AB etch endpoint by ABFILL compensation

Details AB etch endpoint by ABFILL compensation AB etch endpoint by ABFILL compensation

1999-07-15

2002-09-03

Utech, Benjamin L. (Department: 1765)

Semiconductor device manufacturing: process

Chemical etching

C438S690000, C438S704000, C438S706000, C438S745000

Reexamination Certificate

active

06444581

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is directed to the fabrication of semiconductor integrated circuits, and, particularly, to a method and apparatus for controlling an etch process endpoint employed during a shallow trench isolation planarization process by adjusting planarization shapes associated with dummy diffusion structures provided in the semiconductor IC.
2. Discussion of the Prior Art
As described herein, “diffusion shapes” in a Shallow Trench Isolation (STI) manufacturing process is referred to as RX, and “planarization shapes” is referred to as AB.
As known in the industry, the basic steps in a STI manufacturing process include the following: 1) RX photo -which includes a patterning step to define the shallow trench regions to be formed. Par-ticularly, as shown in FIG.
1
(
a
), an RX photoresist layer
11
is patterned over a silicon substrate
10
covered by pad nitride films
19
, to define the areas where trenches are to be formed; 2) RX etch—which includes etching the shallow trench regions. Particularly, as shown in FIG.
1
(
b
), an etching process enables formation of isolation trenches
15
and large isolation trench area
15
a
, which, as shown, isolate active semiconductor regions
18
. As shown in FIG.
1
(
b
), a thin nitride pad layer
19
remains on top of active device active areas
18
; 3) oxide deposition for filling the trench regions (e.g., using tetraethoxysilane “TEOS”). As shown in FIG.
1
(
c
) oxide deposition layer
20
conforms to the surface topography; 4) AB photo—which includes a patterning step defining areas that are to receive AB photoresist
23
, such as shown in FIG.
1
(
d
); 5) Anneal/AB planar resist apply/Anneal—which requires deposition of a planar resist film
25
above the AB photoresist layer
23
which has been subject to reflow anneal process as shown in FIG.
1
(
e
); 6) AB etch
1
(photoresist) to obtain planarity for oxide layer above the trench region as shown in FIG.
1
(
f
). It is understood that during this etch process, exposed areas of oxide enable measurable changes in gas chemistry, enabling an “emission endpoint detection”; 7) AB etch
2
(oxide) for removing the oxide above the oxide layer down to the nitride layer
19
formed on top active areas
18
such as shown in FIG.
1
(
g
); 8) Oxide chem-mech polish (“CMP”) which, as shown in FIG.
1
(
h
), planarizes the device surface by removing all of the oxide layer from on top of the pad nitride layers of active regions
18
; and, finally 9) a pad nitride strip which results in the completed structure comprising active areas
18
and oxide-filled isolation trenches
15
such as illustrated in FIG.
1
(
i
).
The correct endpoint for the AB etch step (Step
7
) is usually determined either by analysis of individual-wafer measurements of trench depth, oxide thickness and AB etch rate; or, by nitride emission endpoint detection. The former method is robust but measurement intensive, as variations in trench depth, TEOS thickness and AB etch time, and each of these contributions must be accounted for on a wafer-by-wafer basis.
FIG. 2
is a plot
10
depicting a typical nitride emission endpoint trace versus etch time. As shown in
FIG. 2
, the nitride endpoint emission signal
12
is weak and relatively flat until a point
14
is reached where the signal rapidly increases in strength indicating nearness to an etch-point limit, i.e., where the underlying nitride becomes exposed. Afterwards, the signal reaches levels off at 16 indicating the endpoint of the etch process. The latter nitride emission endpoint detection method is efficient but not robust, as some diffusion structures must be exposed in order to elicit a nitride emission endpoint signal. These diffusion structures are susceptible to AB etch and (reactive ion etch) RIE-through and to overpolish, both of which can result in defective or unreliable transfer devices.
Thus, the most difficult part of standard Shallow Trench Isolation processing is determining the correct time to stop the oxide etch. An exemplary STI etch process typically comprises a good cross-wafer uniformity for all processes; strong, repeatable emission endpoints for dry etch processes, and a uniform thickness of oxide remaining over all diffusions before chemical-mechanical (“chem-mech”) polishing. But these three conditions are generally not mutually compatible. That is, if oxide is to remain over the pad nitride that covers diffusions, there can be no nitride emission endpoint signal from that nitride.
It would thus be highly desirable to provide a method and apparatus for circumventing the problem of determining the AB etch endpoint by introducing into the design a sufficient quantity of “dummy” diffusion structures that provide a strong endpoint signal.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a technique for circumventing the problem of determining the AB etch endpoint by introducing into the STI design a sufficient quantity of “dummy” diffusion structures with “adjusted” planarization that are made to provide a strong endpoint signal during normal STI fabrication.
It is a further object of the invention to provide a method for determining the AB etch endpoint by introducing into the STI design a sufficient quantity of “dummy” diffusion structures that provide a strong endpoint signal during normal STI fabrication and, that which endpoint signal may be controlled by adjustment of the planarization shapes associated with the dummy diffusion structures.
According to the invention, the technique comprises a step 1) of placing closely-spaced arrays of small, self-similar “dummy” diffusion shapes in the available “whitespace” of a semiconductor design; then, a step 2) generating a set of “planarization shapes” that when rendered in photoresist effectively eliminates the topography associated with the “real” diffusion shapes in the design; then, a step 3) associating a “dummy planarization shape” with each dummy diffusion shape, the dummy planarization shape being constructed so that a certain amount of topography associated with an array of dummy diffusion shapes is retained when the dummy planarization shapes are rendered in photoresist along with the shapes defined in the second step; and, a step 4) proceeding with standard STI processing, with the following modifications: i) insuring that the nominal thickness by which the deposited oxide exceeds the etched trench depth is approximately that same as the retained topography from Step 3); and, ii) terminating the oxide-etch step based on a nitride emission endpoint signal arising from the dummy diffusion shapes.


REFERENCES:
patent: 4240687 (1980-12-01), Bunnell et al.
patent: 4878099 (1989-10-01), Nilarp
patent: 5070602 (1991-12-01), Kaufman
patent: 5234868 (1993-08-01), Cote
patent: 5264387 (1993-11-01), Beyer et al.
patent: 5265378 (1993-11-01), Rostoker
patent: 5358604 (1994-10-01), Lin et al.
patent: 5372673 (1994-12-01), Stager
patent: 5629242 (1997-05-01), Nagashima et al.
patent: 5648670 (1997-07-01), Blanchard
patent: 5705435 (1998-01-01), Chen
patent: 5721172 (1998-02-01), Jang et al.
patent: 5756386 (1998-05-01), Blanchard
patent: 5804084 (1998-09-01), Nasby et al.
patent: 5817567 (1998-10-01), Jang et al.
P. 15 showing a portion of a 1986 technical article describing how the depth of conformally deposited material has different thickness over different density of topography.

Affiliated with

Also associated with

Rating

AB etch endpoint by ABFILL compensation does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with AB etch endpoint by ABFILL compensation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and AB etch endpoint by ABFILL compensation will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2823623