JP3316977B2 - Pattern inspection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for automatically inspecting a pattern formed by a film carrier or the like, and more particularly to an inspection method for judging a defective state in a defective portion of a defective product. is there.

[0002]

2. Description of the Related Art Conventionally, a film carrier used for mounting ICs and LSIs is a polyimide film having a thickness of about 75 to 125 μm, a copper foil is adhered with an adhesive, a photoresist is applied to both sides, and a mask is applied. Exposure, development and etching are performed to form a lead pattern.

After the pattern is formed in this manner, the photoresist is removed, and the surface of the lead is plated with Sn, Au, and solder to complete the film carrier process.
After the completion of this step, the pattern is visually inspected by a human using a microscope. Thus, the fine pattern
Inspection of the eye visually requires skill and results in the abuse of eyes.

On the other hand, as an alternative to the visual inspection, a pattern matching method is used in which a pattern is imaged by a TV camera and an inspection is performed as a reference pattern based on the coincidence rate with the pattern to be measured. It is also possible. However,
As shown by the following equation, the coincidence rate is a measurement in pixel units, and is not suitable for inspection of a fine pattern such as a film carrier. Matching rate = {(whole pixel−unmatched pixel) / (whole pixel)} × 100% Also, in the case of a pattern in which leads of the same shape are continuous, the lead is In some cases, the positions may be shifted by one line, and there is a problem in the case of repeating patterns of the same shape.

Therefore, in order to solve the above problems,
The inventor has previously filed the following method. That is, the pattern to be measured which is determined to be non-defective is defined as a master pattern, and its grayscale image is stored in an image memory, and is binarized from the grayscale histogram of the master pattern stored in this image memory. processed to Masutapata - down of Ejjide - seeking data, the Ejjide - and registers was linearized by the least square method from the data, seeking and centerline L _M and width W of the two straight lines that are located opposite each other This is registered as master pattern information, and this is used as a reference pattern. On the other hand, the edge data of the pattern to be measured is associated with each straight line of the master pattern, and the two are compared and collated.
This is a method in which errors are inspected and non-defective and defective products are determined.

[0006]

However, according to the above method, the edge data obtained along the pattern to be measured and the data of the master pattern are compared to determine a good or defective product. Therefore, if the entire pattern is misaligned, the obtained edge data matches the data of the master pattern, and it is erroneously recognized that there is a misalignment. There was a problem. That is, from the edge data on only one side along the pattern to be measured, it is not possible to recognize whether the defective portion is displaced, or whether the defective portion is protruding or thin, so that the defective state can be accurately determined. Cannot be determined.

[0007]

According to the present invention, a perpendicular line is dropped from the edge data of a pattern to be measured to a straight line of a master pattern corresponding thereto, and the length Δd to the perpendicular line is defined as an error amount. when the error amount Δd is large, Masutapata - measured at the center line L _M of the emission pattern - emissions of opposing Ejjide - Lower the perpendicular from data, the mutually facing Ejjide - from data to the center line L _M By obtaining the distance, the width w of the edge data at the defective part is obtained, and the edge data at the defective part and the edge data of the master pattern are obtained.
The length X of the defective portion of the pattern to be measured is determined from the data, and the length X of the defective portion is used as a criterion for determining the length X of X> kW or X <kW. When the width w of the pattern is determined to be yW> w or w <zW as a criterion for determining the width w of the defective portion, k, y, and z are initially set and the defective state at the defective portion of the pattern to be measured is set. Is determined.

[0008]

The pattern to be measured which is determined to be non-defective is defined as a master pattern, its grayscale image is stored in an image memory, and a binary histogram is obtained from the grayscale histogram of the master pattern stored in this image memory. The edge data of the master pattern is obtained by performing the conversion process, the master pattern is converted into a set of straight lines based on the edge data and registered as a straight line, and the center line L _{M of} the two straight lines facing each other is registered.
And the width W are obtained and registered as master pattern information, and this is used as a reference pattern.

The width w of the edge data at the defective portion is determined from the edge data of the pattern to be measured, and the defect of the pattern to be measured is determined from the edge data at the defective portion and the edge data of the master pattern. The length X of the defective portion is determined, and the length X of the defective portion is determined using X> kW or X <kW as a criterion for determining the length X of the defective portion.
The width w of the defective portion is determined as yW> w or w <zW as a criterion for determining the width w of the defective portion.

The values of k, y, and z are initialized, and a defective state at a defective portion is determined based on the width W of the master pattern and the width of the pattern to be measured.
Even when the defective part is displaced, the position can be corrected and the defective state can be inspected.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail with reference to FIGS. FIG. 1 is a processing flow showing an embodiment of the present invention, and FIGS. 2 to 4 are processing flows of a master pattern and a basic operation diagram of a pattern inspection apparatus showing an embodiment of the present invention, and a system configuration. FIGS. 5 to 13 are explanatory views for explaining a pattern inspection method according to the embodiment of the present invention, and FIGS. 14 to 18 are explanatory views showing an embodiment of the invention.

First, a specific pattern embodying the present invention will be described.
The basic operation diagram shown in FIG. 3 and FIG.
A description will be given based on the system configuration diagram shown in FIG. In this embodiment, a case where a pattern of a TAB tape is used as an example of the pattern to be measured m and the pattern is inspected will be described.

In FIG. 3, a pattern inspection device is
Unwinding unit 1 for feeding out a TAB tape (not shown) film wound on a table 1, fixed-size feeding / positioning unit 2, detecting unit 3 for judging non-defective products, and 3 A feeder for positioning and feeding the tape to the next cutting step; a punching section for cutting the tape of defective products; and a winding section 6 for winding the tape on the reel again. ing.

FIG. 4 shows a system configuration diagram. A system control unit 7 controls the entire system, and
And CRT8 for displaying results and product types, setting, registration,
A keyboard 9 for selecting an operation menu, a sequencer unit 10, a printer 11, an image memory 12, a CPU (measurement unit) 13, a signal selector 14, and the like are controlled.

The sequencer unit 10 controls an operation switch / display lamp 15, a taper unit 16, a tape puncher unit 17, an XY table 18, and the like under the control of the system control unit 7. are doing. 19 is a floppy disk.

The motor 20 in the X direction is
A table controller (3) via the sequencer unit 10
Axis) 22 and the Y-S-P driver 23 control the motor 24 in the Y direction and the motor 25 for the span axis.
The drive of the -Y table 18 is controlled in the X direction, the Y direction, and the span axis direction, respectively. Therefore, the TAB tape is picked up by the cameras 26 and 27 on the XY table 18 and the image is digitally converted by the A / D converter 28 and stored in the image memory 12 in pixel units. Detector 3
Are the cameras 26 and 27 and the moving mechanism, XY
And an A / D converter 28, and the image memory 12.

Next, the operation of the pattern inspection apparatus will be described. TAB tape wound on reel
Under the control of the sequencer unit 10, the camera 2 on the XY table 18 is transferred to the tape planner unit 16 frame by frame.
It is sent out to predetermined positions of 6, 27 and positioned. X-
The Y table 18 includes drive motors 24 in the Y and X axis directions,
21, the positioning is performed by controlling the three axial directions by the span axial direction motor 25.

One frame of TAB positioned at a predetermined position
The pattern of the tape is picked up by two cameras 26 and 27, and the grayscale image is converted into a digital signal by an A / D converter 28 and stored in the image memory 12 in pixel units. On the other hand, in the image memory 12, at the start of the inspection, master pattern information of the TAB tape determined to be non-defective is created and stored. This master pattern M and TA
The B pattern (measured pattern m) is compared in the detection unit 3 by a method described later, and a process for determining a good product or a defective product is performed.

The TAB tape which has been subjected to the process of judging a non-defective product or a defective product for each frame in this way is sequentially sent to the tape puncher unit 17 for each frame, and after the defective product is punched out, The take-up unit 6 takes up the reel again to complete the inspection.

Next, a method of inspecting the TAB pattern so that the detection unit 3 determines that the TAB pattern is a good product or a defective product will be described with reference to FIG. Prior to the inspection, a master pattern M serving as a reference must be created from a TAB tape determined to be non-defective, and registered in the image memory 12 as master pattern information.

Hereinafter, a method of creating the master pattern M will be described with reference to FIG. Black and white cameras 26, 27
5A and 5B show a part of the pattern, and FIG. 6 and FIG. 7 are enlarged views of the gray-scale image of the pattern of the TAB tape determined as non-defective. However, it is first registered once in the image memory 12 (step 4
0). Then, the edge coordinates (Ejjide - data) is 5, as shown in FIG. 6, the grayscale image is white → black, at the midpoint of points has changed black → white, and the threshold value S _L It is extracted assuming that they intersect (step 41).

[0022] Masutapata - Ejjide of emissions M - the data extraction processes binarized it from the density histogram of the entire image, white on both sides of the threshold S _L, prorated in black pixels (i.e., the midpoint ), And the threshold value S _L
X or Y coordinate value (hereinafter referred to as edge data)
N ₁ (4.5, ₂ ), N ₂ (5, 2.5), N ₃ (5.
5, 3) are obtained (step 41).

Next, as shown in FIG. 8, each of the edge data N ₁ , N ₂ , N ₃ ... Of the master pattern M, which is a set of points obtained in this manner, is calculated by the least square method. Are linearized as straight lines A ₁ , A ₂ , A _3.
2). In this case, the edge data N ₁ , N ₂ , N
₃ ... beat perpendicular to a straight line A _1, A _2, A _3, ... from each respective Ejjide - data N _1, N _2, N linear A ₁ from ₃ .... Distance to A ₂ ··· ΔN ₁ , ΔN ₂ , ΔN
₃ ... are, in the case of 0.3 pixel or less, this linearized possible Ejjide - included in the straight line A _1, A _2, ... as data. In this way, Masutapata are extracted as a set of points - Ejjide of emissions M - data _{N 1, N 2, N 3} ·
.. Are linearized, and finally the master pattern M is converted into a set of straight lines A ₁ , A ₂ , A _3.
2).

Next, the width W of the master pattern M is determined. As shown in FIG. 9, obtains a center line L _M from the straight line A ₁ and the straight line A _n, a straight line A ₂ and the line A _n-1 ··· which face each other, the center line de - data is Masutapata - down information (Step 43). This information is used to check the direction of the lead when checking the pattern to be measured m.

When a straight line H _M perpendicular to the obtained center line L _M is drawn, the straight line H _M becomes straight lines A ₁ and A _N , A ₂ and A _N-1.
Are equal to the width W of the master pattern M. That is, the center line L _M both linear A ₁ and A
The trajectory becomes the center locus of the inscribed circle whose diameter is the distance from _n-1 and the diameter becomes the width W of the master pattern M (step 44). The master pattern obtained sequentially in this way
The center lines L M of the straight lines A ₁ , A ₂ , A ₃ ... Of the pattern _M are registered as master pattern information, and a master pattern _M as a reference is created (step 45).

When the master pattern M is created in the procedure shown in FIG. 2 in the manner described above, the next step is to actually execute the TAB pattern.
The pattern m to be measured is inspected according to the procedure shown in FIG. As shown in FIGS. 3 and 4, the TAB tape is wound around a reel, sent out from the unwinding unit 1 frame by frame, and the pattern to be measured m is measured by the cameras 26 and 27. It is imaged. At this time, the captured image becomes a grayscale image as shown in FIG. 5, is digitized by the A / D converter 28, and is temporarily stored in the image memory 12. Image memory 12
The master pattern M stored in the memory and the measured pattern
The pattern m is read out by the CPU 13 of the detection section 3, and after alignment, comparison and collation are performed, and the pattern to be measured m is inspected (step 46).

Here, the pattern to be measured m is defined as a master pattern.
Figure 1 shows the prerequisites for comparing and checking with
Description will be made based on 0. (1) The edge data n ₁ , n ₂ , n ₃ ... Of the master pattern M and the pattern to be measured m are masked by the window 30 as one method for setting an inspection range, for example. Only patterns in 30 are to be inspected (step 47). (2) The pattern m to be measured partially cut out by the window 30 must be the window 30 as shown in FIG.
And the pattern portion having the shape shown by pattern 31 is not recognized as edge data. (3) The inspection always proceeds from the starting point S around the four sides of the window 30 in a certain direction, and finally returns to the starting point S, completing the screen inspection.

Under such preconditions, the pattern to be measured m is inspected in the following procedure. FIG. 11 shows the state at the time of inspection, and the edge data of the master pattern M is shown.
Are linearized by the least squares method, and the straight lines A ₁ , A ₂ ,
A ₃ are indicated by.., Extracted the measured pattern - Ejjide of emissions m - data _{n 1, n 2, n 3} ··· are indicated by black dots (step 48).

Therefore, first, it is necessary to make correspondence between the edge data n ₁ , n ₂ , n ₃ ... Of the pattern to be measured m and the straight line of the master pattern M. . In order to make this association, as shown in FIGS. 11 and 12, the straight lines A ₁ adjacent to each other in the master pattern M are adjacent to each other.
And the straight line A ₂ , and the angle a, which the straight line A ₂ and the straight line A ₃ have,
Calculate straight lines A ₂ ′, A ₃ ′.

Therefore, the master pattern M is represented by the straight line A
₂ ′, A ₃ ′..., And each straight line A ₁ , A ₂ , A _3.
Area (hereinafter, areas A ₁ , A ₂ , A ₃ ...)
), The edge data n ₁ , n ₂ , n ₃ ... Of the pattern to be measured m all correspond to the regions A ₁ , A ₂ , A _3. (Step 49).

Next, the amount of error (difference) Δd between the master pattern M and the pattern to be measured m must be measured. For this purpose, as shown in FIG.
From the edge data n ₁ , n ₂ , n _3.
1 , A ₂ ... Are perpendicular to h ₁ , h ₂ , h _3.
The down distance [Delta] d ([Delta] d ₁ this perpendicular line _{h 1, h 2, h 3} ··· is, crossing the straight line A _1, A ₂ · · · respectively,
When determining the [Delta] d ₂ · · ·), the amount of this distance [Delta] d is an error, i.e., Masutapata - emission M and the measured pattern - a shift amount between the emission m (step 50).

Next, the case where various kinds of inspections are actually performed will be described. When performing an actual inspection, the measured pattern
The width w of the reference m is determined, and this width w is used as a reference master pattern.
The CPU 13 compares the width W of the pattern M with the width W of the non-defective product. At this time, X = kW as a criterion for judging the length X of the defective part as a criterion value for judging a good part and a defective part, and the width w of the pattern to be measured is yW> w or w <zW, CP as a criterion for w
U13 is set as an initial condition (step 5).
1). In the case of this embodiment, k = 2, y = ２, z
= 4/3 is initially set.

Therefore, generally, the judgment coefficients are defined as y and z.
When y <z, the width w of the pattern to be measured m and the width W of the master pattern M satisfy yW <w <zW. When it is outside, the pattern width is determined to be inappropriate and it is determined to be defective.

First, the case where the width w of the pattern to be measured m is determined will be described. FIG. 14 shows the pattern to be measured.
13 shows a state in which the width w of the pattern m is obtained. As shown in FIG. 13, the edge data n ₁ , n ₂ , n
_{From 3} ..., the straight lines A ₁ , A ₂ ... of the master pattern M
・ A vertical line is drawn, and the length to this vertical line Δd (error amount)
(Δd ₁ , Δd ₂ ...) Are obtained, and if the error amount Δd is small, it is determined as a good product as it is.

Edge data n (n
_1, n _2, for the · · n _n), Masutapata - Lower the respective perpendicular to the center line L _M of the emission M, the distance determined, the measured pattern opposing - the emissions m Ejjide -
N (n ₁ , n ₂ ,... N _n ) to width w (w ₁ , w _2.
Nn ) is required.

The obtained width w of the pattern to be measured m
Is compared and calculated by the CPU 13 with the width W of the master pattern M. Initial conditions are set in the CPU 13, and in this embodiment, when the width of the master pattern M is W, a reference value for determining the width w of the pattern to be measured is set. That is, it is determined whether or not it is within the range of (2/3) W <w <(4/3) W (step 51),
If it is within this range, it is determined that it is a good product, and if it is out of the range, it is determined that it is a defective product corresponding to any of thin, thick, protrusion, or chipped.

Next, a description will be given of a case where a defective state is actually determined when a defective product is determined. (1). Inspection of “thinning” FIG. 15 shows that the pattern to be measured m determined to be defective is
This indicates a state at the time of inspection determined to be "thin", and the error amount Δd of the edge data n of the measured pattern m obtained as shown in FIG. and it has Ejjide and - starts checking the data n _1, Ejjide - data _{n 1, n 2, n 3} ··· and the check and final Ejjide - the deviation check to data n _n is occurring Find the length X.

Next, as described above, the master pattern M
Lower the perpendicular respectively to the center line L _M of the distance determined, the measured pattern opposing - the emissions m Ejjide - data _{n (n 1, n 2,} ·· n n) the width w (w ₁ , w ₂
.. _Wn ) are required.

The length X and width w (w ₁ , w ₂ ... W _n ) of the displacement of the measured pattern m thus obtained.
Is compared with the data of the master pattern M in the CPU 13. As a result, assuming that the width W of the master pattern M and the length X and the width w of the deviation of the pattern to be measured m, the deviation length X> 2W (step 52) and the width w <2W / 3.
Only at the time (step 53), it is determined to be "thin" and is determined to be defective (step 54).

(2). FIG. 16 shows that the measured pattern m determined to be defective is
This indicates the state at the time of the inspection determined to be "fat". In the same manner as in the inspection of the thinning, the error amount Δd of the edge data n of the pattern to be measured m is shifted by one pixel or more. and are Ejjide - to start the check from the data n _1, Ejjide -
Data _{n 1, n 2, n 3} ··· and the check and final Ejjide - checks to data n _n by determining the length X of the deviation occurs.

Next, as described above, the master pattern M
Lower the perpendicular respectively to the center line L _M of the distance determined, the measured pattern opposing - the emissions m Ejjide - data _{n (n 1, n 2,} ·· n n) the width w (w ₁ , w ₂
.. _Wn ) are required.

The length X and width w (w ₁ , w ₂ ... W _n ) of the displacement of the pattern to be measured m obtained in this manner.
Is compared with the data of the master pattern M in the CPU 13. As a result, when the width W of the master pattern M and the length X and the width w of the deviation of the pattern to be measured m, the deviation length X> 2W (step 52) and the width w> 4W / 3.
Only at the time (step 55), it is determined to be "fat" and is determined to be defective (step 56).

(3). FIG. 17 shows that the measured pattern m determined to be defective is
This indicates the state at the time of the inspection that is determined to be "chipped", and the pattern to be measured m
The length X at which the displacement occurs is determined.

Next, in the same manner as above, the pattern to be measured is
Of emissions m Ejjide - data _{n (n 1, n 2,} ·· n n) the width _{w (w 1, w 2 ··} w n) is obtained.

The length X and width w (w ₁ , w ₂ ... W _n ) of the displacement of the measured pattern m thus obtained.
Is compared with the data of the master pattern M in the CPU 13. As a result, assuming that the width W of the master pattern M and the length X and the width w of the deviation of the pattern to be measured m, the deviation length X <2W (step 57) and the width w <2W / 3.
Only in the case of (step 58), it is determined that the chip is "missing" and is determined to be defective (step 59).

(4). Inspection of “protrusion” FIG. 18 shows that the measured pattern m determined to be defective is
This indicates the state at the time of inspection determined to be "projection", and the length X at which the pattern to be measured m is displaced is obtained in the same manner as in the inspection in the above (1) to (3). .

Next, in the same manner as above, the pattern to be measured is
Of emissions m Ejjide - data _{n (n 1, n 2,} ·· n n) the width _{w (w 1, w 2 ··} w n) is obtained.

The length X and width w (w ₁ , w ₂ ... W _n ) of the displacement of the measured pattern m thus obtained.
Is the data of the master pattern M in the CPU 13 as well.
Is compared with the data. As a result, when the width W of the master pattern M and the length X and the width w of the deviation of the pattern to be measured m,
Shift length X <2W (step 57) and width w> 4
Only at the time of W / 3 (step 60), it is determined to be "projection" and to be defective (step 61).

[0049]

According to the present invention, a perpendicular line is dropped from the edge data of the pattern to be measured to a straight line of the master pattern corresponding thereto, and the length Δd to this perpendicular line is defined as an error amount. Is large, a perpendicular is dropped from the opposing edge data of the pattern to be measured to the center line LM of the master pattern, and the center line L is moved from the opposing edge data to the center line L _M.
By obtaining the distance to _M , the width w of the edge data at the defective portion is obtained, and the length X of the defective portion of the measured pattern is obtained from the edge data at the defective portion and the edge data of the master pattern. The length X of the defective portion is determined by using X> kW or X <kW as a criterion for determining the length X of the defective portion, and the width w of the pattern to be measured is defined as y
By using W> w or w <zW as a criterion for determining the width w of the defective portion, k, y, and z are initialized to determine a defective state in the defective portion of the pattern to be measured. -The width of the line is used as the inspection standard.
Even if the pattern to be measured is displaced as a whole, the inspection can be performed with the position corrected, and the defective state of the defective part in any defective shape can be inspected accurately.

In particular, as in the prior art, when judging only from one side edge data obtained along a pattern, the projection, thinning, fattening, chipping, etc. must be uniquely defined. Was impossible, but we can define these states exactly.

Since the inspection conditions can be arbitrarily set as the initial conditions, the inspection conditions can be changed as necessary, and the defect states of the defective portions of all the defective products are judged under the same inspection conditions. And accurate test results can be obtained.

[Brief description of the drawings]

FIG. 1 is a processing flow showing an embodiment of the present invention.

FIG. 2 is a processing flow of a master pattern.

FIG. 3 is a basic operation diagram of the pattern inspection apparatus.

FIG. 4 is a system configuration diagram.

FIG. 5 is a part of a pattern to be measured.

FIG. 6 is an explanatory view for obtaining edge data of a master pattern, and is an enlarged view of FIG. 5;

FIG. 7 is an enlarged view of FIG. 6;

FIG. 8 is an explanatory diagram for linearizing a master pattern.

FIG. 9 is an explanatory diagram for obtaining a width W of a master pattern.

FIG. 10 is a pattern to be measured displayed in a window.

FIG. 11 is a diagram showing a relationship between a straight line of a master pattern and edge data of a pattern to be measured.

FIG. 12 is an explanatory diagram for dividing an area of a master pattern.

FIG. 13 is an explanatory diagram for determining a shift (error amount).

FIG. 14 shows an embodiment of the present invention and is an explanatory view of a pattern width inspection.

FIG. 15 shows an embodiment of the present invention and is an explanatory view of thinning inspection.

FIG. 16 shows an embodiment of the present invention and is an explanatory diagram of an inspection for fattening.

FIG. 17 shows an embodiment of the present invention and is an explanatory diagram of chipping inspection.

FIG. 18 shows an embodiment of the present invention and is an explanatory view of inspection of a projection.

[EXPLANATION OF SYMBOLS] M Masutapata - emission A _1, A ₂ ··· Masutapata - linear N ₁ of emissions M, N ₂ · · · Masutapata - down of Ejjide - motor H _M Masutapata - down of the straight line L _M Masutapata - down center line W Masutapata of - width Δd error amount of emissions (deviation) m to be measured patterns - down w measured pattern - emission width n _1, n ₂ ··· measured pattern of - emissions of Ejjide - the data X displacement length 12 Image memory 13 CPU

──────────────────────────────────────────────────の Continued on the front page (58) Fields investigated (Int.Cl. ⁷ , DB name) G06T ⁷ /00-7/60 G06T 1/00 H01L 21/66 G01B 11/00-11/30 G01N 21 / 84-21/958

Claims (5)

(57) [Claims] 1. A non-defective pattern to be measured is stored in an image memory as a gray image of a master pattern, and is binarized from a gray histogram of the master pattern stored in the image memory. Masutapata - down of Ejjide - seeking data, the Ejjide - the method of least squares from the data Masutapata - converts in to a set of straight lines, find the center line L _M of both straight lines lying opposite each other from these lines The center line L is registered as the master pattern information.
To determine the width W of the straight lines facing each other from the length to the point where the straight line intersects at right angles with _M and the two straight lines and register the width W as the master pattern information, to compare with the pattern to be measured. The master pattern which is a reference of the above is prepared, the pattern to be measured is stored in an image memory, the master pattern and the pattern to be measured are aligned, and an inspection range is set. The edge data of the pattern to be measured exists around the inspection range, and the periphery of the inspection range is inspected in a certain direction to check the pattern to be measured.
The edge data of the master pattern is extracted, and a straight line that bisects the angle of the straight lines adjacent to each other in the master pattern is calculated.
The straight line divides the master pattern into regions respectively corresponding to the straight lines, and associates the edge data of the pattern to be measured present in each region with the straight line of the master pattern. In the pattern inspection method for detecting an error between the associated edge data of the pattern to be measured and the edge data of the master pattern to determine a non-defective product or a defective product, A perpendicular line is dropped from the edge data of the pattern to the straight line of the master pattern associated with the pattern, and a length Δd to the perpendicular line is defined as an error amount. When the error amount Δd is large, the center of the master pattern is determined. Ejjide opposing the emission - - the measured pattern line L _M down the perpendicular from data, Ejjide this opposing - by obtaining the distance from the other to the center line L _M, defect sites The width w of the edge data at the defective portion is determined, and the length X of the defective portion of the pattern to be measured is determined from the edge data at the defective portion and the edge data of the master pattern. X is X>
kW or X <kW is used as a criterion for determining the length X of the defective portion, and when the width w of the pattern to be measured is yW
By setting> w or w <zW as a criterion for determining the width w of the defective portion, k, y, and z are initially set to determine a defective state in the defective portion of the pattern to be measured. Inspection method of pattern. 2. Initially setting k = 2 and y = ２, the length of the defective portion is X> 2W and w <2.
2. The pattern inspection method according to claim 1, wherein when the value becomes 2W / 3, the defective state in the defective portion of the pattern to be measured is determined to be thin. 3. Initially setting k = 2 and y = 2/3, the length X of the defective portion is smaller than 2W, and w <
2. The pattern inspection method according to claim 1, wherein when the value becomes 2W / 3, the defective state in the defective portion of the pattern to be measured is determined to be missing. 4. Initially setting k = 2 and z = 4/3, the length X of the defective portion is greater than 2W and w>
2. The pattern inspection method according to claim 1, wherein when the value becomes 4W / 3, the failure state at the defective portion of the pattern to be measured is determined to be fat. 5. Initially setting k = 2 and z = 4/3, the length X of the defective portion is X <2W and w>
2. The pattern inspection method according to claim 1, wherein when the value becomes 4W / 3, the defective state at the defective portion of the pattern to be measured is determined to be a protrusion.