TW494519B - Method for determining overlay measurement uncertainty - Google Patents

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

A7

Background of the Invention 1. The technical field, the present invention is about detection and measurement. For the surname of the jealousy emperor and officials, it is about the uncertainty of measurement in the decision to cover the private edition of the semiconductor. 2 · Relevant technical description The coverage measurement system is used to determine the product. Then, live #, 7 口 口 口 壳, such as integrated circuits. In particular, the overlay measurement is used to determine the alignment of key key features, such as its definition of the Jieji Road device. Misalignment of these characteristics π ^ Misalignment can cause electrical open circuits or short circuits, which can impair product functionality. . In order to ensure that the integrated circuit (ic) production mouth mouth quality α coverage measurement must be completed in ancient, standard installation, with the same accuracy and accuracy (ie, the minimum pattern size Θ 3%, for example, 15〇__18. _ Features It is set to 5n0. The main impact on measurement uncertainty is the dependence on the measured features. For the increasingly small IG t setting, the accurate measurement of key features will rely heavily on instruments, such as precision microscopes and computer algorithms. For accurate measurement readings from a microscope, an accurate and meaningful algorithm is required. The IC device (wafer) is basically fabricated on a semiconductor substrate wafer. The wafer is usually round, and the IC chip is an AI ancient belt. ^ The Japanese film is long and shaped, and is located in a grid across the wafer. During the process, it needs to monitor and align one level of the grid and its subsequent levels to ensure proper masking. And material deposition. This is needed to achieve proper wafer function. In order to monitor and maintain the level-to-level alignment, features are created on the pattern, which can be viewed by a microscope. These are called It is a coverage measurement structure, which includes a bullet (at this alignment level) and a target (to be == this paper size applies the Chinese National Standard (CNS) A4 specification (210X297 male I) — 494519 Α7 Β7 V. Invention Description

Class). Please refer to the top view of the wind measurement structure 10. Structure 10 includes surface features, such as trenches or platforms 16 below the surface, and elevated structures 12 above the surface. 1 Z female features have edge transitions between features that can be used to achieve detection purposes. In the illustrative example of FIG. 1, a structure 12 is a bullet and a target. In order to further illustrate the structure 10 diagram. A trench structure 16 provides a cross section in FIG. 2 in order to distinguish the edges 14 and thus the centerline. It uses an optical microscope, scanning electron microscope or atomic force microscope, so the reflected light or electron system is sensed by a light Or an electronically sensitive device to record and generate an intensity profile (or in the case of an atomic microscope, use a stylus reflection to generate an intensity profile). Figure 3 shows an example of the intensity profile across the bullet and target in one direction (X or y) across the structure 10 (Figures 1 and 2). The measurement purpose in this example is to determine the distance between the bullet and the centerline of the target mark, thereby determining the misregistration of the result vector between the alignment level (bullet to target). This intensity profile represents the edge 14 as a change in intensity. For example, the slope curves 18, 2 0, 2 2 and 24 represent the edges 14 of the structure 12 (bullet), and the slope curves 2 6, 2 8, 30 and 32 represent the trench structure 16 (target ) Edge 1 4. The edges are based on mathematical definitions as the bending points, or the maximum and minimum values obtained from a differential calculation of the structural strength profile. The center distance between the edge pairs of the bullet's slope curves 18 and 2 4 is compared to the target, and a center distance between the edge pairs of the slope curves 26 and 32 is obtained to obtain the X or y direction. Misregistered value. For this wife's edge pairing, the misregistration can be defined as follows: _______ This paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm)

Binding

Line 494519 A7 B7, description of the invention CLbullet_outer ~ CLtarget_outer If you compare the different edges of the S bullet and the target pairing, such as the tilt curve 2 0 and 2 2 compared to 2 8 and 32, you can get a slightly different misregistration, so: CLbullet_outer — CLtarget_outer φ CLbullet_

CL target_inner The above method has disadvantages when it has an asymmetric signal. If a different edge pair is used to determine the centerline, a different position will be obtained. This is due to the change in edge slope of the intensity profile signal. The intensity profile signal cannot be perfectly symmetrical, which means that the slope of the edge will be slightly different between the edge and the combination of edge pairs. This error is the well-known marker-induced offset (M I S), which has been observed to be greater than 100 nm, 3σ (3 standard deviations). Therefore, there is a need for a method to reduce the measurement uncertainty of the coverage measurement using the intensity profile. There is a further need for a method of covering uncertainty that can provide more accurate results and be accurate enough to withstand batch and batch coverage changes in semiconductor devices. SUMMARY OF THE INVENTION According to the present invention, a method for determining measurement uncertainty includes the steps of providing an intensity profile of a bullet and a target feature covering the measurement, determining a position representing the edge of the bullet and the target feature at the intensity cycle, and determining An average centerline between the edge feature of one of the bullet feature and the target feature, calculate an edge pairing distance between the bullet feature and one of the target features and the average centerline of the other bullet feature and the target feature. Between redundant measurements and determining the uncertainty between two different redundant measurements. This paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) 494519 A7 B7 V. Description of the invention (4 In another method, the step of determining an average centerline may include the following steps. • Decide each edge pairing The distance between each halving of each edge pairing distance ', adding the halved edge pairing distance to get a sum, dividing the sum by the number of total edge pairs added. The step of determining the representative edge position may include determining each The position of the edge's maximum slope represents the step of the edge. The step of determining an average centerline may include a step of determining an average bullet centerline, and the step of calculating a redundant measurement includes calculating a step in the target feature. A redundant measurement between the edge pairing distance and the average bullet centerline to determine a target uncertainty step. The step of determining an average centerline may include a step of determining an average target centerline, and the calculation is redundant The remaining measurement step includes calculating an edge pairing distance between the bullet characteristics and a redundant measurement between the average target centerline to determine The step of bullet uncertainty. The step of calculating a redundant measurement may include the steps of determining the difference between an edge pair, halving the distance of the edge pair, and determining the halving distance of the edge pair and the average center. The difference between the lines. The step of determining an uncertainty between two different redundant measurements may include the following steps: determining the difference between two redundant measurements, the difference between the two redundant measurements Divide by 2. The step of providing an intensity profile of the bullet and target feature covering the measurement may include providing a semiconductor device having the target feature formed thereon and forming the bullet thereon. Another type of measurement uncertainty is determined The method includes the following steps. For a coverage measurement, an intensity profile of a target pattern is provided on a first layer, and a bullet contour on a second layer. The bullet pattern and the target pattern include a centerline Roughly symmetrical feature, used on this intensity profile

Binding

This paper size applies to Chinese national standards (ci ^ 7X ^^ (210 χ 29 · 7mm)

5. Description of the invention (5) The edge of the bullet and the target feature at a single position determines an average centerline between the edge pairs of all the bullet features and the edge pairs of all the target features, by taking the bullet feature A half of the distance of a first edge pairing, and subtracting the mean center line of all edge pairs of the target feature, and repeating the second edge pairing of the bullet feature to calculate two redundant measurements of the bullet feature, By taking half the distance of a first edge pairing of the target feature, subtracting the average centerline of all edge pairs of the bullet feature, and repeating the second edge pairing of the target feature, the two features of the target feature are calculated A redundant measurement, and an uncertainty is determined according to the redundant measurement of the target characteristic and the redundant measurement of the bullet characteristic. In another method, the step of determining an average centerline may include the steps of: determining a distance between each edge pair, halving each edge pairing distance, and adding the halved edge pairing distance to obtain a bullet edge The sum of the pairings and the sum of the target edge pairs, and the sum of the bullet edge pairings divided by the total number of added bullet edge pairs, and the sum of the target edge pairings divided by the total number of target edge pairs added. The step of using a single position on the intensity profile to represent the edge of the bullet and the target feature may include the step of using the maximum slope of each edge to determine the single position to represent the edge. The step of determining an uncertainty based on the redundant measurement of the target characteristic and the redundant measurement of the bullet characteristic may include the following steps: determining the difference between the redundant measurement of the target characteristic and the redundant measurement of the bullet characteristic , Dividing the difference between the redundant measurement of the target feature and the redundant measurement of the bullet feature by 2 to determine a target uncertainty and a bullet uncertainty, respectively. The target feature may be formed in the first layer of a semiconductor device. This paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X 297 public love) A7 B7 V. Description of the invention (the trenches) and the bullet The features include an impedance structure formed in the second layer of the semiconductor device. These and other objects, features, and benefits of the present invention will be described in detail through the following specific embodiments, with reference to the accompanying drawings. The present invention will be explained in detail using the following description of the preferred specific embodiments and with reference to the following drawings, where: "Pictures" are shown to contain the bullet layer to be measured and a target Top view of the layer structure; net 2 shows a cross-sectional view of the structure of FIG. 1; FIG. 3 shows an intensity profile generated by detecting reflected light or electrons across the structure of FIG. 1; FIG. 4 shows Figures 丨 and 2 show another diagram of the intensity profile from θ, 口, 口, 口, 口, and 口, which show the characteristics according to the present invention. Figure 5 of the uncertainty of the tentacle projectile is shown as The trench junction shown in Figure 丄 and Figure 2? Another picture of β 5 to degrees, which shows the sign according to the present invention; Figure 6 shows the uncertainty of the nature of the eyes, and it shows a g-shirt with a g-shirt, used in conjunction with the present invention. 7 shows a method for determining uncertainty according to the present invention, A Θ, and a detailed flow chart of various embodiments. The present invention relates to detection and measurement, and more specifically, the measurement uncertainty covered in the conductor manufacturing process. This ache about the decision in the semi-quotation of the wrong explanation example This paper size is suitable for financial purposes _ Tsuk Standard (CNS) A4 specification (_2lGx297 public director) Binding outside 4519 A7 B7 5. Description of the invention (7 to describe. Figure 1 and The structure of 2 and the intensity profile of FIG. 3 will be used to explain in detail the method of calculating the coverage uncertainty according to the present invention. The present invention provides a method that can be used in conventional detection systems. The present invention provides at least one bullet (or target) Two redundant measurements between the edge pairings of s and an average centerline of the edge pairing of the target (or bullet) on the intensity profile to determine the overall misregistration between the stacks of the semiconductor process or inspection Value. Please refer to the specific details in the drawings, where the same reference numbers represent similar or identical components in several drawings. Figure 4 is a scanning transimpedance structure (bullet) 12 and trench structure of Figures 1 and 2 (Target) The intensity profile of 16. The edge slope curve 1 0 2 shows the change in the slope curve along each edge. The edge slope curve represents the change in intensity due to surface features. In another embodiment, 'atomic force microscope' A deviation profile is provided, which can be used in the present invention. The intensity change is relative to a reference line or a threshold intensity of 1.06. The threshold intensity can be selected to be a set value that takes a large intensity, such as the maximum output from an energy source 90% of the intensity. The intensity profile is developed between a data base (y-coordinate equal to 0) and the threshold value, which is selected according to predetermined conditions. The edge slope curve 102 represents the edge of the bullet 12 and / or the structure of the target 16 (Figure 1). The edge pairing is marked with lowercase and uppercase letters. For example, the edge pairing of target 16 is d and D, and the edge pairing of target 16 is c and C. The edge pair of each bullet 12 includes a, A and b, B. Edge pairing includes edges with symmetrically opposite edges on the same stack, that is, bullet layers, such as a, A. The point 1 0 8 on each edge slope curve 102 represents a point with a maximum slope, which is a paper size that applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) 494519 A7 B7

V. Description of the invention (8 Determined by the coverage measurement algorithm. Each point 108 represents the position of the individual edge of the bullet 2 and the target 16 ', that is, the edge of the mark. The coordinate value of the point 108 can be used to explain according to the present invention The calculation of misregistration. Other methods that represent edges in a strong production profile can also be used. According to the present invention, the intensity profile in Figure 4 is used to determine the uncertainty of the bullet ... target centerline (target CL), TcL ' Is calculated by averaging the target edge pair. In this example, this represents the average distance between cc and d D. For the present invention, the distance along an χ axis between an edge pair is labeled aA, bB, cC, and / or d D. For example, the distance between edge d and edge D can be written as d D. The position of an average target centerline τ c L can be written as:

Tcl = [cC / 2 + dD / 2] / 2 This Tcl returns a value on the x-axis, representing the configuration of the average target centerline relative to a reference datum. If more edge matching is needed, a general formula can be used for TcL as follows: where N is the total number of target edge pairs, and L is the χ-axis distance between the edges of each edge pair. An average distance is calculated for each bullet edge pairing '. For example, the average distance between edge pairs b and B is b B / 2 and a and A are a A / 2. These average distances are used to calculate redundant measurements. Because there are two edge pairs for this bullet ', it uses two redundant measurements. Redundant measurement 1 = aA / 2-Ta, and redundant measurement 2 = bB / 2-Ta. These several measurements f represent the X-axis numerical center of the edge pairing of the parent bullets. "11. This paper size applies the Chinese National Standard (CNS) A4 specification (210X 297 mm) 494519 A7 B7. V. The invention description line and the average Difference between target edge pairing centerlines. This redundant measurement is then used to determine the bullet uncertainty according to the present invention. The bullet uncertainty according to the present invention can be calculated using the following formula: (redundant measurement 1-redundant measurement 2) / 2. Please refer to FIG. 5. The intensity profile is used to determine the target uncertainty. A bullet center line B C L is calculated by averaging the bullet edge pairs. In this example, this represents the average between a, A and b, B. In this disclosure, the distance along the X axis between edge pairs is labeled aA, bB, cC, and / or d D, as described above. The position of an average bullet center line B c l can be written as:

Ba = [aA / 2 + bB / 2] / 2 The B G L returns the value on the x axis of the center line of the bullet relative to a reference datum. If more edge matching is needed, a general formula of B C L can be used 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 calculated. For example, the average distance between edge pairs c, C is c C / 2 and for d, D, it is d D / 2. These average distances can be used to calculate redundant measurements. Because there are two edge pairs for this target, it uses two redundant measurements. Redundant measurement 3 = βα-cC / 2, and redundant measurement 4 = βα-dD / 2. These redundant measurements represent the difference between the centerline of the X-axis numerical pairing of the edge pairing of each target and the centerline pairing of the average bullet edge pairing. This redundant measurement is then used to determine the target uncertainty according to the present invention. The target is not correct. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (21 × 297 mm) 494519 A7 B7 V. Description of the invention (10 Qualitative According to the present invention, the following formula can be used to calculate: (Redundant measurement. Measure 3-Redundant measurement 4) / 2. By calculating the coverage uncertainty of the bullet and the target in this way, the present invention reduces the effect of the coverage measurement point of the bad shell. Preferably, the present invention uses the average value to improve Covers uncertainties and provides low sensitivity tracking of batch-to-batch variation in measuring semiconductor devices or other measurements. This uncertainty will make it possible to identify and control processes that help mark degradation Condition (ie, degradation at the appropriate point). Refer to Figure .6, a measuring device 200 includes a microscope 200, such as a tracing electron microscope, an optical microscope, or a Atomic force microscope of a structure or sample 206 to be measured on platform 2 04. — Source 2 0 8 illuminates the structure 2 06. A light-sensing device or electronically sensitive device 209 collects reflections. Strength and store In addition, the reflection (not shown) of the stylus can be used to develop the intensity wheel ^ The reflection data stored in the storage device 210.-The processor 212 is used to perform data calculation according to the present invention.-Monitor 2 4 It can also include a real-time viewing of the structure during operation. In the preferred method, the edge pairing can be optimized by the tool adjusting the measuring device. For example, the signal comparison can be performed by scanning through the focus. For This intensity profile can get more symmetrical signals. For example, if a wafer is placed on the platform 2 0 4 and moved out of the plane determined by the focal point of the microscope 2 0, T ", out of plane In this way, a better surface feature focus can be obtained. This can be achieved when the flattening can be scanned in the z direction. This method is recursive ^. Its generation ~ This scale is applicable to the Chinese country Standard (CNS) A4 specification (210X 297 mm) Binding-13- A7 __— _ B7 V. Description of the invention (11) Measurement results. Then perform a focus control 216 adjustment (Figure 6) to change the microscope 2 G.2 Focus. It produces a second measurement knot The quality is evaluated relative to the first measurement result. If there is improvement, perform further recursive steps to further improve the measurement result. If there is no improvement, perform further adjustments to improve the quality of the measurement. Please refer to ® 7, which shows the general method of the present invention. In step 302, the measurement system, system 2000 provides a sample for measurement. The sample includes a -bullet pattern and a -target pattern. In step 304, an intensity (or deviation from the contour) is obtained from the bullet and the target pattern. In step 306, the intensity contour is processed to determine the maximum tilt point to estimate the edge position of the bullet and the target. * Large slope calculation can be performed using an overlay measurement algorithm, which can be performed like Quaes terTM, provided by Bi_Rad. In step 308, the average target centerline and the average bullet centerline are determined as described above. In step 310, the distance of the edge pairing is calculated and halved. In step 3 12, the redundant measurement is calculated as described above. In step 3 1 4 'the bullet and target uncertainty are calculated. It has been described here that the best method and specific embodiment method for determining the coverage of measurement uncertainty (which is used as an illustration rather than a limitation), it should be noted that corrections and changes can be performed by a person skilled in the art in accordance with the above description. Therefore, it must be understood that changes can be made in the specific embodiments of the invention disclosed under the scope and spirit of the invention as set forth in the scope of the appended patent application. Therefore, the details and uniqueness required by the patent regulations have been proposed here, and those claimed and protected by the patent documents are proposed by the scope of the attached patent application 0 ------------- -14. This paper standard applies China National Standard (CNS) A4 (210 X 297 mm)

Claims (1)

494519 AB c D Patent application scope. A method for determining measurement uncertainty, including the following steps: Provide a strength profile covering the measured bullet and target features; decide to represent the bullet and target features on the strength profile Edge & Determine the / average center line between the edge feature of one of the bullet feature and the target feature; calculate the edge pairing distance between the bullet feature and one of the target features and the other of the bullet feature and the A redundant measurement between the average centerlines of the target features; and determining the uncertainty between two different redundant measurements. 2. The method of item 1 of the patent application range, wherein the step of determining an average centerline includes the following steps: determining the distance between each edge pairing; halving each edge pairing distance; adding the halved edge The pairing distances are used to obtain a sum; and the sum is divided by the total number of edge pairs added. 3. If the method of applying for the scope of the patent application, the method of determining the position of the edge includes the following steps: Determine the position of the maximum slope of each edge to represent the edge. 4. If the method of applying the scope of the patent application, Among them, the step of determining the average centerline includes the step of determining an average bullet centerline, and the step of calculating a redundant measurement includes the redundancy between the target feature's edge-pairing distance and the average bullet centerline. The step of determining the uncertainty of a target by the residual stiffness. / -15- This paper applies the Chinese National Standard (CNS) A4 specification (210X297 mm). The scope of patent application is A8 B8 C8 5. If the method of patent scope is declared, the decision shall be made equal. .¾ includes the step of determining an average target centerline, and the step of calculating a redundant measurement includes the step of determining the bullet uncertainty by measuring a redundant measurement between the distance and the average target centerline on one side of the bullet feature. 6. The method according to item 1 of the patent application scope, wherein the step of calculating a redundant amount includes the following steps: determining a distance between an edge pair; halving the distance of the edge pair; determining the distance of the edge pair The difference between the halving distance and the average centerline. 7. As in the method of the scope of the patent application, the step of determining an uncertainty between two redundant measurements includes the following steps: determining two redundant measurements A difference between the two; Divide the difference between the two redundant measurements by 2. 8. The method of the scope of the patent application 帛 item 'where the bullet and the intensity profile of the target feature are provided The steps include providing a step to form the target feature, and a step to form a half device on which the bullet is formed. Equalization and the fate determine the different covers between the rest of the test in its conductor 9. A decision to determine measurement uncertainty, certainty The method includes the following steps: providing-coverage measurement on the first layer on the second layer-the strength of the bullet pattern is overwhelming, the subscript; the figure L pattern includes features roughly symmetrical to-the center line; Use a single-position bullet on the intensity profile—the bullet and the mesh-16—this paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) The edge of the patent-applied landmark feature; determines an average centerline between all edge pairs of the bullet feature and all edge pairs of the target feature; and the distance of a first edge pair by taking the bullet feature / Half, subtract the average centerline of all edge pairs of the target feature 'and repeat a second edge pairing of the bullet feature to calculate two redundant measurements of the bullet feature; by taking the target feature's The distance of a first edge pairing is half, subtract the average centerline of all edge pairs of the bullet feature 'and repeat a second edge pairing of the target feature to calculate two redundant measurements of the target feature; And an uncertainty is determined according to the redundant measurement of the target feature and the redundant feature of the bullet feature. 10. The method according to item 9 of the scope of patent application, wherein the step of determining an average center line includes the following steps: determining the distance between each edge pairing; halving each edge pairing distance; adding the halved edge pairing Distance to get a sum of the bullet edge pair and a sum of the target edge pair; and divide the sum of the bullet edge pair by the number of added bullet edge pairs' and divide the sum of the target edge pair by the sum The number of target edge pairs. 11. The method according to item 9 of the scope of patent application, wherein the step of using a single position on the intensity profile to represent the bullet and target characteristics is included. -17- This paper size is applicable to China National Standard (CNS) A4 specifications (210 X 297 public love) 494519 A BCD VI. Patent Application Scope • Including the following steps: Use the maximum slope of each edge to determine the single position to represent the edge. 12. The method according to item 1 of the scope of patent application, wherein the step of determining an uncertainty according to the redundant measurement of the target characteristic and the redundant measurement of the bullet characteristic includes the following steps: determining the redundant characteristic of the target characteristic A difference between the redundant measurement and the redundant measurement of the bullet characteristic; dividing the difference between the redundant measurement of the target characteristic and the redundant measurement of the bullet characteristic by 2 to determine a target uncertainty, respectively And a bullet uncertainty. 13. The method of claim 1, wherein the target feature includes a trench formed in the first layer of a semiconductor device, and the bullet feature includes an impedance junction formed in the second layer of the semiconductor device- 18- This paper size applies to China National Standard (CNS) A4 (210X 297mm)