JP2005302890A - Alignment-mark search device and search technique - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem, wherein the moving distance of a search becomes long according to the order of the search, and not only a long time is required for the search but also dust is generated from the device, in response to the moving distance and the defective generation ratio of products is increased, to improve the efficiency of the search, and to minimize the moving distance. <P>SOLUTION: An alignment-mark search device has an imaging section 3 sensing an image in the region of a specified area on a sample, a moving section successively changing an imaging region by relatively moving the sample and the imaging section, and a first storage 4, in which the order of an imaging is stored previously regarding a plurality of the imaging regions on the sample. The search device, further, has an imaging control unit for making the sample imaged, in the order of imaging, by driving the imaging section and the moving section, a detector 6 detecting an alignment mark from the sensed image, and a second storage in which the presence region of the alignment mark is stored. The search device has an arithmetic operation section 8 for computing the detection frequency of past alignment marks in each imaging region and the moving distance required by the moving section from the storage contents of the second storage and an imaging order decision 9 determining the new order of the imaging, on the basis of the result of the computation and updating the order of the imaging of the first storage into the new order of the imaging. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a search device and a search method for detecting an alignment mark on a sample such as a wafer or a glass substrate in manufacturing a semiconductor or a liquid crystal display.

  In the manufacture of conventional semiconductors, liquid crystal displays, and the like, alignment marks are drawn on a substrate, and alignment (positioning) of the substrate is performed using the alignment marks during each process. When detecting an alignment mark, an imaging device that can capture a fixed area, such as an area sensor camera or a line sensor camera, is used. However, the imaging area is limited by the resolution for detecting the alignment mark, resulting in higher accuracy. In recent years, the imaging area tends to be narrowed. Therefore, when the alignment mark cannot be detected at the target position, the substrate or the imaging system is moved to move the imaging area, and the alignment mark is searched. At this time, the method of sequentially moving the imaging region in accordance with a predetermined order to search for the alignment mark may be inefficient, and it takes time to perform the alignment process. In view of this, there has been known a technique in which an area where the alignment mark is frequently detected is preferentially searched based on the past detection history so as to shorten the search time (see, for example, Patent Document 1).

In the manufacture of semiconductors, liquid crystal displays, and the like, with the miniaturization of wiring and the increase in size of chip panels, the demand for prevention of foreign object defects has increased, and there has been a demand for cleanliness during the manufacturing process and inspection. Accompanying this is the relative movement of processing means such as exposure beams and cleaning nozzles, inspection means such as probes and imaging cameras, and processing parts or inspection parts of wafers and substrates in both manufacturing and inspection equipment. Dust generation is a problem. For this reason, the technique which prevents the dust generation from the apparatus at the time of manufacture or an inspection by devising the mechanism of the movable part of a manufacturing apparatus or an inspection apparatus and arrangement of materials is proposed (for example, patent documents 2, (See Patent Document 3).
Japanese Patent Laid-Open No. 10-22201 JP 7-172571 A JP-A-8-323567

However, while high definition is desired, the size of the substrate is increased, and the alignment mark is often arranged near the edge of the substrate. Therefore, even if the substrate is misaligned, the alignment marks on both ends of the substrate may deviate from the imaging area that can be observed with a high-resolution imaging system, and the number of areas for searching increases. it is conceivable that. For this reason, if the search order is simply determined based only on the detection frequency, it is considered that the movement path is wasted and it takes time to detect the alignment mark.
Moreover, even if the configuration of the movable part of the manufacturing apparatus or the inspection apparatus is devised variously, although the level of dust generation is suppressed, if the distance or time for which the movable part moves is long, dust generation occurs accordingly. This is also true for the relative movement of the detection unit during alignment mark detection.
The present invention has been made in consideration of such circumstances, and only considers the time required for the mark search by considering the optimization of the movement path based on the movement distance when determining the search area order for the alignment mark detection. In addition, the present invention provides an alignment search device and a search method capable of preventing the generation of dust from the device at the time of manufacturing and inspection and reducing the product defect rate.

  The present invention provides an imaging unit that captures an image of a predetermined area on a sample, a moving unit that sequentially changes the imaging region by relatively moving the sample and the imaging unit, and imaging a plurality of imaging regions on the sample. A first storage unit that stores the sequence in advance, an imaging control unit that drives the imaging unit and the moving unit to image the sample in the imaging sequence, a detection unit that detects an alignment mark from the captured image, and an alignment mark Based on the second storage unit that stores the existence region, the calculation unit that calculates the detection frequency of the past alignment mark in each imaging region and the movement distance required by the moving unit, from the stored contents of the second storage unit, and the calculation result An alignment mark search device including an imaging order determination unit that determines a new imaging order and updates the imaging order in the first storage unit to a new imaging order is provided.

  According to the present invention, by determining the search order from both the detection frequency and the movement distance based on past alignment mark detection history information, the search time can be shortened and the travel route can be shortened. The generation of dust is suppressed, and the product defect rate can be reduced.

  An alignment mark search apparatus according to the present invention includes an imaging unit that captures an image of a region of a predetermined area on a sample, a moving unit that sequentially changes the imaging region by relatively moving the sample and the imaging unit, and a plurality of samples on the sample A first storage unit that stores an imaging sequence in advance for the imaging region, an imaging control unit that drives the imaging unit and the moving unit to image the sample in the imaging sequence, and a detection unit that detects an alignment mark from the captured image A second storage unit that stores the alignment mark existence region, and a calculation unit that calculates a past alignment mark detection frequency of each imaging region and a moving distance required by the moving unit from the storage content of the second storage unit And an alignment mark search device including an imaging order determination unit that determines a new imaging order based on the calculation result and updates the imaging order in the first storage unit to the new imaging order. To.

Examples of the sample in the present invention include, but are not limited to, a silicon substrate for semiconductor production and a glass substrate for liquid crystal display. The imaging unit is, for example, a device that two-dimensionally captures an object at a predetermined magnification and converts it into an electrical signal. For this, a combination of a lens, a CCD camera, an illumination light source, and the like is preferably used.
In addition, the moving unit moves the sample and the imaging unit relatively two-dimensionally (XY direction). For example, a fixture for fixing the imaging unit and a sample are mounted on a table. Thus, a combination with a motor-driven XY table that moves in a two-dimensional direction can be easily used.
In addition, the first and second storage units, the imaging control unit, the detection unit, the calculation unit, and the imaging order determination unit can be integrally configured by a microcomputer or a personal computer including a CPU, a ROM, and a RAM.

In the present invention, the imaging order determination unit selects an area having the shortest distance from the current imaging area as a next imaging area from a plurality of scheduled imaging areas having similar past detection frequencies, and sets the imaging order. You may decide.
In addition, the imaging order determination unit may determine an imaging scheduled area having a past detection frequency smaller than the current imaging area and the shortest distance from the current imaging area as the next imaging area.
The imaging order determining unit includes an imaging scheduled area in which the past detection frequency is smaller than the current imaging area and the difference is smaller than the first threshold, the past detection frequency is larger than the current imaging area, and When the imaging scheduled area where the difference is larger than the second threshold is mixed, the area having the highest past detection frequency may be determined as the next imaging area.

In addition, the imaging order determination unit uses the detection frequency, the region coordinates, and the detection time for each region from the past data, and when there are a plurality of regions having the same detection frequency and movement distance, the one with the latest detection time is selected. You may make it determine an imaging order by selecting the area | region to include.
The imaging order determination unit uses the detection frequency and region coordinates for each region from past data, and when there are a plurality of regions having the same detection frequency and moving distance, the total detection frequency of the peripheral regions of the candidate regions is calculated. The imaging order may be determined by selecting the most regions.

  In another aspect, the present invention provides an imaging step of capturing an image of a predetermined area on the sample, a moving step of sequentially changing the imaging region by relatively moving the sample and the imaging unit, and a plurality of steps on the sample. A first storage step for storing the imaging order in advance for the imaging region, an imaging control step for driving the imaging unit and the moving unit to image the sample in the imaging order, and a detection step for detecting an alignment mark from the captured image A second storage step for storing the alignment mark existing region, a calculation step for calculating a past alignment mark detection frequency and a moving distance required by the moving unit in each imaging region, and a new imaging based on the calculation result There is provided an alignment mark search method including an imaging order determination step of determining an order and updating an imaging order in a first storage unit to a new imaging order.

Hereinafter, the present invention will be described in detail based on embodiments shown in the drawings.
As shown in FIG. 1, the alignment mark detection apparatus includes an XY table (hereinafter referred to as a stage) 2 on which a substrate 1 is mounted, an imaging unit 3 that captures an image of a predetermined area on the substrate 1, and the substrate 1. A first storage unit 4 that stores the imaging order in advance for the plurality of imaging regions, and an imaging control unit that drives the stage 2 and the imaging unit 3 to image the substrate 1 in the imaging order stored in the first storage unit 4 5, a detection unit 6 that detects an alignment mark from the captured image, a second storage unit 7 that stores an area where the alignment mark exists, and a second after the alignment mark is detected for a plurality of substrates 1. A calculation unit 8 that calculates the detection frequency of the alignment mark of each imaging region and the moving distance required by the stage 2 from the storage content of the storage unit 7, and a new imaging order is determined based on the calculation result, and the first An imaging order determination unit 9 for updating the imaging order of 憶部 4 to a new imaging sequence.

Next, the operation of the alignment mark detection apparatus of the present invention will be described using the flowchart of FIG.
In step S1, the substrate 1 is carried onto the stage 2 by an external transfer device (not shown).
In step S <b> 2, the imaging control unit 5 reads data of the imaging order stored in the first storage unit 4, that is, the alignment mark search order.
In steps S3 and S4, the stage 2 is moved to the target imaging position according to the order data read by the imaging control unit 5 in step S2, and imaging (image capturing) is performed by the imaging unit 3.

  In step S5, the detection unit 6 detects an alignment mark from the obtained image. At this time, if an alignment mark cannot be detected in the image, the process proceeds to step S6. Here, when imaging (search) is completed for all the areas according to the order read in step S2, it is determined as a detection error, the process proceeds to step S14, the error content is stored in the second storage unit 7, and the process proceeds to step S10. move on. If there is still an area to be imaged, the process returns to step S3 to move to the next imaging area, and the steps after step S4 are executed. If an alignment mark is detected, the process proceeds to step S7.

In step S <b> 7, the alignment mark detection position is stored in the second storage unit 7.
In step S8, the substrate 1 is positioned (aligned) based on the detected alignment mark.
In step S9, necessary processing is performed on the positioned substrate 1.
In step S10, the substrate 1 is unloaded from the stage 2 by an external transfer device (not shown).
In step S <b> 11, the calculation unit 8 reads the alignment mark detection data stored in the second storage unit 7.

  In step S12, the calculation frequency of the alignment mark in each imaging region and the required moving distance are calculated from the alignment mark detection data read from the second storage device 7 by the calculation unit 8. Based on the calculation result, the imaging order determination unit 9 calculates an optimal search order (imaging order). If the calculated search order is different from the current search order read in step S2, the search order is changed (updated), the process proceeds to step S13, and the search order data is stored in the first storage unit 4. If the current search order is the same, the process ends.

  Next, the calculation process performed in step S12 will be described. Here, FIG. 3 is a diagram showing the initial state of the search order, FIG. 4 is a diagram showing the area number of the search area, FIG. 5 is a region layout diagram showing the search order in the conventional detection frequency order, and FIG. FIG. 7 is a diagram showing region coordinates, FIG. 8 is an explanatory diagram of data showing the search order in the detection frequency order corresponding to FIG. 5, and FIG. 9 is an explanatory diagram of data indicating a search order when the distance between regions is considered in the detection frequency corresponding to FIG.

  In the initial state, the search order is set so as to perform a spiral search from the central region of the entire search region in the order of numbers (1) to (25) shown in FIG. Based on the result of searching several times in accordance with this search order, the search order is determined in the order of the regions where the alignment marks are frequently detected. Here, A1 to A5, B1 to B5, C1 to C5, D1 to D5, and E1 to E5 in FIG.

When the search order is simply determined in descending order of detection frequency as in the prior art, for example, the detection frequencies for each region shown in FIG. 8 are in the order of numbers (1) to (25) shown in FIG. . At this time, after searching from C4 in (1) to B4 in (7), if an alignment mark is not detected, the process moves to E1 in (8). If no alignment mark is detected, (9) Move to C5. When moving to (10) in the order of the numbers in FIG. 5, it is assumed that each region is a square with a side length of 1, and the total moving distance up to this point is about 18.0. When all the areas are searched, the total movement distance is about 53.7.
In order to reduce the movement to such distant regions, the present invention takes into account the distance between regions when there are a plurality of regions having the same detection frequency or a difference within a certain value when determining the search order. Thus, the search order for shortening the moving distance is calculated.

Below, the rule at the time of determining the search (imaging) order by this invention is demonstrated. It should be noted that an area showing the mode value is set as an initial search (imaging) area, with those arranged in descending order of past detection frequencies as an initial state. Also, numerical values a and b are defined as parameters. Numerical values a and b indicate a threshold value of the frequency value difference, which has a frequency Nt smaller than the frequency Ns of the current position, and the difference between them is less than a, that is, Nt ≦ Ns, and Ns−Nt ≦ a ...... (1)
A is used as the first threshold value to extract the region to be Further, b is provided in order to prevent a region with a long distance and a high frequency from being left when the region selection is performed with distance priority, and has a frequency Nt larger than the frequency Ns of the current position, The difference between them is not less than b, that is, Nt> Ns, Nt−Ns ≧ b (2)
B is used as the second threshold value to extract a region to be

Rule 1: When there is no region where the difference from the frequency of the current position is a or less, that is, when there is no region satisfying the expression (1), the region showing the highest frequency is selected. (If there are multiple areas with the same frequency, select the closest one)
Rule 2: When the difference from the frequency of the current position is a or less, that is, there is at least one area that satisfies the expression (1) and there is no area that satisfies b or more, that is, the expression (2), (1 The region closest to the distance is selected from the regions satisfying the formula. (If there are multiple areas with the same distance, select the one with the highest frequency)
Rule 3: When there is a region where the difference from the frequency of the current position is a or less and there is a region of b or more, that is, there are regions that satisfy equation (1) and regions that satisfy equation (2). If they are mixed, the area showing the highest frequency is selected. (If there are multiple areas with the same frequency, select the closest one)

Rule 4: When the rules 1 to 3 are applied, if the distance and the frequency are the same, the selection is made in ascending order of the numbers in FIG. (Here, as an example, the order shown in FIG. 3 is used. However, a predetermined priority order may be used as a selection condition, or a region including the latest detection history may be prioritized. (A region with a high total value may be given priority.)
Rule 5: When rules 1 to 3 are applied, the selection is made except for the area of detection frequency 0. (Here, as an example, only the region with a detection frequency of 0 is excluded, but it is also possible to automatically determine the conditions, such as by removing the region of what percentage or less of the maximum detection frequency.)

  Next, in FIG. 8 showing the search order determined in the detection frequency order, the search order No. From (7) to No. If attention is paid to (10), the detection frequency is in the range of 13 to 15, but only (8) is in a separated position as shown in FIG. Therefore, in the invention, when changing the search order, the difference in the detection frequency to be referred to can be arbitrarily set. For example, if the parameter a is set to 0, the distance between the regions is taken into consideration only for the region having the same detection frequency. If the parameter a is set to 1, the order is determined for a region having a detection frequency difference of 1. Here, an example in which a fixed value is set is given as an example. However, a statistically calculated parameter such as an average value of detection frequencies may be automatically set.

  In this embodiment, assuming that a = 4 and b = 2, the search order is determined as follows. In this example, when the region closest to the position of B3 in the previous (6) is selected using the region coordinates in FIG. 7, B3 in (6) and B4 in (7) are (1, 0). B3 of (6) and E1 of (8) are the difference of (2,3), B3 of (6) and C5 of (9) are the difference of (2,1), B3 of (6) Since D5 in (10) is the difference between (2, 2), the region B4 in (7) is the closest region. Subsequently, the region closest to B4 in (7) is the region C5 in (9), followed by the region D5 in (10), and finally the region E1 in (8). The search order from (7) to (10) shown in FIG. Assuming that the length of one side of each region is a square, the total movement distance so far is about 14.8.

Thereafter, the procedure as described above is sequentially performed to determine the search order for the alignment marks.
When all regions are searched in this way, the result is as shown in FIG. 9, and the total movement distance is about 45.4. Compared with the case where the search order is determined in the order of simple detection frequency (FIG. 8), the distance of about 8.3 is shortened, and the moving distance is reduced by about 15.4%.

As described above, according to the alignment mark search method of the present invention, the search order is determined in consideration of not only the detection frequency but also the moving distance based on the past alignment mark detection history information, so that an efficient alignment mark search can be performed. Enable. Moreover, generation | occurrence | production of dust generation can be suppressed and the defect generation rate of a product can be reduced by shortening a moving distance.

  In manufacturing a semiconductor or a liquid crystal display, when aligning a wafer or a glass substrate, the orientation of the sample is adjusted using an alignment mark drawn in advance on the sample. Mainly in the field of flat panel displays, while higher definition is desired, the substrate size is increased, the need for searching for alignment marks is increased, and the time required for searching is required to be shortened. Therefore, the present invention is suitably used for a process that requires sample alignment.

It is a block diagram for describing an embodiment of the present invention. It is a flowchart for demonstrating embodiment of this invention. It is a figure which shows the initial state of a search order. It is a figure which shows the area number of a search area | region. It is an area | region arrangement | positioning figure which shows the search order of the conventional detection frequency order. It is an area | region arrangement | positioning figure which shows the search order at the time of considering the distance between area | regions in the detection frequency by this invention. It is a figure which shows an area | region coordinate. It is explanatory drawing of the data which shows the search order of the detection frequency order corresponding to FIG. It is explanatory drawing of the data which shows the search order at the time of considering the distance between area | regions in the detection frequency corresponding to FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Substrate 2 XY table 3 Imaging unit 4 First storage unit 5 Imaging control unit 6 Detection unit 7 Second storage unit 8 Calculation unit 9 Imaging order determination unit

Claims (5)

  An imaging unit that captures an image of a predetermined area on the sample, a moving unit that sequentially changes the imaging region by relatively moving the sample and the imaging unit, and an imaging order for a plurality of imaging regions on the sample are stored in advance A first storage unit, an imaging control unit that drives the imaging unit and the moving unit to image the sample in the imaging order, a detection unit that detects an alignment mark from the captured image, and an area where the alignment mark exists A second storage unit, a calculation unit for calculating a past alignment mark detection frequency of each imaging region and a moving distance required by the moving unit from the stored contents of the second storage unit, and a new imaging order based on the calculation result An alignment mark search device comprising an imaging order determination unit that determines the imaging order of the first storage unit and updates the imaging order to a new imaging order.   The imaging order determination unit selects an area having the shortest distance from the current imaging area as a next imaging area from a plurality of imaging scheduled areas with similar past detection frequencies, and determines the imaging order. The alignment mark search device according to 1.   The imaging order determination unit has a detection frequency difference that is smaller than the first threshold and a distance from the current imaging area is the shortest for an imaging scheduled area whose past detection frequency is smaller than the current imaging area. The alignment mark search device according to claim 1, wherein an imaging scheduled area is determined as a next imaging area.   The imaging order determining unit includes an imaging scheduled area in which the past detection frequency is smaller than the current imaging area and the difference is smaller than the first threshold, the past detection frequency is larger than the current imaging area, and The alignment mark search device according to claim 1, wherein when there is an imaging scheduled area where the difference is larger than the second threshold, the area having the highest past detection frequency is determined as the next imaging area.   An imaging process for imaging an image of a predetermined area on the sample, a moving process for sequentially changing the imaging area by relatively moving the sample and the imaging unit, and an imaging order for a plurality of imaging areas on the sample are stored in advance. A first storage step, an imaging control step of driving the imaging unit and the moving unit to image the sample in the imaging sequence, a detection step of detecting an alignment mark from the captured image, and an alignment mark existing area; A second storage step, a calculation step of calculating a past alignment mark detection frequency of each imaging region and a moving distance required by the moving unit, a new imaging order is determined based on the calculation result, and the first storage unit An alignment mark search method comprising an imaging order determination step of updating an imaging order to a new imaging order.

Images