling of the recessed areas 25 A and 25 B. The protective oxide layer 24 is then removed, preferably by a wet stripping process, e.g., dilute HF.

Referring to FIG. 1E , conventional ion implantation processes are carried out form LDD doped regions including pocket implant regions (not shown) adjacent the gate electrodes. Sidewall spacers e.g., 32 A and 32 B, are then formed adjacent the nitride offset spacers. For example in the exemplary embodiment as shown, oxide liners e.g., 30 A, 30 B are first formed over the nitride offset spacers e.g., 22 A, 22 B by conventional deposition and etch processes followed by formation of an outermost silicon nitride or silicon oxynitride spacer e.g., 32 A and 32 B. It will be appreciated that other types of spacers may be used including outermost oxide spacers or spacers with substantially vertical sidewalls. It will be appreciated that the outermost spacers, when formed of a nitride and/or oxynitride may optionally be formed in a compressive or tensile stress, preferably a tensile stress for the NMOS device and compressive stress for the PMOS device to introduce a strain into the channel region for improved charge mobility. Preferably, the sidewall spacers 32 A and 32 B are formed at a height lower than the hardmask layer level, e.g., 20 A, 20 B by conventional etchback processes to remain intact following a subsequent CMP process explained below. A conventional S/D ion implant is then carried out to form S/D regions (not shown) adjacent the sidewall spacers.

Still referring to FIG. 1E , metal silicide regions are formed in a conventional self-aligned silicided process (salicide) to form metal silicides e.g., e.g., 34 A, 34 B over the source and drain regions. For example the metal suicides may be formed by depositing a metal, e.g., Ti, Co, W, Ni, or Pt, most preferably Co or Ni, over the process surface followed by an RTP annealing process to form the low resistance phase of the silicide e.g., TiSi 2 , CoSi 2 , WSi 2 , NiSi, or PtSi.

Still referring to FIG. 1E a contact etch stop layer 36 , preferably silicon nitride and/or silicon oxynitride is then blanket deposited over the process surface to cover both the NMOS and PMOS device regions. It will be appreciated that the contact etch stop layer 36 may optionally be formed in a compressive or tensile stress, for example a tensile stress to improve the charge mobility of the NMOS device and a compressive stress to improve the charge mobility of the PMOS device by introducing strain into the channel regions e.g., 26 A and 26 B for improved charge mobility.

Referring to FIG. 1F , in an important aspect of the invention, an ILD (dielectric insulating layer) 38 A is formed over the process surface by a conventional CVD or spin-on process, followed by a chemical mechanical polish (CMP) process, where the hardmask layer portions e.g., 20 A and 20 B are removed and the undoped polysilicon upper portion of the gate electrodes e.g., 18 A, 18 B is exposed. The ILD layer 38 A may be formed of P doped silicate glass (PSG), PECVD silicon oxide, PETEOS, BPTEOS, BTEOS, or PTEOS, more preferably PSG.

Referring to FIG. 1G , in an important aspect of the invention, a conventional dry polysilicon etchback process is carried out to etchback a thickness portion of the gate electrodes e.g., 18 A and 18 B, followed by a conventional wet cleaning process, e.g., using SC1 and/or SC2 leaning solutions. It will be appreciated that the remaining thickness portions of the polysilicon gate electrodes following etchback may be between about 100 Angstroms and about 1000 Angstroms in thickness, more preferably between about 200 Angstroms to about 500 Angstroms, preferably being sufficiently thin to allow full metal silicidation of the remaining thickness portions.

Prior to gate electrode silicidation, in an important aspect of the invention, an ion implant process is carried out to dope the respective NMOS and PMOS gate electrodes e.g., 18 A and 18 B with respective conventional N and P dopants to a sufficient concentration to adjust the polysilicon work function to avoid polysilicon/gate dielectric depletion layer formation during device operation. Preferably, the work function of the NMOS gate electrode is adjusted to between about 4.0 and about 4.5 eV while the PMOS gate electrode is adjusted to between about 4.5 and about 5.0 eV. Following ion implantation, a gate electrode silicidation process is carried out to form a metal silicide extending through the remaining thickness portion of the gate electrodes 18 A and 18 B. The same preferred metals and processes as outlined for the source/drain silicidation process is preferably carried out, most preferably forming either CoSi 2 or NiSi gate electrodes.

Referring to FIG. 1H , at least one additional ILD layers e.g., 38 B is deposited over the first ILD layer 38 A, for example using the same preferred oxides as outlined for ILD layer 38 A, followed by planarization and conventional metal damascene formation processes to form wiring interconnects to the S/D regions and gate electrodes. For example, tungsten filled damascenes e.g., 40 A, 40 B, 40 C, 40 D, 40 E, and 40 F are formed including barrier layers e.g., Ti, TiN, Ta, TaN, or WN (not shown) lining the damascene openings.

It will be appreciated that the above process steps may be carried out with respect to the opposite polarity (conductivity type) devices, for example where recessed areas are formed on either side of the NMOS channel region 26 A, and a strained semiconducting silicon alloy with a contracted lattice parameter respect to silicon, e.g., SiC, is used to backfill the recessed area to form a tensile strain on the NMOS device channel region. It will additionally be appreciated that both NMOS and PMOS devices may be formed with respective channel regions in respective tensile and compressive strain to respectively improve electron and hole mobility.

Referring to FIG. 2 is a process flow diagram including several embodiments of the present invention. In process 201 , PMOS and NMOS gate structure are provided on a semiconductor substrate including offset liners. In process 203 , recessed areas are formed adjacent one of the PMOS and NMOS (e.g., PMOS) gate structures. In process 205 , portions of the recessed areas are backfilled with a strained semiconducting silicon alloy, followed by an overlying silicon layer. In process 207 , ion implant processes are carried out before and after formation of offset spacers. In process 209 , first metal silicide regions are formed over S/D regions. In process 211 , a contact etch stop layer is deposited. In process 213 , a first ILD layer is deposited and planarized to expose the polysilicon gate electrodes. In process 215 , the polysilicon gate electrodes are thinned to allow full metal silicidation. In process 217 , the polysilicon gate electrodes are doped to adjust a work function. In process 219 , fully metal silicided gate electrodes are formed. In process 221 , a second ILD layer is formed and metal damascenes formed to contact S/D regions and gate electrodes.

Thus a PMOS and NMOS device pair and method for forming the same has been presented for applying a selected type of stress to the channel region of an FET device by backfilling recessed areas formed in source/drain regions with a strained semiconducting material to improve charge mobility and thereby improving device performance including drive current (I dsat ) and device speed. In addition, by forming fully silicided gate electrodes with a reduced thickness, electrical resistance of the gate electrode is reduced, and the work function is more easily adjusted as well to avoid a depletion effect thereby additionally improving device performance including drive current (I dsat ) and device speed.

The preferred embodiments, aspects, and features of the invention having been described, it will be apparent to those skilled in the art that numerous variations, modifications, and substitutions may be made without departing from the spirit of the invention as disclosed and further claimed below.