WO2001098835A1

WO2001098835A1 - Method for determining overlay measurement uncertainty

Info

Publication number: WO2001098835A1
Application number: PCT/US2001/019577
Authority: WO
Inventors: Christopher Gould; Jurgen Preuninger
Original assignee: Infineon Technologies North America Corp.
Priority date: 2000-06-19
Filing date: 2001-06-19
Publication date: 2001-12-27
Also published as: TW494519B

Abstract

A method for determining measurement uncertainty in accordance with the present invention includes the steps of providing an intensity profile for bullet and target features for an overlay measurement (304), determining locations representing edges of the bullet and target features on the intensity profile (306), determining an average centerline between edge pairs of one of bullet features and target features (308), computing a redundant measurement between an edge pair distance of one of the bullet features and the target features and the average centerline of the other of the bullet features and the target features (312) and determining an uncertainty between two different redundant measurements (314).

Description

METHOD FOR DETERMINING OVERLAY MEASUREMENT UNCERTAINTY

BACKGROUND

1. Technical Field

This disclosure relates to inspection measurements and more particularly, to a method for determining measurement uncertainty for overlays in semiconductor processing.

2. Description of the Related Art

Overlay metrology is used for determining the quality of products, for example, integrated circuits. In particular, overlay metrology is used to determine the alignment of critical features which define, for example, an integrated circuit device. A misalignment of these features can cause electrical opens or shorts thus destroying product functionality. To ensure integrated circuit (IC) product quality, overlay metrology must be done with high accuracy and precision (i.e., 3% of minimum pattern sizes, e.g.± 5 nm for 150 nm - 180 nm features). A major contributor to measurement uncertainty is the dependency on the feature being measured. For increasingly small scale IC devices, accurate measurement of critical features relies heavily on instruments such as precision microscopes and computer algorithms. For an accurate measurement reading from a microscope, a robust and meaningful computer algorithm is necessary.

IC devices (chips) are typically fabricated on a semiconductor substrate wafer. The wafers are usually round while the IC chips are rectangular in shape and positioned in a grid across the wafer. During processing, it is necessary to monitor and align the grid of one level to that of a subsequent level to ensure proper masking and material deposition. This is required to achieve proper chip functionality. To monitor and maintain level to level alignment, features are built in to the chip pattern which are viewed with a microscope. These are called overlay measurement structures and are comprised of a bullet (on the aligning level) and a target (the level to be aligned to). Referring to FIG. 1, a top view of a standard overlay metrology structure 10 is shown. Structure 10 includes surface features such as trenches or plateaus 16 below the surface and elevated structures 12 above the surface. Each feature has edges 14 which may be used to measure dimensions between the features for inspection purposes. In the illustrative example of FIG. 1, a structure 12 is a bullet and a trench structure 16 is a target. To further illustrate structure 10 a cross-sectional view is provided in FIG. 2.

To differentiate between edges 14 and therefore mark centerlines, an optical microscope, scanning electron microscope or atomic force microscope is used such that reflected light or electrons are recorded by a photosensor or electron sensitive device and an intensity profile of the structure is produced (or in the case of atomic microscopy stylus deflections are used to create the intensity profile). FIG. 3 shows an example of an intensity profile across the bullets and targets in one direction (x or y) across structure 10 (FIGS. 1 and 2). The goal of the measurement in this example is to determine a distance between centerlines the bullet and target marks, and hence determine the resultant vector misregistration between the aligning levels (bullet to target).

The intensity profile indicates edges 14 as a change in intensity. For example, sloped curves 18, 20, 22 and 24 indicate edges 14 for structures 12 (bullet), and sloped curves 26, 28, 30 and 32 indicate edges 14 for trench structures 16 (target). Edges are defined mathematically as the points of inflection or the maximum or minimum from a first derivative calculated from the structure intensity profile. A center distance between edge pairs of the bullet, slope curves 18 and 24, is compared to a center distance between edge pairs of the target, slope curves 26 and 32, to give a misregistration value in either x or y direction. For these edge pairs, the misregistration may be defined as : '~' -'bullet_outer " ^ arget_outer

If different edge pairs of the bullet and target are compared, for example, slope curves 20 and 22 relative to 28 and 32, a slightly different misregistration value may result, such that:

*~'J-'bullet_outer " ^ arget_outer ≠ l-'-L'bullet nner " ^ arget nner

The method described above has disadvantages if asymmetric signals are present. If a different edge pair is used to determine the center lines a different location may result. This occurs due to the intensity profile signal's edge slope variation. The intensity profile signal is never perfectly symmetrical which means the slopes of the edges will be slightly different between the edges and sets of edge pairs. This error is known as mark induced shift (MIS) and has been observed to be greater than 10 nm, 3 σ (3 standard deviations).

Therefore, a need exists for a method for reducing measuring uncertainty in overlay measurements using intensity profiles. A further need exists for a method for overlay uncertainty which provides more accurate results and is robust enough to account for lot to lot overlay variation in semiconductor devices . SUMMARY OF THE INVENTION

A method for determining measurement uncertainty in accordance with the present invention includes the steps of providing an intensity profile for bullet and target features for an overlay measurement, determining locations representing edges of the bullet and target features on the intensity profile, determining an average centerline between edge pairs of one of bullet features and target features, computing a redundant measurement between an edge pair distance of one of the bullet features and the target features and the average centerline of the other of the bullet features and the target features and determining an uncertainty between two different redundant measurements.

In other methods, the step of determining an average centerline may include the steps of determining the distance between each edge pair, halving each edge pair distance, adding the halved edge pair distances to obtain a sum and dividing the sum by a number of edge pairs summed. The step of determining locations representing edges may include the step of determining a position of maximum slope for each edge to represent the edge. The step of determining an average centerline may include the steps of determining an average bullet centerline and the step of computing a redundant measurement includes the step of computing a redundant measurement between an edge pair distance of the target features and the average bullet centerline to determine a target uncertainty. The step of determining an average centerline may include the steps of determining an average target centerline and the step of computing a redundant measurement includes the step of computing a redundant measurement between an edge pair distance of the bullet features and the average target centerline to determine a bullet uncertainty. The step of computing a redundant measurement may include the steps of determining a distance between an edge pair, halving the distances of the edge pair and determining a difference between the halved distance of the edge pair and the average centerline. The step of determining an uncertainty between two different redundant measurements may include the steps of determining a difference between two redundant measurements, dividing by two the difference between the two redundant measurements. The step of providing an intensity profile for bullet and target features for an overlay measurement may include the step of providing a semiconductor device having the target features formed thereon and having the bullet to be formed thereon.

Another method for determining measurement uncertainty includes the steps of providing an intensity profile for a target pattern on a first layer and a bullet pattern on a second layer for an overlay measurement, the bullet pattern and the target pattern including features which are substantially symmetrical about a centerline, representing edges of the bullet and target features using a single position on the intensity profile, determining an average centerline between all edge pairs of the bullet features and all edge pairs of the target features, computing two redundant measurements for the bullet features by taking one half a distance of a first edge pair of the bullet features and subtracting the average centerline for all edge pairs of the target features and repeating for a second edge pair of the bullet features, computing two redundant measurements for the target features by taking one half a distance of a first edge pair of the target features and subtracting the average centerline for all edge pairs of the bullet features and repeating for a second edge pair of the target features and determining an uncertainty based on the redundant measurements for the target features and the redundant measurements for the bullet features.

In still other methods, the step of determining an average centerline may include the steps of determining a distance between each edge pair, halving each edge pair distance, adding the halved edge pair distances to obtain a sum for bullet edge pairs and a sum for target edge pairs and dividing the sum for the bullet edge pairs by a number of bullet edge pairs summed and dividing the sum for the target edge pairs by a number of target edge pairs summed. The step of representing edges of the bullet and target features using a single position on the intensity profile may include the step of determining the single position using a maximum slope for each edge to represent the edge. The step of determining an uncertainty based on the redundant measurements for the target features and the redundant measurements for the bullet features may include the steps of determining a difference between the redundant measurements for the target features and the redundant measurements for the bullet features, dividing by two the difference between the redundant measurements for the target features and the redundant measurements for the bullet features to determine a target uncertainty and a bullet uncertainty, respectively. The target features may include trenches formed in the first layer of a semiconductor device and the bullet features include resist structures formed in the second layer of the semiconductor device.

These and other objects, features and advantages of the present invention will become apparent from the following detailed description of illustrative embodiments thereof, which is to be read in connection with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

This disclosure will present in detail the following description of preferred embodiments with reference to the following figures wherein:

FIG. 1 is a top view of a structure including a bullet layer and a target layer to be measured;

FIG. 2 is a cross sectional view of the structure of FIG. 1;

FIG. 3 is a plot of an intensity profile produced by detection of reflected light or electrons across the structure of FIG. 1;

FIG. 4 is another plot of an intensity profile of trench structures shown in FIG. 1 and FIG.2 showing features for determining bullet uncertainty in accordance with the present invention;

FIG. 5 is another plot of an intensity profile of trench structures shown in FIG. 1 and FIG.2 showing features for determining target uncertainty in accordance with the present invention;

FIG. 6 is a block diagram of a measuring apparatus for use with the present invention; and

FIG. 7 is a flow diagram showing a method of determining uncertainty in accordance with the present invention. DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

This disclosure relates to inspection measurements and more particularly, to a method for determining measurement uncertainty for overlays in semiconductor processing. The present invention will be described by way of an illustrative example. The structures of FIGS. 1 and 2 and the intensity profile of FIG. 3 will be used to describe in detail the method for calculating overlay uncertainties in accordance with the present invention. The present invention provides a method which may be employed on conventional inspection systems. The present invention provides at least two redundant measurements between edges pairs of a bullet (or target) and an average centerline of all edge pairs of the target (or bullet) on the intensity profiles to determine an overall misregistration value between layers for semiconductor processing or inspection.

Referring now in specific detail to the drawings in which like reference numerals identify similar or identical elements throughout the several views, FIG. 4 is an intensity profile of a scan across resist structures (bullet) 12 and trench structures (target) 16 of FIGS. 1 and 2. Edge slope curves 102 show variations along each edge slope curve. Edge slope curves indicate a change in intensity due to surface features. In an alternative embodiment, an atomic force microscope provides deflection profiles which may be used with the present invention. The change in intensity is relative to a baseline or threshold intensity 106. The threshold intensity may be chosen as a set value of the maximum intensity, for example 90 % of maximum intensity output from an energy source. The intensity profile is developed between a datum reference (y- coordinate equal to zero) and the threshold which may be chosen in accordance with predetermined criteria. Edge slope curves 102 represent edges of bullet 12 and/or target 16 structures (FIG. 1). Edge pairs are labeled with a lower case and an upper case letter, for example an edge pair for target 16 is d, D, another edge pair for target 16 is c, C. Edge pairs for bullet 12 include a, A and b, B. Edge pairs include edges which have a symmetrical counter part edge on a same layer, i.e, bullet layer, for example a, A. Dots 108 on each edge slop curve 102 indicate a point of maximum slope as determined by an overlay measurement algorithm. Each dot 108 represents the position of the respective edges of bullet 12 and target 16, i.e., mark edges. The coordinate values of dots 108 will be used to illustratively compute misregistration in accordance with the present invention. Other methods of representing the edges in an intensity profile may be used as well.

In accordance with the invention, FIG. 4 the intensity profile is used to determine the bullet uncertainty. A target centerline (Target CL), T_CL, is computed by averaging the target edge pairs. In the example, this means averaging the average distance between c, C and d, D. For the present disclosure, a distance along the x-axis between a edge pairs is designated by aA, bB, cC and/or dD. For example, the distance between edge d and edge D may be written as dD. A location of an average target centerline T_CL may be written as:

T_CL = [cC/2 + dD/2] /2

The T_CL returns a value on the x-axis for the placement of the average target centerline relative to a reference datum. If more edge pairs are desired a general formula may be employed for T_CL as follows: where N is the total number of target edge pairs and L is the x-axis distance between the edges of each edge pair.

For each bullet edge pair, an average distance is computed. For example, the average distance for edge pair b, B is bB/2, and for a, A, it is aA/2. These average distances are used to compute redundant measurements. Since two edge pairs exist for the bullet, two redundant measurements are used.

Redundant Measurement 1 = aA/2 - T_CL; and

Redundant Measurement 2 = bB/2 - T _L-

These redundant measurements, represent a difference between an x-axis value centerline for each edge pair of the bullet and the average target edge pair centerline. The redundant measurements are then used to determine bullet uncertainty, in accordance with the present invention. The bullet uncertainty may be calculated in accordance with the present invention using the following equation:

(Redundant measurement 1 - Redundant measurement 2 )/2.

Referring to FIG. 5, the intensity profile is used to determine the target uncertainty. A bullet centerline B_CL is computed by averaging the bullet edge pairs. In the example, this means averaging the average distance between a, A and b, B. For the present disclosure, a distance along the x-axis between edge pairs is designated by aA, bB, cC and/or dD as described above. A location of an average bullet centerline B_CL m y be written as: B_C = [aA/2 + bB/2] 12

The B_CL returns a value on the x-axis for the placement of the bullet centerline relative to a reference datum. If more edge pairs are desired a general formula may be employed for B_C as follows: where N is the total number of bullet edge pairs and L is the x-axis distance between the edges of each edge pair.

For each target edge pair, an average distance is computed. For example, the average distance for edge pair c, C is cC/2, and for d, D, it is dD/2. These average distances are used to compute redundant measurements. Since two edge pairs exist for the target two redundant measurements are used.

Redundant Measurement 3 = B _L - cC/2 , and

Redundant Measurement 4 = B_CL - dD/2.

These redundant measurements, represent a difference between an x-axis value centerline for each edge pair of the target and the average bullet edge pair centerline. The redundant measurements are then used to determine target uncertainty, in accordance with the present invention. The target uncertainty may be calculated in accordance with the present invention using the following equation:

(Redundant measurement 3 - Redundant measurement 4 )/2.

By computing overlay uncertainties for bullets and targets in this way, the present invention reduces the effects of poor quality overlay measurement points. Advantageously, the present invention uses averaging to improve overlay uncertainties and provides less sensitivity to lot to lot variations in measurements of semiconductor devices or other measurements. The amount of uncertainty introduced into the overlay measurement is determined by processing both the bullet and target. Tracking this uncertainty makes it possible to identify and control the processing conditions which contribute to mark degradation (i.e., degradation in the proper positioning of dots 108). Referring to FIG. 6, a measuring apparatus 200 includes a microscope 202, such as a scanning electron microscope, an optical microscope or an atomic force microscope having a stage 204 for positioning a structure or specimen 206 to be measured. An energy source 208 irradiates structure 206. A photosensitive device or electron sensitive device 209 collects reflected intensities and stores the data in a storage device 210. Alternately, deflections of a stylus (not shown) may be used to develop intensity profiles and deflection data stored in storage device 210. A processor 212 is used to perform calculations on the data in accordance with the present invention. A monitor 214 may also be included for real-time viewing of structure 206 during operation. In preferred methods, the edge pairs may be optimized by tool adjustments to the measuring apparatus. For example, signal comparison may be performed by scanning through focus. More symmetrical signals can be obtained for the intensity profile if, for example, a wafer is placed on stage 204 and moved outside of an isofocal plane as determined by the focus of the microscope 202. In this way, a better focus of surface features may be obtained. This is achieved if stage 204 can scan in the z direction. This method is iterative. A first measurement is made. Then an adjustment is made to a focus control 216 (FIG. 6) to change the focus of microscope 202. A second measurement is made and the quality is assessed against the first measurement. If there is an improvement further iterations are performed to further refine the measurement. If there is no improvement further adjustments may be made to improve the quality of the measurement.

Referring to FIG. 7, a method in accordance with the present invention may be generalized as shown. In step 302, a measurement system 200 is provides along with a specimen to be measured. The specimen including a bullet pattern and a target pattern. In step 304, an intensity (or deflection profile) is obtained of the bullet and target patterns. In step 306, the intensity profile is processed to determine maximum slope points to estimate the locations of edges of the bullet and target. Maximum slope calculation may be performed using an overlay measurement algorithm performed on, for example, a Quaester™ available commercially from Bio-Rad, Inc. In step 308, the average target centerline and the average bullet centerline are determined as described above. In step 310, distances for edge pairs are computed and halved. In step 312, the redundant measurements are calculated as described above. In step 314, bullet and target uncertainties are calculated.

Having described preferred methods and embodiments method for determining overlay measurement uncertainty (which are intended to be illustrative and not limiting), it is noted that modifications and variations can be made by persons skilled in the art in light of the above teachings. It is therefore to be understood that changes may be made in the particular embodiments of the invention disclosed which are within the scope and spirit of the invention as outlined by the appended claims. Having thus described the invention with the details and particularity required by the patent laws, what is claimed and desired protected by Letters Patent is set forth in the appended claims.

Claims

WHAT IS CLAIMED IS:

1. A method for determining measurement uncertainty comprising the steps of: providing an intensity profile for bullet and target features for an overlay measurement; determining locations representing edges of the bullet and target features on the intensity profile; determining an average centerline between edge pairs of one of bullet features and target features; and computing a redundant measurement between an edge pair distance of one of the bullet features and the target features and the average centerline of the other of the bullet features and the target features; and determining an uncertainty between two different redundant measurements.

2. The method as recited in claim 1, wherein the step of determining an average centerline includes the steps of: determining the distance between each edge pair; halving each edge pair distance; adding the halved edge pair distances to obtain a sum; and dividing the sum by a number of edge pairs summed.

3. The method as recited in claim 1, wherein the step of determining locations representing edges includes the step of: determining a position of maximum slope for each edge to represent the edge.

4. The method as recited in claim 1, wherein the step of determining an average centerline includes the steps of determining an average bullet centerline and the step of computing a redundant measurement includes the step of computing a redundant measurement between an edge pair distance of the target features and the average bullet centerline to determine a target uncertainty.

5. The method as recited in claim 1, wherein the step of determining an average centerline includes the steps of determining an average target centerline and the step of computing a redundant measurement includes the step of computing a redundant measurement between an edge pair distance of the bullet features and the average target centerline to determine a bullet uncertainty.

6. The method as recited in claim 1, wherein the step of computing a redundant measurement includes the steps of: determining a distance between an edge pair; halving the distances of the edge pair; determining a difference between the halved distance of the edge pair and the average centerline.

7. The method as recited in claim 1, wherein the step of determining an uncertainty between two different redundant measurements includes the steps of: determining a difference between two redundant measurements; dividing by two the difference between the two redundant measurements.

8. The method as recited in claim 1, wherein the step of providing an intensity profile for bullet and target features for an overlay measurement includes the step of providing a semiconductor device having the target features formed thereon and having the bullet to be formed thereon.

9. A method for determining measurement uncertainty comprising the steps of: providing an intensity profile for a target pattern on a first layer and a bullet pattern on a second layer for an overlay measurement, the bullet pattern and the target pattern including features which are substantiality symmetrical about a centerline; representing edges of the bullet and target features using a single position on the intensity profile; determining an average centerline between all edge pairs of the bullet features and all edge pairs of the target features; and computing two redundant measurements for the bullet features by taking one half a distance of a first edge pair of the bullet features and subtracting the average centerline for all edge pairs of the target features and repeating for a second edge pair of the bullet features; computing two redundant measurements for the target features by taking one half a distance of a first edge pair of the target features and subtracting the average centerline for all edge pairs of the bullet features and repeating for a second edge pair of the target features; and determining an uncertainty based on the redundant measurements for the target features and the redundant measurements for the bullet features.

10. The method as recited in claim 9, wherein the step of determining an average centerline includes the steps of: determining a distance between each edge pair; halving each edge pair distance; adding the halved edge pair distances to obtain a sum for bullet edge pairs and a sum for target edge pairs; and dividing the sum for the bullet edge pairs by a number of bullet edge pairs summed and dividing the sum for the target edge pairs by a number of target edge pairs summed.

11. The method as recited in claim 9, wherein the step of representing edges of the bullet and target features using a single position on the intensity profile includes the step of: determining the single position using a maximum slope for each edge to represent the edge.

12. The method as recited in claim 1, wherein the step of determining an uncertainty based on the redundant measurements for the target features and the redundant measurements for the bullet features includes the steps of: determining a difference between the redundant measurements for the target features and the redundant measurements for the bullet features; dividing by two the difference between the redundant measurements for the target features and the redundant measurements for the bullet features to determine a target uncertainty and a bullet uncertainty, respectively.

13. The method as recited in claim 1, wherein the target features include trenches formed in the first layer of a semiconductor device and the bullet features include resist structures formed in the second layer of the semiconductor device.