KR102378079B1 - Substrate alignment method and apparatus, and mask aligner - Google Patents

Abstract

A substrate pre-alignment method and apparatus and photolithography tool are disclosed. A method of pre-aligning a substrate includes: selecting a reference substrate; acquiring an image of a reference substrate; acquiring an image of the substrate 1 to be pre-aligned; registering the image of the substrate to be pre-aligned (1) to the image of the reference substrate and obtaining an offset parameter with respect to the registered image of the substrate to be pre-aligned; and angularly adjusting the substrate 1 to be pre-aligned based on the offset parameter, and terminating the substrate pre-alignment process. The substrate pre-alignment apparatus includes a positioning apparatus, an image acquisition and analysis apparatus, and a control apparatus, and the photolithography tool includes such a substrate pre-alignment apparatus. The method and apparatus enable pre-alignment of substrates without the need to form specific marks, thereby achieving greater flexibility and helping to increase productivity.

Description

Substrate alignment method and apparatus, and mask aligner

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the field of integrated circuit (IC) manufacturing, and more particularly, to a substrate pre-alignment method and apparatus, and a photolithography tool.

Complex photolithography tools involve operate in coordination to precisely position the substrate on the platform where the substrate is exposed so that the IC pattern on the reticle is precisely moved to a designated location on the substrate surface. There are various sub-systems to do this. However, since the substrate moves in any direction, an offset from the designated location and direction in which the substrate will be exposed is unavoidable. Adding a substrate pre-alignment system to the photolithography tool is a good solution for this orientation adjustment.

Most existing substrate pre-alignment systems rely on position marks formed on the substrate. That is, after accurately centering the substrate (i.e., after aligning its center to a designated point, such as the center of a chuck for carrying the substrate), the substrate is oriented in the desired orientation with the aid of positioning marks. can be accurately rotated to adjust the offset mentioned above. In spite of the TSV (Through-Silicon Via) process, which tends to be complicated, various types of marks are formed on various types of substrates. As a result, dedicated mark identification and position methods cannot handle all existing types of substrates, and versatile ones are required to handle specific tiers of TSV substrates. (requirements) cannot be met. Therefore, there is an urgent need in the art to design a novel substrate pre-alignment method that does not rely on detecting a signal from the substrate mark.

In order to overcome the aforementioned disadvantages, a substrate pre-alignment method and apparatus, and a photolithography tool are provided.

A provided substrate pre-alignment method is used to align a substrate to be pre-aligned to a predetermined position and predetermined orientation, the positioning device comprising: acquiring an image of a reference substrate positioned in a predetermined position and in a predetermined direction on a positioning device; acquiring an image of a substrate to be positioned to be pre-aligned on the positioning device; registering the image of the substrate to be pre-aligned with the image of the reference substrate, and obtaining an offset parameter of the image of the substrate to be pre-aligned from the image of the reference substrate; and adjusting the position and/or orientation of the substrate to be pre-aligned based on the offset parameter to complete the pre-alignment of the substrate to be pre-aligned.

Optionally, the step of registering the image of the substrate to be pre-aligned with the image of the reference substrate comprises:

extracting a plurality of feature points constituting a reference feature point set from the image of the reference substrate;

extracting a plurality of feature points constituting a registration feature point set from the image of the substrate to be pre-aligned; and

and registering the registration feature point set to the reference feature point set so that the registration feature point set substantially coincides with the reference feature point set.

Optionally, the plurality of feature points may include spots and/or corners.

Optionally, the registered feature point set may be registered with the reference feature point set using an Iterative Closest Point (ICP) algorithm.

Optionally, the step of registering the registered feature point set to the reference feature point set by using the Iterative Closest Point (ICP) algorithm includes, for each registered feature point of the registered feature point set, a reference feature point closest to the registered feature point selecting from the reference feature point set; calculating a rigid transformation with a minimum mean distance between each registered feature point and the corresponding nearest reference feature point, and determining an offset parameter; calculating a new transformation point set by applying the offset parameter to the registered feature point set; and repeating until the average distance between the set of transformation points and the set of reference feature points is less than a preset threshold, so that registration is completed.

Optionally, for each registered feature point in the registered feature point set, using a KD tree algorithm to calculate a plurality of nearest reference feature points from the reference feature point set, wherein the plurality of closest reference feature points are the It is close to the registration feature point.

Optionally, the substrate pre-alignment method calculates, for each registered feature point in the registered feature point set, a plurality of closest reference feature points from a reference feature point set close to the registered feature point, and then uses a Random Sample Consensus (RANSAC) algorithm The method may further include selecting a reference feature point closest to the registered feature point from the reference feature point set by removing a wrong reference feature point from the calculated closest reference feature points.

The provided substrate pre-alignment device comprises: a positioning device, which adjusts the position and/or orientation of a substrate to be pre-aligned to be carried on the positioning device; acquiring an image of a reference substrate positioned in a predetermined position and a predetermined orientation on the positioning device and an image of a substrate to be pre-aligned to be mounted on the positioning device, respectively; registering the image of the substrate to be pre-aligned with the image of the reference substrate; and an image acquisition and analysis device configured to calculate, from the image of the reference substrate, an offset parameter of the image of the substrate to be pre-aligned; and respectively connected to the image acquisition and analysis device and the positioning device, using the positioning device to adjust the position and/or the orientation of the substrate to be pre-aligned based on the offset parameter and a configured control device.

Optionally, the positioning device comprises a rotary table and a centering mechanism provided on one side of the rotary table, wherein the rotary table fixedly retains and rotates the substrate to be pre-aligned and the centering mechanism is configured to adjust the substrate to be pre-aligned such that a center of the substrate to be pre-aligned coincides with a center of rotation of the rotation table.

Optionally, said centering mechanism comprises a horizontal guide track and a centering table provided on said horizontal guide track, said centering table to move horizontally on said horizontal guide track ) can be configured.

Optionally, the positioning device further comprises a lifting table provided at the bottom of the turn table, wherein the lifting table is configured to lift or lower the turn table to pre-align the substrate. It may be configured to adjust the vertical height.

Optionally, the image acquisition and analysis device may include a CCD camera configured for an image acquisition and a calculating module.

Optionally, the substrate pre-aligning device further comprises a vision switching shaft, wherein the CCD camera is mounted on the vision switching shaft, the vision switching shaft to be pre-aligned with the substrate to be pre-aligned. and move the CCD camera to be within an imaging region of the CCD camera.

The provided photolithography tool comprises a substrate pre-alignment apparatus as defined above.

Compared with the prior art, the substrate pre-alignment method and apparatus and photolithography tool provided herein registers an image of a substrate to be pre-aligned with an image of a reference substrate using an image registration technique, and the substrate to be pre-aligned. After obtaining the offset parameter for , an angle adjustment of the substrate to be pre-aligned is followed based on the offset parameter. The provided substrate pre-alignment method and apparatus enables pre-alignment of a substrate without the need to form specific marks on the substrate, thereby helping to achieve high versatility and increase productivity.

1 is a schematic diagram of a substrate pre-alignment apparatus according to a specific embodiment of the present invention;
2 is a flow diagram of a method for pre-aligning a substrate according to a specific embodiment of the present invention.
3 shows an image of a reference substrate according to a particular embodiment of the present invention.
4 shows coordinates of feature points in an image of a reference substrate according to a specific embodiment of the present invention.
5 schematically illustrates an initial relationship of relative positions between a reference feature point set and a registered feature point set according to a specific embodiment of the present invention.
6 illustrates a registered feature point set registered in a reference feature point set according to a specific embodiment of the present invention.

The above objects, features and advantages of the present invention will become more apparent from the following detailed description of some specific embodiments which will be described in conjunction with the accompanying drawings. In describing the illustrated embodiments, it should be noted that, for the purpose of facilitating convenience and clarity, the drawings are not necessarily presented to scale and are provided in a very simplified form.

In an embodiment of the present invention, a substrate pre-alignment method and apparatus, and a photolithography tool are provided. The substrate pre-alignment device may include a positioning device, an image acquisition and analysis device, a vision switching shaft, and a control device.

1 , the positioning device includes a rotary table 2 , a centering mechanism disposed on one side of the rotary table 2 , and a lifting table disposed at the bottom of the rotary table 2 . may include The lifting table may be configured to change the vertical height of the pre-aligned substrate 1 disposed on the turn table 2 by raising or lowering the turn table 2 to aid image acquisition of the image acquisition and analysis device. The centering mechanism may be configured to adjust the substrate to be pre-aligned so that the center of the substrate 1 to be pre-aligned coincides with the center of rotation of the rotary table 2 , ie to perform a centering operation on the substrate 1 to be pre-aligned. there is. The rotation table 2 may be configured to hold the substrate 1 to be pre-aligned stationary and rotate it about a center of rotation. In this embodiment, the rotary table 2 preferably holds the substrate 1 to be pre-aligned with a vacuum, but the present invention provides any specific method for holding the substrate 1 to be pre-aligned with the rotary table 2 is not limited to In an alternative embodiment, the rotary table 2 may define a recess in which the substrate may be mounted and secured, or may hold the substrate in a snap-on manner. In addition, the centering mechanism may include a horizontal guide track 4 and a centering table 3 , which allows the centering table 3 to move horizontally on the horizontal guide track 4 . It is disposed on the horizontal guide track 4 so as to bring the coincidence of the center of the substrate 1 and the center of rotation of the rotation table 2 . In particular, the operations that may be performed with the centering mechanism may include: holding of the substrate 1 by the centering table 3 accompanied by the release of the substrate 1 from the rotation table 2 . (retention); and driving the substrate 1 by the centering table 3 to move relative to the turn table 2 so that the center of the substrate 1 coincides with the rotation center of the turn table 2 .

The image acquisition and analysis apparatus may comprise: a CCD camera 5 for capturing an image of the reference substrate and an image of the substrate 1 ; and a calculating module (not shown) for registering the image of the substrate 1 to the image of the reference substrate and calculating an offset parameter based on the registration of the image of the substrate 1 to the reference substrate image. The CCD camera 5 may be mounted on a vision switching shaft (not shown). The CCD camera 5 is moved with the vision switching shaft so that the substrate 1 is within the imaging region of the CCD camera 5 .

The control device may be connected to each of the calculation module and the rotary table 2 and configured to enable rotation of the rotary table 2 by an angle according to an offset parameter calculated by the calculation module.

As shown in Fig. 2, the substrate pre-alignment method according to the present invention includes the steps detailed below.

Step 1: Select a reference substrate and take an image. In this embodiment, the first substrate is defined as a reference substrate and its image is acquired with the CCD camera 5, and the acquired reference image is as shown in FIG. The image of the reference substrate presented in Fig. 3 merely exemplifies this embodiment and does not limit the present invention in any way. Although not limited thereto, the image of the reference substrate generally includes alignment mark(s). Alternatively, it may include other identifiable features other than alignment marks.

Step 2: Acquire an image of substrate 1. This may include the following steps:

A: Load the substrate 1 onto the rotation table 2 and use the centering table 3 to adjust the substrate 1 so that the center of the substrate to be pre-aligned is pre-aligned with the rotation center of the rotation table 2 to do;

B: move the CCD camera 5 radially with respect to the substrate 1 with a vision switching shaft until the substrate 1 is positioned within the imaging area of the CCD camera 5, and move the substrate ( 1) adjusting (ie, focusing) to a vertical height suitable for imaging with a CCD camera;

C: Rotating the substrate 1 360 degrees with the rotation table 2 simultaneously with the CCD camera 5 capturing the substrate 1 image.

Step 3: Register the image of the substrate 1 to the image of the reference substrate and determine the offset parameter of the registered image of the substrate to be pre-aligned.

The feature point may be extracted from the image of the reference substrate using an image processing program executed on the calculation module, and may be defined as a reference feature point set as shown in FIG. 4 , which indicates the coordinates of the feature point corresponding to the reference substrate . The horizontal and vertical coordinates are measured in pixels. Likewise, the feature points can be extracted from the image of the substrate 1 and can be defined as a set of registered feature points. The feature points may be selected from spots or corners of the image, or preferably both. Preferably, in this embodiment, spots and corners of the image are selected. Here, the spots correspond to the etched marks, and the corners themselves correspond to the marks of the substrate 1 . 5 shows the relationship of the initial relative positions between the reference feature point set and the registered feature point set in this embodiment, where A represents the reference feature point set and B represents the registered feature point set. As can be seen, there is a significant deviation between A and B.

An image algorithm may be used to register the registered feature point set to the reference feature point set. In this embodiment, the image algorithm is preferably an iterative closest point (ICP) algorithm. The registration may include the following steps: for each feature point in the registered feature point set, finding a feature point from a reference feature point set closest to the registered feature point; calculating a rigid transformation with a minimum mean distance between each registered feature point and a corresponding nearest reference feature point, and determining an offset parameter; calculating a new transformation point set by applying the offset parameter to a registered feature point set; and performing iteration until the average distance between the set of transformation points and the set of reference feature points is smaller than a preset threshold, so that registration is completed.

Preferably, using a KD tree (or k-dimensional tree, a data structure used to partition a data space into k dimensions) algorithm, a set of reference features for each feature in the set of registered feature points By calculating the feature points closest to , effectively increased computational efficiency can be obtained. In fact, for each feature point in the registered feature point set, there are a number of closest feature points including erroneous feature points that cannot be matched. To solve this problem, in this embodiment, a random sample consensus (RANSAC) algorithm may be used to remove erroneous feature points from the identified closest reference feature points, which can bring a significant increase in matching accuracy. there is.

로 표현되는 등록 특징점 세트 P와, 각 등록 특징점이

로 표현되고 j=1, 2, ..., M(M은 자연수)인 등록 특징점 세트 Q를 가정하면, 그들 간의 유클리드 거리(Euclidean distance)는

로 표현될 수 있고 목적 함수는

로 표현될 수 있다. i=1, 2, ..., N(N은 자연수)인

에 대해, R은 P의 변환 행렬(transformation matrix)을 나타내고, T는 Q의 변환 행렬이므로, 등록은 목적 함수를 최소화하는 변환 행렬 R 및 T를 얻는 것을 목적으로 한다. 더 높은 매칭 정확도를 얻기 위해, 이 실시예에서, RANSAC 알고리즘이 목적 함수

를 최소화하는

에 대한 최적 해를 얻는 데 사용될 수 있다. 여기서, RANSAC 알고리즘을 이용하여 최적 해를 얻는 것은 공지된 기술이므로 이에 대한 자세한 설명은 생략한다. 기판 (1)의 이미지에 대한 회전 파라미터는 결과적인 변환 행렬(R 및 T)로부터 계산될 수 있으며, 파라미터는 전술한 오프셋 파라미터로 정의된다. 도 6은 본 실시예에 따라 기준 특징점 세트에 등록된 등록 특징점 세트를 도시하며, 여기서 C는 기준 특징점 세트를 나타내고, D는 등록된 등록 특징점 세트를 나타낸다. 알 수 있듯이, C와 D 사이에는 오프셋이 거의 없으며 서로 거의 일치한다. 4단계에서, 오프셋 파라미터에 기초한 각도 보상(angular compensation)이 기판(1)에 대해 이뤄지고, 기판 사전 정렬 프로세스가 종료된다. 구체적으로, 계산 모듈은 오프셋 파라미터를 제어 장치에 전달할 수 있고, 제어 장치는 그 후 오프셋 파라미터에 기초하여 각도 보상을 달성하기 위해 회전 테이블(2)의 회전을 제어할 수 있다.For example, each registered feature point

A set of registered feature points P represented by , and each registered feature point

Assuming a set of registered feature points Q expressed as and j = 1, 2, ..., M (M is a natural number), the Euclidean distance between them

can be expressed as and the objective function is

can be expressed as i = 1, 2, ..., N (N is a natural number)

Since R denotes a transformation matrix of P and T is a transformation matrix of Q , registration aims at obtaining transformation matrices R and T that minimize the objective function. In order to obtain higher matching accuracy, in this embodiment, the RANSAC algorithm is

to minimize

can be used to obtain an optimal solution to Here, since obtaining an optimal solution using the RANSAC algorithm is a known technique, a detailed description thereof will be omitted. The rotation parameters for the image of the substrate 1 can be calculated from the resulting transformation matrices R and T , the parameters defined as the offset parameters described above. 6 illustrates a set of registered feature points registered in a reference feature set according to the present embodiment, where C denotes a reference feature set and D denotes a registered registered feature set. As can be seen, there is very little offset between C and D and they almost coincide with each other. In step 4, angular compensation based on the offset parameter is made for the substrate 1, and the substrate pre-alignment process is ended. Specifically, the calculation module may transmit the offset parameter to the control device, and the control device may then control the rotation of the rotary table 2 to achieve angular compensation based on the offset parameter.

It should be noted that the offset parameter in step 3 may also include a parameter of translation with respect to the substrate 1 . Accordingly, step 4 may further comprise causing the substrate 1 to translate with respect to the rotary table 2 with the centering table 3 based on the translation parameters.

Finally, it is determined whether the position of the substrate 1 meets the requirements. If so, the dictionary sort is considered successful. Otherwise, the above steps are repeated until a successful pre-alignment is achieved.

Also provided herein is a photolithography tool comprising the substrate pre-alignment apparatus discussed above.

In summary, in the substrate pre-alignment method and apparatus and photolithography tool to be provided herein, the image of the substrate 1 to be pre-aligned is registered to the image of the reference substrate using the ICP registration algorithm, and the KD tree and the RANSAC algorithm are Used for registration to obtain improved matching accuracy. Calculate the offset parameters required by the substrate 1 to achieve registration. The calculated offset parameter with respect to the substrate 1 can be used by the control device to adjust the position of the substrate 1 . This enables pre-alignment of the substrate 1 without the need to form specific marks thereon, thereby achieving high flexibility and increased productivity.

Obviously, those skilled in the art can make various modifications and variations of the present invention without departing from the spirit and scope of the present invention. Accordingly, it is intended that the present invention cover all such modifications and variations provided they come within the scope of the appended claims and their equivalents.

1,
1: substrate to be pre-aligned,
2: turn table,
3: centering table,
4: horizontal guide track;
5: CCD camera.

Claims (14)

A method of pre-aligning a substrate for adjusting a substrate to be pre-aligned in a predetermined position and in a predetermined direction, the method comprising:
acquiring an image of a reference substrate positioned on the positioning device in a predetermined position and in a predetermined direction;
acquiring an image of a substrate to be pre-aligned positioned on the positioning device;
registering the image of the substrate to be pre-aligned with the image of the reference substrate, and obtaining an offset parameter of the image of the substrate to be pre-aligned from the image of the reference substrate; and
adjusting the position and/or orientation of the substrate to be pre-aligned based on the offset parameter to complete the pre-alignment of the substrate to be pre-aligned;
A substrate pre-alignment method comprising a. According to claim 1,
Registering the image of the substrate to be pre-aligned with the image of the reference substrate comprises:
extracting a plurality of feature points constituting a reference feature point set from the image of the reference substrate;
extracting a plurality of feature points constituting a set of registered feature points from the image of the substrate to be pre-aligned; and
registering the registered feature point set to the reference feature point set so that the registered feature point set coincides with the reference feature point set;
A substrate pre-alignment method comprising a. 3. The method of claim 2,
wherein the plurality of feature points include spots and/or corners. 3. The method of claim 2,
The registered feature point set is registered in the reference feature point set using an Iterative Closest Point (ICP) algorithm. 5. The method of claim 4,
The step of registering the registered feature point set to the reference feature point set using the ICP algorithm comprises:
for each registered feature point in the registered feature point set, selecting a reference feature point closest to the registered feature point from the reference feature point set;
calculating a rigid body transformation with a minimum average distance between each registered feature point and the corresponding nearest reference feature point, and determining an offset parameter;
calculating a new set of transformation points by applying the offset parameter to the set of registered feature points; and
repeating until the average distance between the set of transformation points and the set of reference feature points is smaller than a preset threshold so that registration is completed;
A substrate pre-alignment method comprising a. 6. The method of claim 5,
For each registered feature point in the registered feature point set, a KD tree algorithm is used to calculate a plurality of closest reference feature points from the reference feature point set, and the plurality of closest reference feature points are the substrate dictionary close to the registered feature point. sort method. 6. The method of claim 5,
For each registered feature point in the registered feature point set, after calculating a plurality of closest reference feature points from the reference feature point set close to the registered feature point, the closest reference feature points calculated by using a Random Sample Consensus (RANSAC) algorithm and removing erroneous reference feature points from the reference feature point set and selecting a reference feature point closest to the registered feature point from the reference feature point set. A substrate pre-alignment apparatus comprising:
a positioning device for adjusting the position and/or orientation of a substrate to be pre-aligned mounted on the positioning device;
Acquire an image of a reference substrate positioned on the positioning device in a predetermined position and a predetermined direction and an image of a substrate to be pre-aligned mounted on the positioning device, respectively, and set the image of the substrate to be pre-aligned to the reference substrate an image acquisition and analysis device configured to register with the image of and calculate an offset parameter of the image of the substrate to be pre-aligned from the image of the reference substrate; and
a control device coupled to each of the image acquisition and analysis device and the positioning device and configured to adjust the position and/or the orientation of the substrate to be pre-aligned based on the offset parameter by using the positioning device;
A substrate pre-alignment device comprising a. 9. The method of claim 8,
The positioning device includes a rotary table and a centering mechanism provided on one side of the rotary table, wherein the rotary table is configured to hold the substrate to be pre-aligned fixedly and to rotate the substrate to be pre-aligned; and the centering mechanism is configured to adjust the substrate to be pre-aligned such that a center of the substrate to be pre-aligned coincides with a center of rotation of the rotary table. 10. The method of claim 9,
The centering mechanism includes a horizontal guide track and a centering table provided on the horizontal guide track, the centering table configured to move horizontally on the horizontal guide track. 10. The method of claim 9,
The positioning device further comprises a lifting table provided under the rotary table, the lifting table being configured to adjust a vertical height of the substrate to be pre-aligned by lifting or lowering the rotary table. 10. The method of claim 9,
The image acquisition and analysis device includes a CCD camera and a calculation module, wherein the CCD camera includes an image of a reference substrate positioned on the positioning device in a predetermined position and in a predetermined direction, and a dictionary mounted on the positioning device. a substrate to obtain an image of the substrate to be aligned, the calculation module registering the image of the substrate to be pre-aligned to the image of the reference substrate and calculating an offset parameter of the image of the substrate to be pre-aligned from the image of the reference substrate pre-alignment device. 13. The method of claim 12,
a substrate further comprising a vision switching shaft, wherein the CCD camera is mounted on the vision switching shaft, the vision switching shaft configured to move the CCD camera such that the substrate to be pre-aligned is within an imaging area of the CCD camera pre-alignment device. delete

Images