Method for forming a superjunction device with improved ruggedness

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation and claims a benefit and priority to a U.S. patent application Ser. No. 11/586,901, filed on Oct. 26, 2006, entitled SUPERJUNCTION DEVICE WITH IMPROVED RUGGEDNESS, assigned to the same assignee, the U.S. patent application Ser. No. 11/586,901 is a divisional and claims a benefit and priority to U.S. patent application Ser. No. 10/968,499, filed Oct. 19, 2004, entitled SUPERJUNCTION DEVICE WITH IMPROVED RUGGEDNESS, which claims the benefit and priority to U.S. Provisional Application Ser. No. 60/513,174, filed Oct. 21, 2003, entitled SUPERJUNCTION DEVICE WITH IMPROVED RUGGEDNESS, all of which are incorporated herein by reference.

FIELD OF THE INVENTION

This invention relates to superjunction devices and a process for their manufacture, and more specifically to the increase of the ruggedness of superjunction devices, and the increase of the process window tolerances for such devices.

BACKGROUND OF THE INVENTION

Superjunction devices possess the advantage of significantly reduced for the same high breakdown voltage (BV) of a conventional MOSFET. The superjunction is comprised of a multi-layer, for example, a six-layer sequence of implant and epitaxy to form spaced P-columns which is used to balance the charge in the N type drift region epi which receives the columns. The same reticle is used repetitively on the six layers to generate the P-column.

The charge balance is critical with a small process window. Exceeding this window on the P-type side (that is, having an excessive P charge in the P columns) leads to the BV falling below the spec. Exceeding this window on the N-type side leads to high BV but can lead to ruggedness reduction.

Device ruggedness can be enhanced by structural modifications that force the current to flow through the P-column rather than outside it. Such structures are shown in copending application Ser. No. 60/417,212, filed Oct. 8, 2002 and assigned to the assignee of the present invention, and which is incorporated herein by reference. In that case, the top-most portion only of the P columns had a higher and unbalanced P concentration (charge) than the remainder of the columns, which have a balanced concentration against the surrounding N type body. This caused avalanche current at the top of the columns to be diverted from under the MOSFET source regions (the R_b′ region) and toward the axis of the column.

BRIEF DESCRIPTION OF THE INVENTION

The invention proposes a different modification. Instead of using the same design for all the layers, the topmost layer design is modified with a slightly larger feature (diameter) and thus increased volume and P charge, solely in the active area such that the BV of the active area cells is reduced selectively and also so that the current flows into or toward the axis of the P column, thus improving the ruggedness. The lower 5 layers and the termination can then be optimized for maximum BV. The use of the separate upper or 6^th layer design will allow the realization of high termination BV, relatively lower active area BV and current flow in the P-column. The conjunction of these 3 factors will improve the ruggedness and increase the process window tolerance since it reduces the dependence of the EAS on the device BV. Note that while a six layer design has been chosen to illustrate the invention, any number of layers can be used.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a cross-section of a small portion of the active area of a superjunction device, which employs the present invention.

FIG. 2 is a cross-section of FIG. 1 taken across section line 2-2 in FIG. 1.

FIG. 3 shows the process step of forming an enlarged volume P section at the top of the P columns in accordance with the invention.

DETAILED DESCRIPTION OF THE INVENTION

Referring to FIGS. 1 to 3, a silicon wafer (sometimes referred to as a die or chip) is formed of an N⁺ wafer 10 which receives a series of epitaxial layers N₆ to N₁ which are sequentially formed. After the formation of each layer, an implant and diffusion is carried out to form P regions (P₆ to P₁ respectively). In the prior art, the implant and diffusions are identical in size and concentration and which are charge balanced to the surrounding charge of layers N₆ to N₁ respectively, which are each of the same concentrations.

Each of P regions P₆ to P₁ are aligned to one another to form a continuous column or “pylon”.

A MOSgated structure is then formed atop each column, shown in FIG. 1 as P channel regions 20, 21 and 22 which conventionally receive N⁺ source regions 24, 25 and 26 respectively. A gate oxide 27 is deposited over the inversion areas of each of channel regions 20, 21 and 26 respectively and a conductive polysilicon gate 28 is formed over each of gate oxide regions 27. An LTO insulation layer 29 is formed over gates 28 and a source electrode 30 is formed over the layer 29 and contacts each of sources 24, 25, 26 and the inner channel of regions 20, 21 and 22. Note that the channel regions 20, 21 and 22 may be polygonal cells or stripes; and columns P₆ to P₁ have corresponding circular or stripe shapes. A drain electrode 40 is attached to the bottom of N⁺ region 10.

In accordance with the invention, the uppermost P regions P₁ have a greater diameter then that of the underlying regions P₂ to P₆, so that the top of the columns will have a greater P charge than that of the surrounding N₁ layer. The top-most column may have an increased diameter of only a few percent over that of the lower columns. By way of example, if the lower column elements P₂ to P₆ have a diameter, after diffusion of 5 microns, the top P region P₁ may have a diameter of 5.1 microns (2% greater) to obtain the benefits of the invention.

FIG. 3 shows the implant and diffusion of the top P region P₁. Thus, the layer N₁ is deposited atop layer N₂ and its P regions P₂. A mask 50 is then formed atop layer N₁ with windows 51, 52 aligned with the center of region P₂. A boron or other P species implant and diffusion is then carried out to form the enlarged diameter regions P₁ aligned to the tops of the P columns. However, the window diameter for windows 51 and 52 are larger than the implant windows in the mask for regions P₂ to P₆, creating the enlarged diameter top region P₁. Alternately, the diffusion process is carried out for a longer period of time to form enlarged regions P₁.

While the windows 51 and 52 are circular (FIG. 2), other shapes can be used for windows 51 and 52, such as elongated stripes, rectangles, ovals, or circles with projecting fingers, and the like to produce a larger P volume at the top of each column. In addition, regions P₁ need not be formed on every column over region P₂. Some column may include a P₁ region that is the same size as region P₂ or other regions in the P columns. These columns may be interspersed throughout the semiconductor device to obtain particular characteristics for the device.

Further, while the description above contemplates identical diameters (or widths) for P regions P₂ to P₆, they may be continuously tapered or stepped down in diameter from a larger diameter for regions P₂ to a smaller diameter for regions P₆. In addition, a number of upper P regions may be enlarged to some extent, and be in charge imbalance with the surrounding N type material. For example, the topmost two or three P regions may be enlarged in comparison to the lower P regions, and be in charge imbalance with the surrounding N type material.

Although the present invention has been described in relation to particular embodiments thereof, many other variations and modifications and other uses will become apparent to those skilled in the art. It is preferred, therefore, that the present invention be limited not by the specific disclosure herein.