US 4,981,813 AGrant

Pad Oxide Protect Sealed Interface Isolation Process

Issue Date:1991-01-01

•2 Claims

•4 Drawing Sheets

Abstract

Field oxide regions are formed between active regions of a silicon substrate by forming over the substrate a sandwich of silicon dioxide, silicon nitride and silicon dioxide layers, opening the layers to expose a portion of the silicon substrate, removing a layer of the exposed substrate, forming side wall spacers on the edges of the opening, removing a layer of the silicon substrate exposed between the side wall spacers, and then reaching the exposed substrate for the thermal oxidation of the exposed substrate for forming the field oxide region. In those structures in which the field oxide is buried in the substrate as shown in FIG. 12, it may be feasible to use thicker field oxide regions and thereby to reduce the need for the heavily doped surface layer under the field oxide.

Metadata

Assignee

SGS-Thomson Microelectronics, Inc.

Inventors

Frank R. Bryant
Yu-Pin Han
Fu-Tai Liou
Tsiu C. Chan

Application Information

Application Number:US 2793431

Filing Date:1988-12-02

Priority Date:1987-02-24

Art Unit:134

Classifications

IPC:

H01L 2176H01L 21304

Field of Search:

156437644;648;649;653;66263;64;69;72;73

Patent Drawings (4 sheets)

Description

Brief Description of the Drawing

FIGS. 1-5 show the silicon workpiece at various stages of its processing in accordance with one embodiment of the invention,

FIGS. 6-10 show the silicon workpiece at various stages of its processing in accordance with another embodiment of the invention, and

FIGS. 11 and 12 show the silicon workpiece at the final two stages of a variation of the process illustrated in FIGS. 6-10.

Detailed Description

In FIG. 1, there is shown a portion of a silicon monocrystalline wafer, designed to be diced into a plurality of silicon chips each embodying an integrated circuit, after suitable processing. To this end, each chip is to include a plurality of active regions in which are to be formed one or more active circuit elements and which are to be separated from one another by relatively thick field oxide regions, as is well known in the art.

The drawing will focus on the formation of only one of such thick field oxide regions to simplify the disclosure.

As shown in FIG. 1, the silicon wafer has formed on one of its major surfaces a succession of layers including a thermally grown silicon dioxide layer 12, a silicon nitride layer 14, and a silicon dioxide layer 16. The silicon dioxide layer 12 is often described as a pad layer and it primarily serves as a stress release layer and etch stop layer of nitride etch between the silicon substrate and the silicon nitride layer that minimizes defects in the silicon substrate.

Various techniques are available for forming such an arrangement. Advantagerously, the silicon nitride layer 14 is formed by a low pressure chemical vapor deposition process and the silicon dioxide layer 14 is formed by a chemical vapor deposition process. Typical dimensions found suitable are about 200 Angstroms for layer 12, 800 Angstroms for layer 14, and 1000 Angstroms for layer 16. There then is formed an aperture or opening in the multiple layers, slightly larger than that desired for the localized field oxide region desired. Typically, the aperture would be about 1.2 microns wide and considerably longer lengthwise, the direction normal to the plane of the drawing. To improve the packing density of active elements in a chip, the width of the field oxide region should be as narrow as the technology makes practice.

Various techniques are available for providing the desired aperture, generally involving photolithographic patterning of a suitable photoresist mask and subsequent etching of the pattern exposed by the mask. Since the aperture desirably should have essentially vertical side walls to save space, ion milling or anisotropic reactive ion etching (RIE) generally will be performed at the present stage of the technology.

In FIG. 2, there is shown the workpiece 10 with an opening 18 which penetrates to the silicon substrate.

Next there is deposited over the workpiece a conformed coating of silicon nitride 20, typically also deposited by LP CVD and the resultant is shown in FIG. 3. The thickness of the layer 20 will determine how much the aperture 18 will be narrowed so it is chosen appropriately. A thickness of between 2000 and 2500 Angstroms typical for the particular embodiment being described.

This is then followed by an isotropic etching only until the silicon nitride layer 20 has been removed from the bottom of the central portion of the original aperture so , that the etching leaves the side wall spacers 22A and 22B along the original edges of the aperture, as shown in FIG. 4. As is seen in FIG. 4, these spacers result in a narrowing of the width of the exposed silicons in the apertures.

Thereafter the workpiece is heated in the manner usual for forming the field oxide, which typically involves heating the workpiece in an oxidizing atmosphere for a suitable time for achieving the desired thickness, typically at least several thousands Angstroms.

In FIG. 5, on an expanded scale, there is shown one edge of the resultant field oxide region 24. As seen in this figure, the upper surface of the field oxide will rise significantly above the surface of the original silicon surface. This may pose a disadvantage particularly when the interconnect wiring for the integrated circuit comprises one or more levels of metallization over the silicon chip which are desirably as planar.

FIGS. 6-10 show a modification of the basic process described which permits a field oxide thickness of the same thickness but with a reduction in the amount the top surface of the field oxide rises above the surface of the silicon substrate along the active regions between the field oxide region.

In these figures, FIG. 6 corresponds to FIG. 1, but as shown in FIG. 7, the depth of the aperture 28 shown in FIG. 7 penetrates below the original level of the silicon substrate. To this end, after baring the silicon surface by removing layers 12, 14 and 16, the etching is continued either by a separate step or by a continuation of the same process when ion milling is used to form the aperture, to remove a layer of the desired thickness of the silicon substrate, shown as 29 in FIG. 7. A thickness of about a thousand Angstroms is suitable.

Further processing proceeds as before, involving deposition of a conformed silicon nitride layer 20 as shown in FIG. 8, patterning the layer 20 to expose the silicon substrate but leaving side wall spacers 22A, 22B as seen in FIG. 9, and then heating in an oxidizing atmosphere for the requisite time to form the field oxide 24 as shown in FIG. 10.

A further reduction in the effective height of the field oxide can be achieved by the modifications illustrated in FIG. 11. To this end, after formation of the side wall spacers, as illustrated in FIG. 9, an additional layer of the exposed silicon substrate is removed, as designated by numeral 30 in FIG. 11. Again a layer of a thickness of about a thousand Angstroms should be suitable. This layer can be removed as the earlier layer 29 was removed, for example, either by ion milling or reactive ion etching.

The resultant is shown in FIG. 12 where the upper surface of the field oxide region 24 is shown as extending relatively less above the surface of the silicon surface along the active regions between field oxide regions.

It can be appreciated that the particular dimensions described are merely illustrative. Additionally, it should be apparent that other materials could be used for the various layers 12, 14-16 and 20 so long as they function appropriately. In particular, it is feasible to omit layer 16 in the process described at some sacrifice in the profile of the side spacers left after etching of the conformed layer.

Accordingly, the specific embodiments which have been described should be viewed merely as illustrative of the general principles involved and various other embodiments may be devised by a worker in the art without departing from the spirit and scope of the invention. For example, there may be omitted the step of etching the surface layer 29 as shown in FIG. 7 although there is later removed the surface layer 30 as depicted in FIG. 11.

Additionally, as known to workers in the art, it is usual to include in the substrate below the field oxide a heavily doped surface layer to reduce the possibility o surface inversion. The processes described can be modified to include before the thermal oxidation step one or more ion implantation steps in the usual fashion to form such heavily doped layers if desired. Moreover, when the field oxide is used to overlie the junction between P-wells and N-wells in a substrate to be used for CMOS devices, it may be desirable to include a heavily doped N-type surface layer in the N-well and a heavily doped P-type surface layer in the P-well.

Claims

What is claimed:

1. A process for forming oxide regions between active regions of a silicon substrate used in the manufacture of integrated circuits comprising the steps of: forming a first layer of silicon oxide on a first surface of the silicon substrate; providing a first layer of silicon nitride over the first layer of silicon oxide; forming a second layer of silicon oxide over the first layer silicon nitride; forming apertures through the second layer of silicon oxide, the first layer of silicon nitride, the first layer of silicon oxide, and extending to a first bottom portion below the first surface of the silicon substrate by a predetermined amount, said apertures having substantially vertical sidewalls; forming spacer portions of silicon nitride on the aperture sidewalls leaving the first bottom portion of the silicon substrate exposed; etching a second bottom portion into the first bottom portion, said second bottom portion having substantially vertical sidewalls; and oxidizing the second bottom portion of the silicon substrate.

2. The process of claim 1 wherein the step of forming sidewall spacer portions of silicon nitride comprises the steps of depositing a second layer of silicon nitride over the second layer of silicon oxide and partially filling the apertures; and etching the second layer of silicon nitride for exposing the second layer of silicon oxide and the second bottom portion of the silicon substrate while leaving sidewall spacer portions of the silicon nitride on the sidewall of the apertures.

Patent Citations (16)

Patent	Date	Inventor
US4272308	1981-06-01	Vourshney
US4495025	1985-01-01	Haskell
US4526631	1985-07-01	Silvestri et al.
US4534824	1985-08-01	Chen
US4560421	1985-02-01	Maeda et al.
US4579621	1986-04-01	Hine
US4585513	1986-04-01	Gale et al.
US4626281	1987-01-01	Buiguez et al.
US4645564	1987-02-01	Morie et al.
US4686000	1987-08-01	Heath
US4690729	1987-09-01	Douglas
US4758530	1988-07-01	Schubert
US4868136	1989-09-01	Ravaglia
EPX284456	1988-09-01
JPX63842	1982-04-01
NLX3418638	1984-11-01

Non-Patent Literature (2)

Swada et al. "Electrical Properties . . . Swami Technology" IEEE Trans. on Elec. Dev. vol. ED-32 No. 11, 11/85, pp. 2243-2248.
Abstract 273, Inulshi et al. Extended Abs. vol. 86-2 Oct. 19-24, 1986.