CROSS-REFERENCES TO RELATED APPLICATIONS

This application is a division of U.S. application Ser. No. 11/201,400, filed Aug. 9, 2005, which is incorporated by reference in its entirety for all purposes.

BACKGROUND OF THE INVENTION

The invention relates in general to semiconductor power field effect transistors (FETs), and more particularly to a structure and method for forming an improved inter-poly dielectric (IPD) in a shielded gate FET.

Shielded gate trench FETs are advantageous in that the shield electrode reduces the gate-drain capacitance (Cgd) and improves the breakdown voltage of the transistor. FIG. 1 is a simplified cross section view of a conventional shielded gate trench MOSFET. Trench 110 includes a shield electrode 114 directly below a gate electrode 122. Shield electrode 114 is insulated from adjacent silicon regions by a shield dielectric 112 which is generally thicker than gate dielectric 120. The gate and shield electrodes are insulated from one another by a dielectric layer 116 commonly referred to as inter-poly dielectric or IPD. The IPD layer must be of sufficient quality and thickness to support the required voltage between the gate and shield electrodes.

The conventional shielded gate FET of FIG. 1 suffers from a number of drawbacks. First, gate electrode 122 has sharp bottom comers which together with the flat top surface of shield electrode 114 leads to high electric fields in these regions. Second, conventional methods for forming the IPD typically introduce an oxide layer on the mesas between trenches that must be removed at some point after the gate electrode has been formed. In removing this oxide some etching of the gate oxide down the trench sidewalls inevitably occurs, which may result in gate shorts and gate leakage. Other known techniques tie formation of the IPD to formation of the gate dielectric and thus the IPD thickness is limited to a set multiple of the gate dielectric thickness. This does not allow independent optimization of the gate dielectric and the IPD. The largest differential in thickness between the IPD and the gate dielectric achieved has been about three to one (i.e., for a given target gate dielectric thickness, the largest IPD thickness that has been achieved is about three times greater than that of the target gate dielectric thickness).

Thus, there is a need for a structure and method of forming a shielded gate trench FET with improved IPD and gate dielectric.

BRIEF SUMMARY OF THE INVENTION

In accordance with an embodiment of the invention, a method of forming a FET includes the following steps. A trench is formed in a silicon region of a first conductivity type. A shield electrode is formed in a lower portion of the trench. An inter-poly dielectric (IPD) including a layer of thermal oxide and a layer of conformal dielectric is formed along an upper surface of the shield electrode. A gate dielectric lining at least upper trench sidewalls is formed. A gate electrode is formed in the trench. The gate electrode is insulated from the shield electrode by the IPD.

In one embodiment, the IPD has a concave upper surface.

In another embodiment, a ratio of a thickness of the IPD to a thickness of the gate dielectric is greater than three to one.

In yet another embodiment, the gate dielectric is formed after forming the IPD.

In another embodiment, the IPD is formed as follows. A layer of thermal oxide is formed along upper sidewalls of the trench and along an upper surface of the shield electrode. The trench is filled with a conformal layer of dielectric. The conformal layer of dielectric and the layer of thermal oxide are partially removed such that the IPD comprising a remaining portion of the thermal dielectric layer and a remaining portion of the conformal layer of dielectric is formed.

In accordance with another embodiment of the invention, a FET includes a trench extending into a silicon region of a first conductivity type. A shield electrode insulated from the silicon region by a shield dielectric extends in a lower portion of the trench. A gate electrode is in the trench over, but insulated from, the shield electrode by an inter-poly dielectric (IPD). The IPD comprises a conformal layer of dielectric and a thermal oxide layer.

In one embodiment, the gate electrode is insulated from the silicon region by a gate dielectric extending along upper trench sidewalls, and a ratio of a thickness of the IPD to a thickness of the gate dielectric is greater than three to one.

In another embodiment, the IPD has a concave profile along its upper surface.

In yet another embodiment, the gate electrode has a concave profile along its lower surface.

In another embodiment, the conformal layer of dielectric is surrounded by the thermal dielectric along its lower surface and sidewalls.

In another embodiment, the conformal layer of dielectric has a convex lower surface and a concave upper surface.

In another embodiment, the shield electrode has a rounded top surface.

In another embodiment, the silicon region includes an epitaxial layer of the first conductivity type extending over a substrate of the first conductivity type, well regions of a second conductivity type in the epitaxial layer, and source regions of the first conductivity in the well regions. The source regions flank the trench.

In one embodiment, the trench extends into and terminates within the epitaxial layer.

In another embodiment, the trench extends through the epitaxial layer and terminates within the substrate.

The following detailed description and the accompanying drawings provide a better understanding of the nature and advantages of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a simplified cross section view of a conventional shielded gate trench MOSFET; and

FIGS. 2A-2L are simplified cross section views at various stages of an exemplary process sequence for forming a shielded gate trench FET with improved inter-poly dielectric and gate dielectric, in accordance with an embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

In accordance with the present invention, an IPD layer of a shielded gate FET is formed by performing a thermal oxidation followed by deposition of a conformal layer of dielectric. A gate dielectric is then formed, followed by formation of the gate electrode. This method decouples the process for forming the IPD layer from that for forming the gate dielectric, enabling each of these dielectric layers to be independently optimized. A high quality, thick IPD can thus be formed to support the required voltage between the gate and shield electrodes without compromising the gate dielectric quality or thickness. Further, the conformal layer of dielectric helps obtain a smooth concave profile (i.e., like the inside of a bowl) along the top surface of the IPD which in turn results in a smooth bottom profile for the gate electrode. This helps reduce the electric field as compared to the conventional FET structure in FIG. 1 where the gate electrode has sharp lower corners.

FIGS. 2A-2L are simplified cross section views at various stages of an exemplary process sequence for forming a shielded gate trench FET with improved IPD and gate dielectric, in accordance with an embodiment of the invention. In FIG. 2A, a trench 202 is formed in silicon region 204 using conventional masking and silicon etch techniques. In one embodiment, silicon region 204 comprises a relatively lightly doped n-type epitaxial layer extending over a highly conductive n-type substrate. In this embodiment, depending on the design goals, the trench can be etched to terminate within the epitaxial layer or extend deeper to terminate within the substrate. In FIG. 2B, a shield dielectric layer 206 (e.g., comprising oxide) lining the trench sidewalls and bottom as well as the surface of the mesa regions adjacent the trench is formed using known techniques. In one embodiment, the shield oxide is formed using high temperature (e.g., 1,150Â° C.) dry oxidation. The high temperature helps round out the trench bottom corners and form about 1,250 â„« of shield oxide sufficient to support device breakdown.

In FIG. 2C, a layer of polysilicon is deposited to fill the trench using conventional polysilicon deposition techniques. The deposited polysilicon is then recessed into the trench to thereby form shield electrode 208. The exposed portions of the shield dielectric layer are then thinned down, to for example 650 â„«. In an embodiment where the shield electrode comprises oxide, a wet buffered oxide etch is used to thin down the shield oxide. The remaining exposed portion 206a of the shield dielectric helps prevent excessive growth of oxide along the trench sidewalls and mesa surfaces in the subsequent thermal oxidation step, and helps control the profile of the trenched cavity. Excessive growth of thermal oxide can lead to formation of voids in the subsequent deposited conformal dielectric. In one embodiment, the exposed portions of shield dielectric 206 are completely removed or alternatively the entire shield dielectric 206 is left intact.

In FIG. 2D, a thermal oxidation step is carried out to form a thermal oxide layer 210 along the trench sidewalls, over the surface of the mesa regions adjacent the trench, and over shield electrode 208. The thermal oxidation advantageously oxidizes an upper portion of shield electrode 208 resulting in a rounded-top profile. The rounded top helps minimize electric field in the regions between shield electrode 208 and the gate electrode formed later. In one embodiment, thermal oxide layer 210 is formed by performing a low temperature, wet oxidation followed by a high temperature, dry oxidation. In another embodiment, the thermal oxidation yields a thermal oxide layer 210 with a thickness in the range of 1000-1500 â„«. In yet another embodiment, a low temperature thermal oxidation (e.g., about 850Â° C.) is carried out so that a thicker thermal oxide layer is formed along the top surface of the shield electrode than along the trench sidewalls and over the mesa surfaces. In such embodiment, a thickness ratio in the range of 1.5:1 to 2:1 would be desirable. In one specific embodiment, the thermal oxidation yields a thermal oxide layer having a thickness of about 1,850 â„« along the top surface of the shield electrode and a thickness of about 1,250 â„« along the trench sidewalls and the mesa surfaces.

In FIG. 2E, a conformal layer of dielectric 212 (e.g., comprising oxide) is deposited to fill the trench. In one embodiment, conformal dielectric layer 212 is deposited using a sub-atmospheric chemical vapor deposition (SACVD) tetraethylorthosilicate (TEOS)/Ozone process at a temperature of about 510Â° C. and pressure of about 480 Torr. The deposited oxide fills the trench completely without voiding.

In FIG. 2F, conformal dielectric layer 212 and thermal oxide layer 210 are etched down into the trench to the desired depth. All the dielectric over the mesa surfaces and along upper trench sidewalls is completely removed, and an inter-poly dielectric (IPD) layer 214 having a concave top surface remains over shield electrode 208. IPD layer 214 thus comprises a stack of thermal dielectric layer and conformal dielectric layer. In one embodiment, a uniform etch back of the dielectric stack is carried out in recessing the dielectric stack into the trench. A dry anisotropic plasma etch or a wet etch can be carried out to achieve the desired thickness for the IPD and to ensure that the oxide along the trench sidewalls and over the mesa is completely removed. A conventional densification step can also be carried out to densify the SACVD oxide. In one embodiment, a dry etch and a subsequent densification are carried out, followed by a wet etch.

In FIG. 2G, a gate dielectric layer 216 (e.g., from oxide) extending along trench sidewalls, over the IPD layer, and over the mesa regions adjacent the trench is formed using conventional techniques. Because the IPD formation is completely decoupled from the gate dielectric formation, the gate dielectric can be independently optimized to have the desired characteristics. In FIG. 2H, using conventional techniques, a layer of polysilicon is deposited to fill the trench, and is then etched back to form a recessed gate electrode 218 in trench 202.

In FIG. 2I, the portion of gate dielectric 216 extending over the mesa is etched back to a thickness suitable for body implant and source implant. A conventional blanket body implant and drive in process is performed to form p-type body regions 220 along an upper portion of silicon region 204. A conventional source implant together with a masking layer (not shown) is then used to form source regions 222 flanking trench 202. In FIG. 2J, a pre-metal dielectric layer 224 (e.g., comprising BPSG) is formed over the structure using conventional techniques. In FIG. 2K, using a masking layer (not shown), dielectric layer 224 is partially removed to expose surfaces of body regions 220 and source regions 222 as defined by the masking layer. A conventional silicon etch (e.g., dry etch) is then carried out to recess the exposed surface regions. The recessed silicon regions thus form contact openings 226.

In FIG. 2L, a blanket heavy body implant is carried out to form self-aligned p-type heavy body regions 228 in body regions 220. A reflow of dielectric 224 is then carried out to obtain a better aspect ratio for the contact openings and a better step coverage for a metal layer 226 which is formed in a subsequent step to electrically contact heavy body regions 228 and source regions 222. A horizontally extending dashed line is included in FIG. 2L to illustrate the embodiment wherein an epitaxial layer 203 extends over a substrate 201, and trench 202 extends through epitaxial layer 203 and terminates in substrate 201. Alternatively, trench 202 may be terminated within epitaxial layer 203. Also, the process sequence depicted by cross section views in FIGS. 2A-2L is merely exemplary and the various steps can be modified or carried out in a different sequence than that shown. In particular, any one of a number of known process steps can be carried out instead of those depicted by FIGS. 2I-2L to obtain a shielded gate structure with the desired features and characteristics.

In accordance with the structure and method of the present invention, an improved IPD layer comprising a film stack of a thermally grown dielectric and a conformal deposited dielectric is obtained. The thermally grown dielectric provides a high quality dielectric and serves to round off the top comers of the shield electrode. The conformal deposited dielectric (1) fills in the gaps on the upper side of the shield electrode, (2) serves as a hard mask to protect the thermal dielectric over the shield electrode while the dielectric along the trench sidewalls is being etched, and (3) creates a smooth concave profile over which the gate electrode is formed. The rounded profile along the top surface of the shield electrode and the bottom of the gate electrode result in lower electric fields in these localized areas.

Further, the method of forming the IPD in accordance with the invention serves to decouple formation of the IPD from that of the gate oxide so that the IPD and gate dielectric can be independently optimized to, for example, achieve a thick, high quality IPD and a thin, high quality gate dielectric. In one embodiment, optimum transistor characteristics are obtained by using a ratio of IPD thickness to gate dielectric thickness of greater than about five to one. For example, an IPD thickness of greater than about 2,000 â„«, and a gate dielectric thickness of less than about 400 â„« is found to provide optimum transistor characteristics. The ability to form a thin gate dielectric can advantageously be used to achieve a lower on resistance.

The various structures and methods of the present invention may be combined with one or more of a number of charge spreading techniques as well as other shielded gate structures and manufacturing processes disclosed in the commonly assigned application Ser. No. 11/026,276, filed Dec. 29, 2004, and incorporated herein by reference in its entirety, to achieve an even lower on-resistance, higher blocking capability and higher efficiency, among other advantages and features.

Although a number of specific embodiments are shown and described above, embodiments of the invention are not limited thereto. For example, it is understood that the doping polarities of the structures shown and described could be reversed and/or the doping concentrations of the various elements could be altered without departing from the invention. The process sequence depicted by FIGS. 2A-2L is for forming an n-channel FET, however modifying this process sequence to form a p-channel FET would be obvious to one skilled in the art in view of this disclosure. Also, while the various embodiments described above are implemented in conventional silicon, these embodiments and their obvious variants can also be implemented in silicon carbide, gallium arsenide, gallium nitride, diamond or other semiconductor materials. Further, the cross-section views of the different embodiments may not be to scale, and as such are not intended to limit the possible variations in the layout design of the corresponding structures. Also, the FET shown and its obvious variants can be formed in stripe or cellular architecture including hexagonal or square shaped transistor cells. Moreover, the features of one or more embodiments of the invention may be combined with one or more features of other embodiments of the invention without departing from the scope of the invention. Hence, the scope of this invention should not be limited to the embodiments described, but are instead defined by the following claims.