WO2002041389A2

WO2002041389A2 - A method for monitoring line width of electronic circuit patterns

Info

Publication number: WO2002041389A2
Application number: PCT/EP2001/013305
Authority: WO
Inventors: David H. Ziger
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2000-11-16
Filing date: 2001-11-14
Publication date: 2002-05-23
Also published as: WO2002041389A3; JP2004514292A; EP1334407A2

Abstract

To measure line width at best focus in a projected circuit pattern, a set pattern is formed on a substrate, and then an array of test patterns formed adjacent the set pattern. The focus varies by predetermined steps along the array, with the focus of a test pattern at an intermediate position along the array corresponding to the focus of the set pattern. By scanning the test patterns and measuring the line width at each test pattern, data can be obtained to provide a curve indicative of focus relative to line width which shows either a maximum or a minimum line width at that focus, dependent on exposure data. This information can be used to change the programming of a stepper for the focus of a projector to obtain focus independent line width and/or to determine the stability of line widths independent of machine focus and/or to determine the stability of line widths independent of the machine focus.

Description

A method for monitoring line width of electronic circuit patterns

This invention relates to the production of electronic circuits on substrates and in particular to the production of conductor lines of such circuits.

BACKGROUND OF THE INVENTION In the production of electronic circuits, it is desired that the line width of the pattern features be maintained at, or very close to, a critical dimension - defined in the circuit specification. The line width is a function of focus of the projected pattern, processing of the exposed substrate, and other possible variations.

SUMMARY OF THE PRESENT INVENTION

With the present invention, a test pattern is exposed at nominal exposure and focus conditions at several locations on a substrate. At each location there are a plurality of individual patterns, each exposed at a slightly different focus. On scanning the test pattern a series of values are obtained for line width in each of the individual patterns. These results are resolved into data which can be regressed to, for example a second order equation, or some other equation, as desired, to provide a polynomial. The supposed line width versus focus exhibits a convex curve, though a concave curve is also possible depending on the relative exposure dosing. The critical dimension is calculated for each location, and the appropriate focus. This information can be then used to control the stepper so that the ideal focus can be provided to give the critical dimension. Additionally, the line width best focus enables the user to decouple source of line width variation due to process variations from focus variations.

Broadly, the invention comprises forming a first pattern on a substrate at a normal focus, forming a test array adjacent the first pattern, the array comprising a plurality of test patterns each of a different predetermined focus, the focus range extending along the array and the focus at an intermediate position equal to the focus of the first pattern, measuring the line width in each test pattern and comparing with the line width to produce a curve indicative of line width versus focus; and to produce a focus value for an extreme line width; determining the difference between the focus for this line width and the focus for the first pattern. This difference can be used to program a stepper on the pattern projector to maintain a focus for an extreme line width production.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates diagrammatically an example layout of a substrate;

Figure 2 illustrates a curve obtained from the resolving of the scanned line width, showing line width relative to focus;

Figure 3 is a table giving an example of maximum critical dimension mapping across a wafer substrate. Considering Figure 1, the wafer 12 is initially scanned, and the focus measured, by the scanner, at various positions 8, to determine that there is correct positioning of the wafer. This is conventional. This provides a focus value - hereafter referred to as the nominal value. No actual exposure of a pattern occurs.

Following the initial scanning, a series of set patterns 10 are exposed on the wafer at nominal exposure conditions and at the nominal focus conditions. A series of test arrays 14 are exposed on the wafer at positions adjacent the patterns 10. In the test arrays 14 a plurality of test patterns I6^l-I6^'n are exposed at slightly different foci, to form what can be termed focus arrays. The set patterns 10 provide a nominal or datum focus as provided by the stepper for the projector. In an array 14, as an example, the focus of test pattern 16^"1 will be lower than the nominal value and at 16^"" where the nominal focus will be higher than the value, the intermediate test patterns 16 having foci stepping from that of 16^"1 to that of 16^"n.

Normally one set pattern is formed and then the adjacent array, the process continued until all set patterns and test arrays are formed.

The same feature is measured on both the nominal exposure set patterns 10 and the focus arrays 14. At nominal focus and exposure, the average line width and variation is calculated. For the focus arrays, the data are regressed versus focus and the critical dimension is calculated at each test pattern 16. Also, the focus at which the critical dimension is found is recorded.

The average critical dimension at best focus is a direct measurement of the machine focus effect on line width when compared to a line width exposed at normal conditions. Changes in line width at best focus are a measure of process effect, or exposure effect, independent of machine focus.

Figure 2 illustrates one example of a program for a particular wafer substrate. The nominal focus for the set patterns 10 is given a focus value as 0. The foci for the test patterns 16 is varied from -0.2 μm to +0.02 μm. The critical dimension is indicated by the dotted line 20. The curve 22 is obtained from the data obtained from the patterns 16 resolved as above. It will be seen that the critical dimension is obtained at a slightly negative focus; almost -0.1 μm. This information can be applied to the stepper so that processing of wafers with the particular set up will proceed with the stepper maintaining the focus slightly reduced from the nominal. The stepper will maintain this focus independent of any other variations. In addition to providing for a focus independent critical dimension for lines, the invention also provides information concerning other features of a process. The table of Figure 3 gives values for line width at various positions across a wafer, the left and top values giving Y and X positions. Generally the critical dimension or line width varies from 0.191 to 0.200, but shows one locality with a major difference, 0.178. This may indicate that there is a fault in the process, such as in the coating of the substrate, or in development or some other possible variable. Any discrepancy is not caused by focus effects. Also, by looking at the values in the table, it is possible to determine whether the line width data is within specifications.

Without the use of the present invention, the stepper controlling the focus of the pattern projector is programmed to maintain the best possible focus, although this may not result in the optimum line width. With the present invention, it is possible to determine when an adjustment needs to be made to the stepper program, stepper exposure dose, or other process step, so that optimum line width and product quality is obtained. This allows for process control avoiding confounding focus effects with process and exposure influences on line width.

Broadly, the method is to produce a first, set pattern, in the conventional manner, the stepper focusing the pattern projection at the best focus. Thus a series of test patterns in a test array are formed, the focus of each test pattern differing from the rest.

Conveniently, a central pattern is at the same focus as that of the first, set, pattern, the focus stepping up or down for the remaining test patterns in the array. The line width is measured for each test pattern in an array and from this it is possible to determine the best focus for optimum line width. This information is used to control the stepper, in effect giving a bias to the stepper control.

The number of set patterns 10, and the number of arrays 14 can vary. The size of the substrate will influence the number to some extent. Also, the number of test patterns 16 in an array can also vary. However, it is desirable that the set patterns 10 and arrays 16 be spaced apart over the area of the substrate and the set patterns 10 and arrays 16 be fairly closely associated. At the limit, only one set pattern and one array is essential, but normally a plurality of set patterns and arrays will be used.

Depending upon exposure dosage, either a maximum line width or a minimum line width can be detected. The process enables the day-to-day focus effects on line widths to be separated from other factors.

The process of the invention can be used solely to monitor a product line, to provide information, for use by an operator, for example, or it can be used to directly control the stepper, if desired.

Claims

CLAIMS:

1. A method for monitoring line width in circuit patterns produced by a projector having a stepper for adjusting the focus of the projector, comprising: initially focusing on a substrate 12 to determine a nominal focus; forming at least one set pattern 10 on said substrate 12 by projection at said nominal focus; forming a test array 14 adjacent each set pattern, each said array comprising a series of test patterns 16^l-16ⁿ, the focus of each pattern 16 differing from adjacent patterns and extending in a predetermined step arrangement from one end of an array to the other; the focus of an intermediate test pattern corresponding to the focus of the adjacent set pattern; measuring the line width in each of said test patterns 16^"- 16" and comparing with the focus at the set pattern 10, to produce a curve indicative of line width dependent on focus to produce a focus value for a critical dimension line width; determining the differences between said focus value and the focus for said set patterns ; using said difference as desired to program data to obtain a critical dimension.

2. A method as claimed in claim 1 , comprising forming a plurality of set patterns 10 spaced apart on said substrate 12.

3. A method as claimed in claim 2, including positioning said set patterns 10 adjacent a center of the substrate 12 and adjacent to a periphery of the substrate 12.

4. A method as claimed in claim 1 , each test array 14 comprising a linear array.

5. A method as claimed in claim 4, including scanning the test patterns I6^'l-I6ⁿ in an array 14 in sequence.

6. A method as claimed in claim 5, including scanning the test arrays 14 in sequence.

7. A method as claimed in claim 1, said critical dimension a maximum line width.

8. A method as claimed in claim 1, said critical dimension a minimum line width.