Shielded Gate Trench FET with the Shield and Gate Electrodes Connected Together in Non-active Region

Abstract

A field effect transistor (FET) in a semiconductor die including an active region housing active cells, a non-active region with no active cells therein, a drift region of a first conductivity type, a body region of a second conductivity type over the drift region, and a plurality of trenches extending through the body region and into the drift region. Each trench includes a shield electrode and a gate electrode, the shield electrode being disposed below the gate electrode. The FET further includes source regions of the first conductivity type in the body region adjacent to each trench, heavy body regions of the second conductivity type in the body regions adjacent the source regions, and a source interconnect layer contacting the source regions and heavy body regions. The shield electrode and the gate electrode extend out of each trench and into the non-active region where the shield electrode and gate electrode are electrically connected together by a gate interconnect layer.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. application Ser. No. 12/268,616, filed Nov. 11, 2008, which is a continuation of U.S. application Ser. No. 11/938,583, filed Nov. 12, 2007, now U.S. Pat. No. 7,473,603, which is a division of U.S. application Ser. No. 11/471,279, filed Jun. 19, 2006, now U.S. Pat. No. 7,319,256, all of which are incorporated herein by reference in their entirety for all purposes.

BACKGROUND

The present invention relates in general to semiconductor power field effect transistors (FETs) and in particular to shielded gate trench FETs with their shield and gate electrodes connected together.

Shielded gate trench FETs are advantageous over conventional FETs 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 sectional view of a conventional shielded gate trench MOSFET. An n-type epitaxial layer 102 extends over n+ substrate 100. N+ source regions 108 and p+ heavy body regions 106 are formed in a p-type body region 104 which is in turn formed in epitaxial layer 102. Trench 110 extends through body region 104 and terminates in the drift region. Trench 110 includes a shield electrode 114 below a gate electrode 122. Gate electrode 122 is insulated from its adjacent silicon regions by gate dielectric 120, and Shield electrode 114 is insulated from its adjacent silicon regions by a shield dielectric 112 which is thicker than gate dielectric 120.

The gate and shield electrodes are insulated from one another by a dielectric layer 116 also referred to as inter-electrode dielectric or IED. IED layer 116 must be of sufficient quality and thickness to support the potential difference that may exist between shield electrode 114 and gate electrode 122. In addition, interface trap charges and dielectric trap charges in IED layer 116 or at the interface between the shield electrode 114 and IED layer 116 are associated primarily with the methods for forming the IED layer.

The IED is typically formed by various processing methods. However, insuring a high-quality IED that is sufficiently robust and reliable enough to provide the required electrical characteristics results in complicated processes for forming the shielded gate trench FET. Accordingly, there is a need for structure and method of forming shielded gate trench FET that eliminate the need for a high-quality IED while maintaining or improving such electrical characteristics as on-resistance.

BRIEF SUMMARY

According to an embodiment of the invention, a field effect transistor (FET) includes a plurality of trenches extending into a semiconductor region. Each trench includes a gate electrode and a shield electrode with an inter-electrode dielectric therebetween. A body region extends between each pair of adjacent trenches, and source regions extend in each body region adjacent to the trenches. A first interconnect layer contacts the source and body regions. The plurality of trenches extends in an active region of the FET, and the shield electrode and gate electrode extend out of each trench and into a non-active region of the FET where the shield electrodes and gate electrodes are electrically connected together by a second interconnect layer. The electrical connection between the shield and gate electrodes is made through periodic contact openings formed in a gate runner region of the non-active region.

In one embodiment, the FET of claim further includes a substrate and a drift region bounded by the body regions and the substrate, and the plurality of trenches terminate within the drift region. In another embodiment, the plurality of trenches extends through the drift region and terminates within the substrate.

In yet another embodiment, the shield electrode and the gate electrode in each trench are electrically connected together by an additional connection through the IED in each trench.

In still another embodiment, the first interconnect layer is a source interconnect layer and the second interconnect layer is a gate interconnect layer.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a cross sectional view of a conventional shielded gate trench MOSFET;

FIGS. 2A-2H are simplified cross sectional views at various steps of a process for forming a shielded gate trench FET according to an embodiment of the invention; and

FIG. 3 is an isometric view of a portion of a gate runner in a shielded gate trench FET, according to an embodiment of the invention.

DETAILED DESCRIPTION

FIGS. 2A-2H are simplified cross sectional views at various steps of a process for forming a shielded gate trench FET according to an embodiment of the invention. In FIGS. 2A-2H, the left cross section views depict the sequence of steps leading to formation of the shield gate trench FET structure in the active region, and the right cross section views depict corresponding views of a transition region from active region to non-active region (from right to left). In this disclosure, “active region” represents areas of a die housing the active cells, and “non-active region” represents areas of the die which do not include any active cells. The non-active region includes the termination region extending along the perimeter of the die and the gate runners extending along the perimeter or middle of the die or along both the perimeter and middle of the die.

In FIG. 2A, using conventional techniques, trench 210 is formed in a semiconductor region 202, and then a shield dielectric 212 (e.g., comprising oxide) is formed lining the trench sidewalls and bottom surface and extending over mesa regions adjacent the trench. The right cross section view in each of FIGS. 2A-2H is through the center of the trench in the left cross section view, along a dimension perpendicular to the left cross section view. Thus, the right cross section view shows the trench of the left cross section view terminating at the edge of the active region. Also, the cross section views are not to scale, and in particular, the physical dimensions (e.g., thickness) of the same layers or regions in the right and the left cross section views may not appear the same. For example, in FIG. 2A, shield dielectric 212 appears thinner in the right cross section view than the left.

As shown in the right cross section view of FIG. 2A, shield dielectric 212 extends along the bottom surface of trench 210, and at the edge of the active region, extends up and out of trench 210 and over silicon region 202. In one embodiment semiconductor region 202 includes an n-type epitaxial layer (not shown) formed over a highly doped n-type substrate (not shown), and trench 202 extends into and terminates within epitaxial layer. In another variation, trench 202 extends through the epitaxial layer and terminates within the substrate.

In FIG. 2B, shield electrode 214 is formed along a bottom portion of trench 210 and is made electrically accessible in the non-active region of the die, as follows. Using known techniques, a conductive material (e.g., comprising doped or undoped polysilicon) is first formed filling the trench and extending over the mesa regions, and subsequently recessed deep into trench 210 to form shield electrode 214.

During recessing of the conductive material, a mask 211 is used to protect portions of the conductive material extending in the non-active region of the die. As a result, shield electrode 214 is thicker inside trench 210 than over the mesa surfaces in the non-active region of the die, as depicted in the right cross section view in FIG. 2B. Further mask 211 is applied such that, at the edge of the active region, the shield electrode extends out of trench 210 and over the mesa surface of the non-active region. Shield electrode 214 inside trench 210 is thus made available for electrical connectivity in the non-active region of the die.

In FIG. 2C, using known methods, shield dielectric 212 is completely removed from along trench sidewalls and over mesa surfaces in the active region, as depicted by the right cross section view. The shield dielectric is thus recessed below the top surface of shield electrode 214. In one embodiment, shield electrode 214 is recessed so that its top surface becomes co-planar with that of the shield dielectric layer 212. This provides a planar surface for the subsequent formation of gate/inter-electrode dielectric layer.

In FIG. 2D, a gate dielectric layer 216 extending along upper trench sidewalls is formed using conventional techniques. In one embodiment, gate dielectric 216 is formed using conventional oxidation of silicon. This process also results in oxidation of shield electrode 214 thus forming an inter-electrode dielectric (IED) layer over gate electrode 214. As shown in the right cross section view, dielectric layer 216 extends along all exposed surfaces of the shield electrode 214 in the active and non-active regions. As further discussed below, the additional process steps typically required for forming a high-quality IED are eliminated.

In FIG. 2E, recessed gate electrode 222 is formed in trench 210 and is made electrically accessible in the non-active region as follows. Using conventional techniques, a second conductive layer (e.g., comprising doped polysilicon) is formed filling trench 210 and extending over the mesa surfaces in the active and non-active regions of the die. The second conductive layer is then recessed into trench 210 to form gate electrode 222.

During recessing of the second conductive layer, a mask 219 is used to protect portions of the second conductive material extending in the non-active region of the die. As a result, gate electrode 222 is thicker inside trench 210 than over the mesa surfaces in the non-active region of the die, as depicted in the right cross section view in FIG. 2B. Further mask 219 is applied such that, at the edge of the active region, the recessed gate electrode 222 extends out of trench 210 and over the mesa surface of the non-active region. Gate electrode 222 inside trench 210 is thus made available for electrical connectivity in the non-active region of the die. Note that mask 219 does not extend over the entire shield electrode 214 in the non-active region. As will be seen, this facilitates contacting both the gate electrode and shield electrode through the same contact opening.

In FIG. 2E, p-type body regions 204 are formed in semiconductor region 202 using conventional body implant and drive in techniques. Highly doped n-type source regions 208 are then formed in body regions 216 adjacent trench 210 using conventional source implant techniques.

In FIG. 2F, a dielectric layer 224, such as BPSG, is formed over the structure using known techniques. In FIG. 2G, dielectric layer 224 is patterned and etched to form source/body contact openings in the active region, followed by a dielectric flow. As shown in the left cross section, a dielectric dome 225 extending fully over gate electrode 222 and partially over source regions 208 is formed. P-type heavy body regions 206 are then formed in exposed semiconductor regions 202 using conventional implant techniques. The same masking/etching process for forming contact openings in the active region is used to form a contact opening 221 in dielectric layer 224 in the non-active region in order to expose a surface region and sidewall of gate electrode 222 and a surface region of shield electrode 214, as shown in the right cross section view.

In FIG. 2H, an interconnect layer (e.g., comprising metal) is formed over the structure and then patterned to form source/body interconnect 226A and gate interconnect 226B. As shown in the left cross section view, source/body interconnect 226A contacts source regions 208 and heavy body regions 106 but is insulated from gate electrode 222 by dielectric dome 224. As shown in the right cross section view, gate metal 226B contacts both shield electrode 214 and gate electrode 222 through contact opening 221, thus shorting the two electrodes to one another.

Thus, contrary to conventional shielded gate FETs wherein the shield electrode either floats (i.e., is electrically unbiased) or is biased to the source potential (e.g., ground potential), in the FET embodiment shown in FIG. 2H, the shield electrode is connected and biased to the same potential as the gate electrode. In conventional FETs where the shield electrode is floating or connected to ground potential, a high-quality IED is typically required to support the potential difference between the shield and gate electrodes. However, electrically connecting together the shield and gate electrodes eliminates the need for a high-quality IED. The shield electrode, although biased to the gate potential, still serves as a charge balance structure enabling the reduction of the on resistance for the same breakdown voltage. Thus, a low on-resistance for the same breakdown voltage is obtained while the process steps associated with forming a high quality IED are eliminated. Theoretically, such a structure would not even need an IED, but the IED is formed naturally during the formation of gate dielectric. Thus, a high performance transistor is formed using a simple manufacturing process.

The electrical contact between the gate and shield electrodes may be formed in any non-active region, such as in the termination or edge regions of the die, or in the middle of the die where the gate runners extend as shown in FIG. 3. FIG. 3 is an isometric view of a portion of a gate runner in a shielded gate trench FET, according to an embodiment of the invention. The upper layers (e.g., gate interconnect layer 326B and dielectric layer 324) are peeled back in order to reveal the underlying structures. As shown, trenches 310 extending in parallel in the active region 341 terminate on either side of the gate runner region 340.

The gate runner region 340 is structurally symmetrical about line 3-3, with each half being structurally similar to that shown in FIG. 2H. Shield dielectric 312 extends out of the rows of trenches 310 and onto the mesa surface in gate runner region 340. Likewise, each of shield electrode 314, inter-electrode dielectric 316 and gate electrode 322 extend out of the rows of trenches 310 and onto the mesa surface in gate runner region 340. Regions 311 represent the mesas between adjacent trenches in the active region 341.

Contact openings 321 expose surface areas of shield electrode 314 to which gate interconnect layer 326B (e.g., comprising metal) makes electrical contact. Additionally, gate interconnect layer 326B makes electrical contact with surface areas 332 of gate electrodes 322 exposed through dielectric layer 324. It is desirable to minimize the gate resistance in order to minimize the delay in biasing the individual gate electrodes inside the trenches. For the same reasons, it is desirable to minimize the delay in biasing the individual shield electrodes inside the trenches. Accordingly, the frequency and shape of contact openings 321 in gate runner region 340 can be optimized to minimize the resistance and thus the delay from the gate pad to each of the gate and shield electrodes. The delay in biasing the shield and gate electrodes can be further reduced by forming the gate electrode to shield electrode contacts in both the gate runner regions and in the termination or edge regions of the die.

The shield and gate electrodes may be electrically connected in other ways according to other embodiments of the invention. For example, the IED in each trench may be etched in certain places before forming the gate electrode over the IED. In this embodiment, contact openings as shown in FIGS. 2H and 3 would not be necessary, and a gate interconnect contact to the gate electrode in each trench would also be coupled to the corresponding shield electrode through shorts in the IED. According to the other embodiments, gate and shield electrode contacts may be formed through openings in the IED and through contact openings formed in the non-active regions such as the termination and gate runner regions. The elimination of the need to form a high-quality IED results in a simplified and more controllable process for forming shielded gate trench MOSFETs with improved drain-to-source on-resistance R_DSon.

The principles of the invention may be applied to any shielded gate FET structures such as those shown in FIGS. 3A, 3B, 4A, 4C, 6-8, 9A-9C, 11, 12, 15, 16, 24 and 26A-26C of patent application Ser. No. 11/026,276, titled “Power Semiconductor Devices and Methods of Manufacture,” which disclosure is incorporated herein by reference in its entirety for all purposes.

While the above provides a complete description of the preferred embodiments of the invention, many alternatives, modifications, and equivalents are possible. Those skilled in the art will appreciate that the same techniques can apply to other types of super junction structures as well as more broadly to other kinds of devices including lateral devices. For example, while embodiments of the invention are described in the context of n-channel MOSFETs, the principles of the invention may be applied to p-channel MOSFETs by merely reversing the conductivity type of the various regions. Therefore, the above description should not be taken as limiting the scope of the invention, which is defined by the appended claims.

Claims

1-5. (canceled)

6. A field effect transistor (FET) in a semiconductor die comprising:

an active region housing active cells;

a non-active region with no active cells therein;

a drift region of a first conductivity type;

a body region of a second conductivity type over the drift region;

a plurality of trenches extending through the body region and into the drift region, each trench including a shield electrode and a gate electrode, the shield electrode being disposed below the gate electrode;

source regions of the first conductivity type in the body region adjacent to each trench;

heavy body regions of the second conductivity type in the body regions adjacent the source regions; and

a source interconnect layer contacting the source regions and heavy body regions;

wherein the shield electrode and the gate electrode extend out of each trench and into the non-active region where the shield electrode and gate electrode are electrically connected together by a gate interconnect layer.

7. The FET of claim 6 wherein the gate electrode is recessed in the trench to below a top surface of the source regions, the FET further comprising a dielectric material over the gate electrode for insulating the gate electrode and the source interconnect layer from one another.

8. The FET of claim 6 further comprising a substrate of the first conductivity type, the drift region extending over the substrate, wherein each trench terminates within the drift region.

9. The FET of claim 6 further comprising a substrate of the first conductivity type, the drift region extending over the substrate, wherein each trench extends through the drift region and terminates within the substrate.

10. The FET of claim 6 further comprising an inter-electrode dielectric between the shield electrode and the gate electrode in each trench, wherein the shield electrode is electrically connected to the gate electrode by an additional connection through the inter-electrode dielectric.

11. The FET of claim 6 wherein the non-active region includes a gate runner region extending through a middle portion of the die, the shield electrode and gate electrode extending out of each trench and into the gate runner region where the shield electrode and gate electrode are electrically connected together by the gate interconnect layer.

12. The FET of claim 6 wherein the non-active region includes a termination region extending along a perimeter of the die, the shield electrode and gate electrode extending out of each trench and into the termination region where the shield electrode and gate electrode are electrically connected together by the gate interconnect layer.