Method of making transistor with strained source/drain

[0001] This invention relates to the field of semiconductor manufacturing, and more specifically to a method of manufacturing transistors and devices.

[0002] Silicon Germanium (SiGe) alloy has been used in non-recessed source/drain regions (such as raised source/drain) for a shallow junction to suppress the short channel effect. The short channel effect is a well-known phenomenon in which the threshold voltage becomes less predictable as the gate dimensions are reduced.

[0003] U.S. patent application No. 2003/0080361 describes a process for manufacturing an improved PMOS transistor. Recesses are etched into a layer of epitaxial silicon after formation of spacers adjacent the gate electrode, and implantation of dopant. Source and drain films are deposited in the recesses. The source and drain films are made of an alloy of silicon and germanium. The alloy is epitaxially deposited on the layer of silicon after spacer formation. The alloy has a lattice having the same structure as the structure of the lattice of the layer of silicon. However, due to the inclusion of the germanium, the lattice of the alloy has a larger spacing than the spacing of the lattice of the layer of silicon. The larger spacing creates a stress in a channel of the transistor between the source and drain films. Silicon under a biaxially stressed film, such as SiGe, enhances carrier mobility to improve current performance.

[0004] However, the short channel effect was increased as a result of the higher temperature conditions of epitaxially forming the silicon/geranium alloy in the source drain regions. An improved fabrication method that can improve device performance and the processing window is desired.

[0005] In some embodiments, a method of fabricating a transistor comprises the steps of: forming a gate electrode above a substrate made of a first semiconductor material having a first lattice spacing, forming recesses in the semiconductor substrate at respective locations where a source region and a drain region are to be formed, epitaxially growing a second semiconductor material having a second lattice spacing different from the first lattice spacing in the recesses, and implanting a dopant in the second semiconductor material to form a pocket or lightly doped drain implant after the growing step.

[0006] In some embodiments, a method of fabricating a CMOS device including a PMOS transistor and an NMOS transistor, comprises the steps of: forming gate electrodes for the PMOS transistor and for the NMOS transistor above a substrate made of a first semiconductor material having a first lattice spacing, forming recesses in the semiconductor substrate at respective locations adjacent the gate electrode of the PMOS transistor, where a source region and a drain region are to be formed, epitaxially growing a second semiconductor material having a second lattice spacing different from the first lattice spacing in the recesses, and implanting dopants in the PMOS and NMOS transistors to form a pocket or lightly doped drain implant after the growing step.

[0019] The following detailed description of a preferred embodiment of the invention is to be considered together with the accompanying drawings wherein like numbers refer to like parts and further wherein the drawings are to be considered part of the entire written description. Terms used to describe the preferred structure and process embodiments have traditional meaning in the art. Relative terms such as “horizontal”, vertical, “up”, “down”, “top”, “bottom” should be construed to refer to the orientation as described or as shown in the drawing figure under discussion. The drawing figures are not to scale and certain features may be shown exaggerated in scale or in somewhat schematic form in the interest of clarity and conciseness.

[0020] U.S. patent application Ser. No. 10/002,465 filed Nov. 1, 2001 (U.S. Patent Application Publication No. U.S. 2003/0080361 A1, May 1, 2003) is incorporated by reference herein in its entirety, as though fully set forth below.

[0021] FIGS. 1–12 show an exemplary method of forming a PMOS transistor having strained source/drain regions, and a CMOS device including the PMOS transistor. In the examples, the Pocket and/or lightly doped drain (LDD) implants are processed after selective epitaxial SiGe deposition in the PMOS source/drain region recesses.

[0022] Referring to FIG. 1, two gate electrodes 10 separated by a shallow trench isolation structure (STI) 15 are formed on the surface of a semiconductor substrate 20 made of a material having a first lattice spacing. In some embodiments, the substrate 20 comprises an epitaxial silicon layer formed on a monocrystalline silicon wafer substrate. The gate electrodes 10 may be formed of polycrystalline silicon, for example. A hard mask layer 25 of a material such as silicon oxynitride or SiN may be formed over the two gate electrodes 10. Subsequently, an oxide layer 30 (e.g., SiO, TEOS, or RTO oxide) and a silicon nitride 35 masking layer are applied over the surface of the substrate 20.

[0023] In some embodiments, another oxide layer (not shown) is deposited above the SiN layer 35, for forming ONO spacers.

[0024] Referring to FIG. 2, a highly selective silicon nitride etch back process is employed to produce a first set of silicon nitride spacers 40 and silicon oxide spacers 45 adjacent the gate electrodes 10.

[0025] If the optional second oxide layer (not shown) was provided above the SiN layer 35, the second oxide layer is removed by the SiN etch back process of this step, except for a small residual layer above the spacers 40. Then an extra step is performed to remove the remainder of the second oxide layer, for example, by an HF dip.

[0026] Referring to FIG. 3, a P+ implant photomask 50 comprising a photoresist is next applied over the NMOS transistor area. The exposed region defines the area available for subsequent recess etching. Thus, the recess etching area is only defined for the PMOS transistor.

[0027] Referring to FIG. 4, recessed areas 60 for the PMOS transistor are then formed in the semiconductor substrate 20, for example, by using a known etching method to selectively etch the oxide and substrate, but not the spacers 40.

[0028] Referring to FIG. 5, the photoresist 50 is then stripped, to expose the NMOS transistor region. Unlike the PMOS transistor region, there are no recesses in the source and drain regions of the NMOS transistor.

[0029] Referring to FIG. 6, source and drain regions 65 are formed by epitaxial deposition of SiGe alloy in the recesses 60, for example by known CVD methods at high temperature, e.g. 650˜850° C. The lattice spacing of the SiGe alloy is different from the first lattice spacing of the semiconductor material of substrate 20. The germanium present in the combination of the silicon and the germanium may be about 15 atomic percent. It is known that epitaxy can be maintained with a germanium concentration of up to about 20 atomic percent of the combination of the silicon and germanium by volume. It is also understood that other materials can optionally be incorporated into the SiGe alloy.

[0030] Referring to FIG. 7, the silicon nitride spacers 40 and gate hard mask layer 25 are subsequently removed, for example by a phosphoric acid (H₃PO₄) etch back process. An H₃PO₄ process may be preferred, because it reduces the critical dimension (CD) bias of the PFET polysilicon after the hard mask is removed, and provides surface roughness control.

[0031] Referring to FIG. 8, the resultant PMOS transistor source/drain extension (SDE) region, including gate electrode 10A, silicon oxide spacers 45 and source/drain regions 65, is completed by P-pocket/P-LDD implant 70. This may be done, for example, using standard methods. The NMOS transistor is completed by an N-pocket/N-LDD implant 75. In some embodiments, as shown in FIG. 8, the oxide spacers 45 are left in place while the pocket/LDD implants are performed.

[0032] Referring to FIG. 9, in some alternative embodiments, the silicon oxide spacers 45 are removed before the pocket/LDD implants are performed.

[0033] Referring to FIG. 10, silicon oxide 80 and silicon nitride 85 spacer layers are applied over the PMOS and NMOS transistors of FIG. 8 or FIG. 9. The silicon oxide layer 80 is typically applied via known LPCVD or PECVD procedures using TEOS as a source. The silicon nitride layer 85 is typically applied using well known LPCVD or PECVD procedures. In some embodiments, the oxide layer 80 may include oxide from spacers 45 that is not removed, as shown in FIG. 8.

[0034] Referring to FIG. 11, a second set of silicon nitride spacers 90 and silicon oxide spacers 95 are then formed adjacent each gate electrode 10A–B using a highly selective silicon nitride etch back process.

[0035] Referring to FIG. 12, a PMOS transistor 100 is formed by selectively performing a P+ dopant implant in the area defined by gate electrode 10A and source and drain regions 65 of FIG. 8. The NMOS device 110 can be formed by selectively performing an N+ dopant implant in the area defined by gate electrode 10B and adjacent surfaces of the semiconductor material. This completes the CMOS structure 105. Although an example of a CMOS structure 105 is provided, the method described above may be used to fabricate any PMOS transistor, for example, to be included in logic, input/output (I/O), static random access memory (SRAM), or the like.

[0036] The process described in FIGS. 1–12, wherein LDD/Pocket implantation occurs after the formation of the Epitaxial SiGe alloy source and drain regions, provides for better control of implantation in a transistor with strained source/drain, especially the LDD/Pocket implants, in the final structure. This control improves the processing window and provides more flexible tuning of a transistor and enhanced performance of a CMOS device. Because the implants are performed after the epitaxial growth of the SiGe material in the source/drain region recesses, there is no impact on the thermal budget of the SiGe growth. The examples described above are compatible with existing processes using either SiN spacers or composite spacers.

[0037] Although the invention has been described in terms of exemplary embodiments, it is not limited thereto. Rather, the invention should be construed broadly, to include other variants and embodiments, which may be made by those skilled in the art without departing from the scope and range of equivalents of the appended claims.