WO2016095282A1 - 晶片定位装置及方法 - Google Patents

晶片定位装置及方法 Download PDF

Info

Publication number
WO2016095282A1
WO2016095282A1 PCT/CN2014/095845 CN2014095845W WO2016095282A1 WO 2016095282 A1 WO2016095282 A1 WO 2016095282A1 CN 2014095845 W CN2014095845 W CN 2014095845W WO 2016095282 A1 WO2016095282 A1 WO 2016095282A1
Authority
WO
WIPO (PCT)
Prior art keywords
wafer
edge
lens
curve
ccd sensor
Prior art date
Application number
PCT/CN2014/095845
Other languages
English (en)
French (fr)
Inventor
李靖
Original Assignee
北京北方微电子基地设备工艺研究中心有限责任公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京北方微电子基地设备工艺研究中心有限责任公司 filed Critical 北京北方微电子基地设备工艺研究中心有限责任公司
Publication of WO2016095282A1 publication Critical patent/WO2016095282A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/67—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere
    • H01L21/68—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for positioning, orientation or alignment

Definitions

  • the present invention relates to the field of semiconductor device fabrication, and more particularly to a wafer positioning device and method.
  • FIG. 1 is a common wafer transfer system for semiconductor processes.
  • the transport system is composed of a cassette loading device 101 (LoadPort), a vacuum cassette lifting device 102 (VCE), a vacuum robot 103, and a transport chamber 104, and is capable of transferring the wafer to the reaction chamber. 105.
  • the wafer In order to ensure the stability of the process, the wafer needs to be delivered to a predetermined position in the reaction chamber with extremely high precision.
  • the theoretical transmission accuracy of the current manipulator can meet the set requirements, but in the actual transmission process, various uncertainties including mechanical vibration, mounting accuracy, etc., cause the wafer to be offset from the center of the finger of the robot.
  • the detection of the offset cannot be performed by the robot, so in order to ensure that the wafer can accurately reach the predetermined position of the reaction chamber to ensure the quality of the process, the wafer must be calibrated to eliminate the deviation. Therefore, a wafer aligning device is usually added to first determine the position of the wafer, and then fine-tune the position of the wafer to eliminate the deviation.
  • the position of the wafer there are mainly two ways to determine the position of the wafer: 1) using a plurality of sets of lenses to converge the light of the scattered light source on a CCD (Charge Coupled Device) sensor to determine the edge position of the wafer, and Further determine the position of the wafer; 2) The edge of the wafer is determined from the projection of the wafer edge on the CCD sensor using a parallel light source and the position of the wafer is further determined.
  • the first method requires high stability of the light source and the device is bulky.
  • the second method requires high installation accuracy of the light source, and the manufacturing cost of the parallel light source is also high. Therefore, it is an urgent problem to study a wafer positioning device that requires less light source.
  • a wafer positioning apparatus for achieving the objects of the present invention includes a rotary support structure and a wafer position acquisition structure for placing a wafer.
  • the wafer position acquisition structure includes a light source, a lens, a CCD sensor, and a data processing component, and an upper surface of the rotation support structure is interposed between the lens and the CCD sensor, wherein: the light source includes an LED;
  • the lens is an elongated lens, and the elongated lens is capable of diverging light emitted by the LED into an elongated spot having an area greater than or equal to a receiving surface area of the CCD sensor;
  • the CCD sensor is configured to receive from the strip a light above and output a signal corresponding to the intensity of the light to the data processing component;
  • the data processing component receives an output signal of the CCD sensor, and processes the received signal to obtain the wafer s position.
  • the LEDs are at least two
  • the light source further comprises a substrate
  • the at least two LEDs are fixed on the substrate.
  • the at least two LEDs are connected in parallel with the same power supply.
  • the at least two LEDs are arranged equidistantly on the substrate in a direction that coincides with a longitudinal direction of the elongated lens.
  • the substrate is horizontally disposed and parallel to the elongated lens.
  • the present invention also provides a wafer positioning method, which adopts A wafer positioning apparatus according to any of the above aspects of the invention is used.
  • the method comprises the following steps: S100, transferring a wafer to a rotating support structure, the rotation of the rotating support structure drives the edge of the wafer to sequentially pass through the wafer position collecting structure; S200, the CCD sensor collects light from above according to a preset sampling frequency, so as to Obtaining a first number of wafer edge position projection curves during one rotation of the wafer and transmitting the same to a data processing component; S300, the data processing component subtracts each of the edge position projection curves from a uniform light curve Obtaining a first number of difference curves; S400, performing an edge search on the difference curve to obtain edge pixels of the wafer corresponding to each of the difference curves, and according to preset pixel points and actual length values Corresponding relationship determines an actual length of the wafer edge corresponding to the edge pixel point; S500, determining an actual center
  • the wafer positioning method further includes the following steps: S600, determining an eccentric angle and an eccentricity of the wafer according to a center position of the rotating support structure and an actual center coordinate of the wafer.
  • the method for positioning a wafer further includes the step of determining a correspondence between the preset pixel point and an actual length value, and specifically comprising the steps of: performing a disc difference by performing steps S100 to S300 by replacing the wafer with a disk of a specified radius.
  • the preset number of disks of the specified radius includes at least two disks having different radii.
  • step S400 an edge search is performed on the difference curve to obtain edge pixel points of the wafer corresponding to each of the difference curves, specifically including the following steps: S410, filtering the difference curve Processing, obtaining a filter curve; S420, according to a preset local The window width and the local window threshold are used to locally compare the filter curve to obtain local information of the edge pixel point. S430: Perform an edge search using the OSTU algorithm in the local information to obtain the edge pixel.
  • the wafer positioning method determines a pixel point satisfying the following formula as a pixel point included in the partial information:
  • i denotes the i-th pixel point in the filter curve
  • f(i) denotes the intensity value of the i-th pixel point
  • l is the local window width
  • r is the local window threshold
  • the wafer positioning device of the present invention uses a common LED as a light source, and uses a long lens to diverge the light emitted by the LED into a long spot for detection by the CCD sensor, and the requirements of the entire device for the light source are greatly reduced, that is, using an ordinary LED
  • the actual position of the wafer can be determined, thereby greatly reducing equipment costs.
  • the light source can be processed by using only one lens, and therefore, the overall structure of the wafer positioning device is simple and easy to install.
  • the wafer positioning method provided by the present invention realizes the positioning of the wafer by using the wafer positioning device provided by the present invention, thereby realizing the positioning of the wafer by using a common light source, and the method is small in calculation amount, and the wafer position is determined accurately.
  • FIG. 1 is a schematic diagram of a wafer transfer system for a semiconductor process
  • FIG. 2 is a schematic structural view of a specific embodiment of a wafer positioning device of the present invention.
  • FIG. 3 is a graph showing the intensity of light received by a CCD sensor when no wafer is placed on the rotating support structure according to an embodiment of the present invention
  • FIG. 4 is a graph showing the intensity of light received by a CCD sensor when a wafer is placed on a rotating support structure according to an embodiment of the present invention
  • FIG. 5 is a flow chart of a specific embodiment of a wafer positioning method of the present invention.
  • FIG. 6 is a schematic diagram of a uniform light curve of a specific embodiment of a wafer positioning method of the present invention.
  • FIG. 7 is a schematic diagram showing a projection curve of a wafer edge position according to an embodiment of the wafer positioning method of the present invention.
  • Figure 8 is a schematic diagram showing the difference curve of Figure 6 and Figure 7;
  • FIG. 9 is a schematic diagram of a filter curve after the difference curve of FIG. 8 is filtered.
  • a wafer positioning apparatus includes a rotating support structure 001 for placing a wafer 100 and a wafer position collecting structure including a light source 201, a lens 202, a CCD sensor 203, and data processing.
  • Component (not shown).
  • the data processing component can be disposed in the upper computer of the manufacturing device of the semiconductor device, such as a computer controlled by the process, and the main function thereof is to receive the output signal of the CCD sensor 203 and process the signal to obtain the position of the wafer. .
  • the light source 201 includes an LED; the lens 202 is an elongated lens, and the elongated lens is capable of diverging the light emitted by the LED into an elongated spot having an area greater than or equal to the receiving surface area of the CCD sensor 203.
  • the CCD sensor 203 receives the intensity of the unobstructed light, as shown in FIG. 3 (the abscissa is the pixel point and the ordinate is the light intensity), and the light intensity varies greatly without obvious regularity.
  • the abscissa is the pixel point and the ordinate is the light intensity
  • the intensity of the light received by the CCD sensor 203 is as shown in FIG. 3
  • the data processing component can determine the edge position of the calibrated wafer 100 based on the varying light intensity, and can ultimately determine the position of the wafer 100 using a plurality of light intensity curves acquired by the wafer 100 for one revolution or more than one week. By comparing the actual position of the wafer 100 with the desired position of the wafer 100, the difference between the two is determined (ie, determining the positional deviation between the actual position of the wafer 100 relative to its desired position, ie, the wafer 100 is required to be moved to the desired position.
  • the amount of movement is moved, and the wafer 100 is moved using an adjustment member such as a robot to finely adjust the actual position of the wafer 100, thereby achieving calibration of the wafer 100.
  • an adjustment member such as a robot to finely adjust the actual position of the wafer 100, thereby achieving calibration of the wafer 100.
  • the center of the support member for supporting the wafer 100 in the rotary support structure 001 coincides with the center of the desired position of the wafer 100
  • the center of the actual position of the wafer 100 and the support member in the rotary support structure 001 can be compared by comparing the center of the actual position of the wafer 100.
  • the center of the determination determines the amount of movement required to move the wafer 100 to the desired location.
  • the wafer positioning device uses a common LED as a light source, and uses a long lens to diverge the light emitted by the LED into a long strip for detection by the CCD sensor 203.
  • the overall device requirements for the light source 201 are greatly reduced, i.e., the actual position of the wafer 100 can be determined using conventional LEDs, thereby greatly reducing equipment costs.
  • the light source 201 can be processed by using only one lens. Therefore, the overall structure of the wafer positioning device is simple and easy to install.
  • the lens 202 can use a semi-cylindrical elongated convex lens.
  • the substrate 204 may be disposed in the light source 201, and The light source 201 is fixedly disposed on the substrate 204.
  • the brightness of the edge of the long strip may be insufficient, that is, the brightness of the edge of the long strip will be much smaller than the brightness of the center of the strip, so that when the strip is long When the center and edge brightness values are too large, the final calibration result of the wafer 100 is affected.
  • at least two LEDs may be disposed on the substrate 204, and preferably two or more LEDs are evenly distributed on the substrate 204 to achieve light intensity. A more even effect.
  • the detection distance i.e., the distance between the light source 201 and the CCD sensor 203
  • a single LED can also be used.
  • the light source 201 includes at least two LEDs such that the at least two LEDs are arranged in a single row on the substrate 204 equidistant from each other, and the alignment direction is consistent with the longer direction of the elongated lens.
  • the center position of the LED light source 201 can be substantially coincident with the center position of the CCD sensor 203, so that the CCD sensor 203 can obtain a better light intensity curve.
  • the at least two LEDs can be connected in parallel with the same power supply. Since the intensity of the light emitted by the LED is related to the magnitude of the current flowing through it, the same power supply can be used to make the light intensity of the at least two LEDs have the same trend at any interval, so as to facilitate the subsequent light intensity.
  • the edge position of the wafer 100 is determined.
  • the shape of the substrate 204 can be set in accordance with the structure of the manufacturing apparatus of the semiconductor device. As an implementation, the substrate 204 can be placed horizontally and in parallel with the elongated lens.
  • the present invention also provides a wafer positioning method that uses the aforementioned wafer positioning device to determine the position of the wafer. As shown in FIG. 5, the wafer positioning method includes the following steps:
  • the wafer is transferred to the rotating support structure, and the rotating support structure rotates and drives the wafer to rotate along with the same, so that the circumferential edge of the wafer sequentially passes through the wafer position collecting structure.
  • the wafer positioning method is performed during the fabrication of the semiconductor device, which is part of the wafer calibration process in the fabrication process.
  • the wafer to be determined (to be calibrated) is generally placed by the robot to the rotating support structure of the device. on.
  • the light source in the wafer positioning device is in an open state during this process.
  • the present invention determines the edge position of the wafer by blocking the light from the light source by the wafer to be positioned.
  • the wafer circumferential direction is obtained by means of the CCD sensor.
  • the light intensity curve when each edge position point reaches the wafer position collecting structure that is, the occlusion of the light at the edge position point when each edge position point in the circumferential direction of the wafer reaches the wafer position collecting structure, to determine the specific point of the edge position point position.
  • the determination is based on: at the edge position of the wafer, the light intensity will jump.
  • the CCD sensor acquires a first number of edge position projection curves of the wafer rotation for one week according to a preset sampling frequency, and transmits the data to the data processing component.
  • the preset sampling frequency is determined in combination with the rotation speed of the rotating support structure, and the general rotation support structure is a uniform rotation, and the combination of the two determines an image of how many frames the CCD samples during the one-rotation of the wafer, that is, The number of frames sampled by the CCD during the one-rotation of the wafer corresponds to the first number, which is an integer greater than or equal to two. It can be understood that when the sampling frequency is high, the more the number of edge position projection curves of the same wafer to be calibrated is obtained, the determination of the wafer position is more accurate. However, there is no proportional relationship between the accuracy of the wafer position determination and the sampling frequency. Therefore, in consideration of the processing speed of the data processing unit, the sampling frequency of the CCD sensor can be maximized.
  • the CCD sensor with a sampling frequency of 5 MHz can be used, and in practical applications, the single frame sampling time of the CCD sensor can be used.
  • the interval is set to 1ms, which is 1KHz, so that the CCD sensor can acquire about 1500 frames of images by rotating the CCD sensor.
  • the first number is 1500.
  • the rotating support structure in the wafer positioning device can also be rotated at a non-uniform speed, as long as it matches the sampling frequency of the CCD sensor to obtain an image of a sufficient number of frames during one rotation of the wafer.
  • the so-called enough frames mean that according to the obtained frames The image determines the position of the wafer.
  • the data processing component subtracts the projection curve of each edge position from the uniform light curve to obtain a first quantity difference curve.
  • the light obtained by using the LED used in the wafer positioning device of the present invention as a light source is not uniform or parallel light.
  • the intensity of the light received by the CCD sensor has a similar trend, but since the unevenness is too obvious, the edge value cannot be calculated directly by a commonly used algorithm. Therefore, it needs to be properly processed to get the correct result.
  • the uniform light curve referred to here can also be said to be an ambient light curve received by the CCD sensor, which is obtained by the following steps.
  • the CCD sensor collects a plurality of light intensity curve images without wafer occlusion.
  • the collected multiple light intensity curves are then averaged. For example, if an image of m frames is acquired, there are m light intensity curves without occlusion of the wafer, and the CCD sensor has a total of n pixels.
  • i represents a single pixel in the CCD sensor, that is, the value of i ranges from 1 to n
  • j represents a single frame in the number of acquisition frames, that is, the value of j ranges from 1 to m.
  • f(i,j) is used to represent the intensity value of the i-th pixel in the single-frame image acquired in the jth time
  • the calculation formula of the uniform light curve ie, the ambient light curve
  • the ambient light intensity value of each pixel point can be obtained, and the ambient light intensity value is a value obtained by homogenizing the light intensity received by the CCD sensor.
  • the edge position projection curve obtained in the same device is shown in Fig. 7, and the difference between the curve light intensity shown in Fig. 6 and Fig. 7 is made.
  • a difference curve as shown in Fig. 8 is obtained.
  • the abscissas in FIGS. 6, 7, and 8 all indicate pixel points, and the ordinate indicates light intensity.
  • S500 Determine an actual center coordinate of the wafer according to the obtained first number of actual edge lengths of the wafer.
  • the obtained first number of wafer edge actual lengths can be fitted using a least squares method to obtain the actual center coordinates of the wafer, thereby determining the position of the wafer.
  • the wafer positioning method provided by the embodiment of the invention uses the LED as a light source, and homogenizes the obtained light intensity of the ambient light to obtain an ambient light curve, and performs the difference between the ambient light curve and the light intensity blocked by the wafer, and then performs the edge.
  • the search obtains the edge pixels of the wafer, thereby enabling the determination of the position of the wafer using a common light source, and the algorithm is simple and the calculation amount is small.
  • the discs are subjected to steps S100 to S300 instead of the wafers of the specified radius to obtain the difference curve of the disc, and the edge of the disc is further searched for the edge of the first number of discs.
  • Corresponding disc pixel point position The disc of the specified radius mentioned here is generally a mechanically finished aluminum alloy disc which can be closely matched with the support member, that is, the center of the support and the center of the support member when the aluminum alloy disc is used for inspection. The points are coincident.
  • the true wafer has a radius of 100 mm, and the calibration range is 5 mm.
  • the mechanically finished disc radius can be designed to be 105 mm, 102 mm, 100 mm, 98 mm, and 95 mm. And select a preset number of discs with a specified radius, Wherein, the selected disc includes at least a disc having a specified radius of a maximum calibration range, a minimum calibration range, and a real wafer radius.
  • the first number of disc pixel point positions are arithmetically averaged to obtain the pixel point position corresponding to the specified radius.
  • the specified discs having different radii are sequentially placed to obtain a preset number of pixel point positions corresponding to the preset number of discs of the specified radius.
  • the preset number is the number of selected specified radius discs.
  • the determination experiment of the correspondence relationship between the pixel point and the actual length value may be performed after the device is installed, and the pixel point and the actual length value may be performed for the wafer of the same specification.
  • the determination experiment of the correspondence relationship, but this determination step is not performed once in each wafer position determination. If there is already a corresponding relationship between the pixel points of the corresponding specifications and the actual length values, the same experiment is not required.
  • this correspondence is performed for a specific device or even a specific specification of the wafer, which can accurately characterize the actual length corresponding to the pixel point, so that the overall wafer position determination is more accurate. In practical applications, this also eliminates the difference between devices due to installation or assembly accuracy, and plays an initial role.
  • step S400 an edge search is performed on the difference curve to obtain edge pixel points of the wafer corresponding to each difference curve, which may be performed by using a local maximum inter-class difference method, and specifically includes the following steps:
  • the purpose of the filtering process is to filter out image noise and random interference as much as possible while keeping the contrast information of the image as much as possible.
  • Median filtering, averaging, or a combination of multiple filtering algorithms can be used to filter out noise and increase contrast.
  • the difference curve in FIG. 8 in the foregoing embodiment is subjected to filtering processing as shown in FIG. 9.
  • S420 Perform local comparison on the filter curve according to the preset partial window width and the local window threshold to obtain local information of the edge pixel.
  • a pixel that satisfies the following formula can be determined as a pixel included in the local information:
  • i denotes the i-th pixel point in the filter curve
  • f(i) denotes the intensity value of the i-th pixel point
  • l is the local window width
  • r is the local window threshold, i+2 ⁇ k ⁇ i+l -2.
  • k is a variable indicating a value in the range of i+2 and i+l-2.
  • r can be a value 115
  • l can take the value of 11. But this value is only a typical learning result, not applicable to all situations.
  • the local window width and the local window threshold can be obtained by self-learning in the first calibration (first position determination) of the same specification wafer, as described below:
  • the CCD sensor is started to collect the light intensity curve (edge position projection curve), and the window width and threshold are used for the light intensity curve.
  • Self-learning Let the window width be l, the window threshold be r, where l the preset value is odd and l>5; for the pixel point i, its light intensity value is f(i), which will satisfy the acquired light intensity curve ( 2) consecutive pixel points are considered partial windows. If a plurality of local windows (the same light intensity curve) satisfying the requirements are obtained in the search, the starting pixel points and the number of the partial windows are recorded, if multiple partial windows are The starting pixels are very close.
  • the window width is increased by 1 and the matching is performed again until the positions of the starting pixels are not close. If there are still multiple partial windows that meet the requirements at this time, add the window threshold to 1, and match again until there is only one matching window. If there is no matching window, the window threshold and the window length are both reduced by 1, and then again. Make a match.
  • the maximum number of times that can be matched can be preset, and the self-learning is stopped when the number of matching reaches the maximum value, and the input of the local window width and the partial window threshold is selected, or the device is inspected.
  • S430 Perform an edge search using the OSTU algorithm in the local information to obtain edge pixel points.
  • the algorithm is improved based on the traditional edge search algorithm maximum inter-class difference method (OSTU) method.
  • OSTU edge search algorithm maximum inter-class difference method
  • the traditional OSTU has higher accuracy for edge recognition, it requires a certain amount of time due to a large number of operations, which affects the sampling interval between frames, resulting in lower precision.
  • the local maximum inter-class difference method of the present invention makes the edge value no longer have strict requirements on the mechanical structure and optical structure of the manufacturing device of the semiconductor device, and the implementation cost is reduced.
  • the OSTU algorithm uses the OSTU algorithm to calculate the local information, so that the number of pixels that need to be calculated is reduced from the number of pixels of the CCD sensor (generally 2048) to the local window length (generally 15), so that the number of cycles is greatly reduced (required) It takes less than 0.5% of the original time, and because it does not need to get all the pre-processed data to match, the final edge value can be obtained before the single-frame image acquisition of the CCD sensor is completed, which greatly shortens the sampling interval. It is beneficial to improve the final fitting accuracy. At the same time, since only partial pixel information is used, problems caused by some noise and interference can be effectively eliminated.

Abstract

一种晶片定位装置及方法。其中该装置用于放置晶片(100)的旋转支撑结构(001)和晶片位置采集结构，在该晶片位置采集结构中：光源(201)包括LED；透镜(202)为长条形透镜，且长条形透镜能够将LED发出的光发散为面积大于等于CCD传感器(203)的接收面面积的长条形光斑；CCD传感器(203)用于接收来自其上方的光线，并将与该光线的强度相对应的信号输出至数据处理部件；数据处理部件接收CCD传感器(203)的输出信号，并对其进行处理，得到晶片的位置。整个装置对光源的要求降低，降低了设备成本。且只使用一个透镜对光源进行处理，设备整体结构精简，安装方便。其配合晶片定位方法实现了通过普通光源进行晶片位置的确定。该方法计算量小，晶片位置确定准确。

Description

晶片定位装置及方法 技术领域
本发明涉及半导体器件制造领域，特别是涉及一种晶片定位装置及方法。
背景技术
在半导体的制程工艺中(比如8英寸晶片)，待处理的晶片需要从大气环境中逐步传送到反应腔室中进行例如刻蚀工艺(Etch)、物理气相沉积(PVD)等工艺处理。把晶片传送到反应腔室，需要一个由一系列的大气设备和真空设备等组成的晶片传输系统，图1即是常见的半导体工艺用晶片传输系统。该传输系统由片盒装载装置101(LoadPort)、真空片盒升降装置102(VCE)、真空机械手103(Vacuum Robot)、传输腔室104(Transport Chamber)构成，并能够将晶片传输到反应腔室105中。
为了保证工艺处理的稳定性，晶片需要以极其高的精度传送至反应腔中的预定位置。当前机械手的理论传输精度能够达到设定的要求，但是在实际的传输过程中，会有包括机械振动、安装精度等在内的各种不确定的原因导致晶片相对于机械手的手指中心发生偏移，而该偏移量的检测是无法通过机械手完成的，因此为了保证晶片能够精确的到达反应腔的预定位置以保证工艺质量，就必须对晶片进行校准以消除该偏差。因此，通常会增加一个晶片校准装置，以便先确定晶片的位置，再对晶片进行位置的微调以消除偏差。
传统技术中主要有以下两种方式来确定晶片的位置：1)使用多组透镜将散光源的光在CCD(Charge Coupled Device，电荷藕合器件图像)传感器上汇聚成一点确定晶片边缘位置，并进一步确定晶片的位置；2) 使用平行光源根据晶片边缘在CCD传感器上的投影确定晶片的边缘，并进一步确定晶片的位置。其中第一种方法对光源的稳定性要求很高，且设备体积庞大。而第二种方法对光源的安装精度要求高，且平行光源的制造成本也较高。因此，研究一种对光源要求较低的晶片定位装置是一个亟待解决的问题。
发明内容
基于此，有必要针对传统技术中晶片定位装置中对光源要求很高的问题，提供一种对设备光源要求低，且晶片位置确定准确的晶片定位装置及方法。
为达到本发明目的提供的一种晶片定位装置，其包括用于放置晶片的旋转支撑结构和晶片位置采集结构。所述晶片位置采集结构包括光源、透镜、CCD传感器和数据处理部件，并且所述旋转支撑结构的上表面介于所述透镜和所述CCD传感器之间，其中：所述光源包括LED；所述透镜为长条形透镜，且所述长条形透镜能够将所述LED发出的光发散为面积大于等于所述CCD传感器的接收面面积的长条形光斑；所述CCD传感器用于接收来自其上方的光线，并将与该光线的强度相对应的信号输出至所述数据处理部件；所述数据处理部件接收所述CCD传感器的输出信号，并对所接收的信号进行处理，得到所述晶片的位置。
其中，所述LED至少为两个，所述光源还包括基板，所述至少两个LED固定在所述基板上。
其中，所述至少两个LED与同一个供电电源并联连接。
其中，所述至少两个LED在所述基板上以与所述长条形透镜的长度方向相一致的方向等距排列。
其中，所述基板水平设置，且与所述长条形透镜平行。
作为本发明的另一个方面，本发明还提供一种晶片定位方法，其采 用了本发明上述任一方案所述的晶片定位装置。该方法包括如下步骤：S100，将晶片传送至旋转支撑结构上，旋转支撑结构旋转带动晶片的边缘依次通过晶片位置采集结构；S200，CCD传感器按照预设采样频率采集来自其上方的光线，以便在所述晶片旋转一周的过程中获取第一数量的晶片边缘位置投影曲线，并将其传输至数据处理部件；S300，所述数据处理部件对每条所述边缘位置投影曲线与均匀光曲线相减得到第一数量的差值曲线；S400，对所述差值曲线进行边缘搜索，得到每条所述差值曲线对应的所述晶片的边缘像素点，并根据预设的像素点与实际长度值的对应关系确定所述边缘像素点对应的晶片边缘实际长度；S500，根据得到的第一数量的所述晶片边缘实际长度确定所述晶片的实际圆心坐标。
其中，该晶片定位方法还包括以下步骤：S600，根据所述旋转支撑结构的圆心位置与所述晶片的实际圆心坐标确定所述晶片的偏心角和偏心距。
其中，该晶片定位方法还包括确定所述预设的像素点与实际长度值的对应关系的步骤，具体包括以下步骤：用指定半径的圆盘代替所述晶片执行步骤S100～S300得到圆盘差值曲线，并继续对所述圆盘差值曲线进行边缘搜索得到第一数量的所述圆盘的边缘对应的圆盘像素点位置；对第一数量的所述圆盘像素点位置进行算术平均，得到所述指定半径对应的像素点位置；求得预设数量的指定半径的圆盘对应的预设数量的像素点位置；对预设数量的像素点位置及对应的指定半径进行直线拟合，得到预设的像素点与实际长度值的对应关系；所述预设数量的指定半径的圆盘包括至少两个半径不同的圆盘。
其中，在步骤S400中，对所述差值曲线进行边缘搜索，得到每条所述差值曲线对应的所述晶片的边缘像素点，具体包括以下步骤：S410，对所述差值曲线进行滤波处理，得到滤波曲线；S420，根据预设的局部 窗口宽度和局部窗口阈值对所述滤波曲线进行局部比较，获取所述边缘像素点所在的局部信息；S430，在所述局部信息中使用OSTU算法进行边缘搜索，得到所述边缘像素点。
其中，该晶片定位方法将满足以下公式的像素点确定为所述局部信息中包含的像素点：
Figure PCTCN2014095845-appb-000001
其中，i表示所述滤波曲线中的第i个像素点；f(i)表示第i个像素点的光强值；l为局部窗口宽度；r为局部窗口阈值；i+2≤k≤i+l-2。
本发明的有益效果如下：
本发明的晶片定位装置，使用了普通的LED作为光源，使用长条透镜将LED发出的光发散成为长条形光斑供CCD传感器检测，整个装置对光源的要求大大降低，即，使用普通LED即可确定晶片的实际位置，从而大大降低设备成本。并且，配合该光源可仅使用一个透镜来对其进行处理，因此，该晶片定位装置的整体结构简单、安装方便。
类似地，本发明提供的晶片定位方法，由于采用了本发明提供的晶片定位装置进行晶片的定位，因而实现了使用普通光源对晶片进行定位，且该方法计算量小，晶片位置确定准确。
附图说明
图1为半导体工艺用的晶片传输系统示意图；
图2为本发明的晶片定位装置的一具体实施例的结构示意图；
图3为本发明一实施例中旋转支撑结构上不放置晶片时CCD传感器所接收到的光线的强度曲线图；
图4为本发明一实施例中旋转支撑结构上放置晶片时CCD传感器所接收到的光线的强度曲线图；
图5为本发明的晶片定位方法的一具体实施例的流程图；
图6为本发明的晶片定位方法的一具体实施例的均匀光曲线示意图；
图7为本发明的晶片定位方法的一具体实施例晶片边缘位置投影曲线示意图；
图8为图6与图7做差后的差值曲线示意图；
图9为图8的差值曲线经过滤波后的滤波曲线示意图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图对本发明的晶片定位装置及方法的具体实施方式进行说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
本发明一实施例的晶片定位装置，如图2所示，包括用于放置晶片100的旋转支撑结构001和晶片位置采集结构，晶片位置采集结构包括光源201、透镜202、CCD传感器203和数据处理部件(未示出)。其中，数据处理部件可设置在半导体器件的制造设备的上位机中，如设置在工艺控制的计算机中等，其主要作用是接收CCD传感器203的输出信号，并对该信号进行处理，得到晶片的位置。光源201包括LED；透镜202为长条形透镜，且长条形透镜能够将LED发出的光发散为面积大于等于CCD传感器203的接收面面积的长条形光斑。
此处需要说明的是，旋转支撑结构001上没有放置晶片100时， CCD传感器203上接收到无遮挡的光线强度，如图3所示(横坐标为像素点，纵坐标为光线强度)，光线强度变化很大且无明显的规律。而当旋转支撑结构001上放置有晶片100时，晶片100的一部分会遮挡住到达CCD传感器203的光线，从而CCD传感器203接收到的光线会发生变化。有遮挡时，CCD传感器203接收到的光线强度，如图4所示(横坐标为像素点，纵坐标为光线强度)，其中，被遮挡部分的光线强度明显减弱。数据处理部件可根据变化的光线强度确定被校准晶片100的边缘位置，并可利用晶片100旋转一周或者大于一周所采集到的多个光线强度曲线最终确定晶片100的位置。通过比较晶片100的实际位置和晶片100的期望位置，确定二者之差(即，确定晶片100的实际位置相对于其期望位置之间的位置偏差，也即将晶片100移动至预期位置所需的移动量)，并使用诸如机械手等的调整部件移动晶片100，以对晶片100的实际位置进行微调，从而实现对晶片100的校准。当然，如果旋转支撑结构001中的用于支撑晶片100的支撑部件的中心与晶片100的期望位置的中心相重合，则可以通过比较晶片100的实际位置的中心与旋转支撑结构001中的支撑部件的中心，确定将晶片100移动至预期位置所需的移动量。
本发明实施例提供的晶片定位装置，使用普通LED作为光源，使用长条透镜将LED发出的光发散成为长条形光斑供CCD传感器203检测。整个装置对光源201的要求大大降低，即，使用普通LED即可确定晶片100的实际位置，从而大大降低设备成本。并且，配合该光源201可仅使用一个透镜来对其进行处理，因此，该晶片定位装置的整体结构简单、安装方便。
此处需要进一步说明的是，透镜202可使用半圆柱形的长条形的凸透镜。
具体的，在光源结构设置方面，可在光源201中设置基板204，将 光源201固定设置在基板204上。同时，如果使用单LED可能会造成长条形光斑的边缘的亮度不够，即长条形光斑的边缘的亮度值会比长条形光斑的中心的亮度值小很多，这样，当长条形光斑的中心与边缘的亮度值有过大的差距时，会影响晶片100的最终校正结果。为此，为保证长条形光斑的边缘的亮度，可在基板204上设置至少两个LED，并且最好是使两个或者两个以上的LED在基板204上均匀分布，以使光线强度达到更为均匀的效果。当然，若检测距离(即光源201与CCD传感器203之间的距离)较小时，也可使用单LED。
作为一种可实施方式，光源201包括至少两个LED，使所述至少两个LED彼此等距地在基板204上排列成单排，且排列方向与长条形透镜的较长方向一致。同时，可以将LED光源201的中心位置与CCD传感器203的中心位置大致重合，以使CCD传感器203能够获得更好的光线强度曲线。
较佳地，可将所述至少两个LED与同一个供电电源并联连接。因为LED发出的光线强度与流经其的电流的大小相关，因此使用同一个电源供电，可使所述至少两个LED的光线强度在任意间隔时刻的变化趋势大致相同，以便于后续通过光线强度确定晶片100的边缘位置。
其中，基板204的形状可配合半导体器件的制造设备的结构设置。作为一种可实施方式，可将基板204水平设置，且与长条形透镜平行。
本发明还提供一种晶片定位方法，其采用前述的晶片定位装置确定晶片的位置。如图5所示，该晶片定位方法包括如下步骤：
S100，将晶片传送至旋转支撑结构上，旋转支撑结构旋转并带动晶片随其一同旋转，以使晶片周向上的边缘依次通过晶片位置采集结构。
此处需要说明的是，此晶片定位方法是在半导体器件的制作过程中进行的，其是制作工艺中的晶片校准环节的一部分。制作过程中，一般通过机械手将待确定位置(待校准)的晶片放置到设备的旋转支撑结构 上。当然，此过程中晶片定位装置中的光源要处于开启状态。如前面所述，本发明是借助待定位晶片对来自光源的光线的遮挡而确定晶片的边缘位置，在晶片定位装置中的旋转支撑结构带动晶片旋转的过程中，借助CCD传感器获取晶片周向上的各个边缘位置点到达晶片位置采集结构时的光线强度曲线，即，获取晶片周向上的各个边缘位置点到达晶片位置采集结构时该边缘位置点对光线的遮挡情况，来确定该边缘位置点的具体位置。其中，该确定依据为：在晶片的边缘位置处，光线强度会发生跳变。
S200，CCD传感器按照预设采样频率获取晶片旋转一周的第一数量的边缘位置投影曲线，并传输至数据处理部件。
其中，所述的预设采样频率是结合旋转支撑结构的旋转速度确定的，且一般旋转支撑结构为匀速转动，两者结合确定在晶片旋转一周的过程中CCD采样多少帧的图像，即，在晶片旋转一周的过程中CCD所采样的帧数即对应为第一数量，其为大于等于2的整数。可以理解，采样频率较高时，获得的同一个待校准晶片的边缘位置投影曲线的个数越多，晶片位置的确定会更准确。但是晶片位置确定的准确性和采样频率之间并不是正比关系。因此，在考虑数据处理部件的处理速度的情况下，可尽量提高CCD传感器的采样频率。
作为一种可实施方式，在旋转支撑结构匀速转动的旋转速度为240°/s的情况下，可采用采样频率为5MHz的CCD传感器，且在实际应用中，可以将CCD传感器的单帧采样时间间隔设置为1ms，即1KHz，这样，晶片旋转一圈CCD传感器可以采集约1500帧图像。在此，第一数量为1500。
在其他实施例中，晶片定位装置中的旋转支撑结构也可采用非匀速转动，其只要和CCD传感器采样频率相配合在晶片旋转一周的过程中获得足够多帧的图像即可。所谓足够多帧，是指根据所获得的这些帧的 图像即可确定晶片的位置。
S300，数据处理部件对每条边缘位置投影曲线与均匀光曲线相减得到第一数量的差值曲线。
此处需要说明的是，本发明的晶片定位装置中使用的LED作为光源得到的光不是均匀的，也不是平行光。有晶片和无晶片时，CCD传感器上接收到的光线强度大致的趋势相同，但是由于不均匀性太明显，无法直接通过常用的算法对其计算得到边缘值。因此需要对其进行适当的处理，以得到正确的结果。此处所说的均匀光曲线，也可以说是CCD传感器所接收的环境光曲线，其通过如下步骤获得。
首先CCD传感器采集多帧没有晶片遮挡的光线强度曲线图像。之后再对所采集的多个光线强度曲线进行平均处理。如，设采集了m帧的图像，则有m个没有晶片遮挡的光线强度曲线，设CCD传感器共有n个像素。下面用i表示CCD传感器中的单个像素，即i的取值范围为1～n；用j表示采集帧数中的单帧，即j的取值范围为1～m。同时使用f(i,j)来表示在第j次采集的单帧图像中，第i个像素的光强值，则均匀光曲线(即环境光曲线)的计算公式为：
Figure PCTCN2014095845-appb-000002
依此公式即可得出每个像素点的环境光强值，该环境光强值是对CCD传感器接收到的光线强度进行均匀化后的数值。如在其中一次检测中求得的均匀光曲线如图6所示，在同一设备中某次获得的边缘位置投影曲线如图7所示，图6与图7所示的曲线光线强度做差值得到如图8所示的差值曲线。其中，图6、图7和图8中的横坐标均表示像素点，纵坐标表示光线强度。
S400，对差值曲线进行边缘搜索，得到每条差值曲线对应的晶片的 边缘像素点，并根据预设的像素点与实际长度值的对应关系，确定边缘像素点对应的晶片边缘实际长度，即确定该边缘像素点所对应的晶片边缘的实际坐标。
S500，根据得到的第一数量的晶片边缘实际长度确定晶片的实际圆心坐标。此处可对得到的第一数量的晶片边缘实际长度使用最小二乘法进行拟合，得到晶片实际的圆心坐标，从而确定晶片的位置。
本发明实施例提供的晶片定位方法，使用LED做光源，并对获取的环境光的光线强度进行均匀化获得环境光曲线，将环境光曲线与有晶片遮挡的光线强度做差处理后再进行边缘搜索，得到晶片的边缘像素点，从而能够实现使用普通光源对晶片位置的确定，且算法简单，计算量小。
进一步的，还可以继续执行以下步骤：
S600，根据旋转支撑结构的圆心位置与晶片的实际圆心坐标确定晶片的偏心角和偏心距。以此，使晶片的定位更加准确，从而使后续的晶片位置调节更加精细。在实际应用中，可将计算得出的晶片位置的调节量直接用于机械手，以对晶片的位置进行调整。
需要说明的是，预设的像素点与实际长度值的对应关系可针对每一半导体器件的制作设备进行实际检测获得，具体包括以下步骤：
启动晶片定位装置后，首先，用指定半径的圆盘代替晶片执行步骤S100～S300得到圆盘的差值曲线，并继续对圆盘的差值曲线进行边缘搜索得到第一数量的圆盘的边缘所对应的圆盘像素点位置。此处所说的指定半径的圆盘一般为机械精加工的铝合金材质圆盘，该圆盘能够与支撑件紧密配合，即认为使用该铝合金圆盘进行检测的时候其中心与支撑件的中心点是重合的。例如真实的晶片的半径为100mm，而支持校准的范围为5mm，则可以将机械精加工的圆盘半径的规格设计为105mm，102mm，100mm，98mm，95mm。且选取的预设数量的指定半径的圆盘， 其中，所选取的圆盘中至少包括指定半径为最大校准范围、最小校准范围以及真实晶片半径这三种规格的圆盘。
其次，对第一数量的圆盘像素点位置进行算术平均，得到指定半径所对应的像素点位置。
然后，依次放入半径不同的指定圆盘，求得预设数量的指定半径的圆盘所对应的预设数量的像素点位置。
最后，对预设数量的像素点位置及对应的指定半径进行直线拟合，得到预设的像素点与实际长度值的对应关系。所述预设数量即为所选取的指定半径圆盘的数量。
需要说明的是，对于同一半导体器件的制作设备，可在设备安装完成后进行一次像素点与实际长度值的对应关系的确定实验，也可以针对同一规格的晶片进行一次像素点与实际长度值的对应关系的确定实验，但是此确定步骤不是每次晶片位置确定中都要进行一次，如果已经有对应规格的像素点与实际长度值的对应关系，则无需再进行相同的实验。
此对应关系的确定是针对具体的设备甚至具体规格的晶片进行的，其可准确的表征像素点所对应的实际长度，使整体晶片位置确定更加精确。在实际应用中，以此也可以消除因为安装或者装配精度带来的设备间的差异，起到了一个初始化的作用。
具体地，步骤S400中，对差值曲线进行边缘搜索，得到每条差值曲线对应的晶片的边缘像素点，可采用局部最大类间差值法进行，具体包括以下步骤：
S410，对差值曲线进行滤波处理，得到滤波曲线。
滤波处理的目的是为了尽可能多的滤除图像噪声与随机干扰，同时尽量保持图像的对比度信息。可采用中值滤波处理、均值滤波或者多个滤波算法的组合等，以滤除噪声并增加对比度。如，前述实施例中的图8中的差值曲线经过滤波处理后如图9所示。
S420，根据预设的局部窗口宽度和局部窗口阈值对滤波曲线进行局部比较，获取边缘像素点所在的局部信息。
对于局部信息，可将满足以下公式的像素点确定为局部信息中包含的像素点：
Figure PCTCN2014095845-appb-000003
其中，i表示滤波曲线中的第i个像素点；f(i)表示第i个像素点的光强值；l为局部窗口宽度；r为局部窗口阈值，i+2≤k≤i+l-2。其中，k为变量，表示在i+2和i+l-2范围内取值。至于r,l的取值范围，由于r和l是通过自学习得到的，条件的不同，其自学习得到的结果也会不同，因此，没有特定的取值范围；例如，r可以取值为115，l可以取值为11.但是该取值只是一个典型的学习结果，并不适用于所有的情况。
对于局部窗口宽度和局部窗口阈值可在同一规格晶片的第一次校准(第一次位置确定)中通过自学习获得，具体按以下描述进行：
在已经获取均匀光曲线(环境光曲线)之后，在校准装置中放入晶片或者相似遮挡物，启动CCD传感器采集光线强度曲线(边缘位置投影曲线)，使用光线强度曲线对其进行窗口宽度与阈值的自学习。设窗口宽度为l,窗口阈值为r，其中l预设值为奇数且l>5；设对于像素点i，其光强值为f(i)，将采集到的光线强度曲线中满足式(2)的连续像素点认为局部窗口。若在搜索中得到了多个满足要求的局部窗口(同一光线强度曲线)，记录该局部窗口的起始像素点和数量，若多个局部窗口的 其起始像素很接近，如起始像素点之间间隔的像素点数小于等于2时，则将窗口宽度加1，重新进行匹配，直至起始像素点的位置均不接近为止。若此时仍有多个满足要求的局部窗口，则将窗口阈值加1，再次进行匹配，直至仅有一个匹配窗口为止，若没有匹配窗口，则将窗口阈值和窗口长度均减1，然后再次进行匹配。同时，可预先设定可进行匹配的最大次数，并在匹配次数达到最大值时停止自学习，选择人为输入局部窗口宽度和局部窗口阈值，或者对设备进行检修等。
在学习完成之后，需要再次采集一帧图像获取边缘位置投影曲线，并使用学习得到的窗口阈值和窗口宽度对其进行校验，检测该情况下边缘是否出现了较大的波动，如果波动较大，则需要重新进行校准学习，否则，则接受该学习的结果。
S430，在局部信息中使用OSTU算法进行边缘搜索，得到边缘像素点。
该算法基于传统的边缘搜索算法最大类间差值法(OSTU)法改进得来。因为传统的OSTU虽然对于边缘识别的准确度较高，但是由于涉及到大量的运算，使得计算结果的得出需要一定的时间，这会影响到每帧间的采样时间间隔，从而导致精度的降低。而采用本发明的局部最大类间差值法使得边缘值不再对半导体器件的制作设备的机械结构和光学结构有严格的要求，降低实现成本。且使用OSTU算法对局部信息进行计算，使得原来需要计算的像素点由CCD传感器的像素个数(一般为2048)降低为局部窗口长度(一般为15)，使得循环的次数大大减小(所需耗费的时间不足原来的0.5％)，而且由于不需要得到所有的预处理数据即可进行匹配，因此可以在CCD传感器单帧图像采集完毕之前即得到最终的边缘值，极大的缩短了采样间隔，有利于提高最终拟合精度。同时，由于只使用了局部的像素信息，可以有效的排除某些噪声和干扰引起的问题。
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。

Claims (10)

  1. 一种晶片定位装置，包括用于放置晶片的旋转支撑结构和晶片位置采集结构，其特征在于，所述晶片位置采集结构包括光源、透镜、CCD传感器和数据处理部件，并且所述旋转支撑结构的上表面介于所述透镜和所述CCD传感器之间，其中：
    所述光源包括LED；
    所述透镜为长条形透镜，且所述长条形透镜能够将所述LED发出的光发散为面积大于等于所述CCD传感器的接收面面积的长条形光斑；
    所述CCD传感器用于接收来自其上方的光线，并将与该光线的强度相对应的信号输出至所述数据处理部件；
    所述数据处理部件接收所述CCD传感器的输出信号，并对所接收的信号进行处理，得到所述晶片的位置。
  2. 根据权利要求1所述的晶片定位装置，其特征在于，所述LED至少为两个，所述光源还包括基板，所述至少两个LED固定在所述基板上。
  3. 根据权利要求2所述的晶片定位装置，其特征在于，所述至少两个LED与同一个供电电源并联连接。
  4. 根据权利要求2或3所述的晶片定位装置，其特征在于，所述至少两个LED在所述基板上以与所述长条形透镜的长度方向相一致的方向等距排列。
  5. 根据权利要求2所述的晶片定位装置，其特征在于，所述基板水平设置，且与所述长条形透镜平行。
  6. 一种晶片定位方法，其特征在于，采用权利要求1至5任意一项所述的晶片定位装置，包括如下步骤：
    S100，将晶片传送至旋转支撑结构上，旋转支撑结构旋转带动晶片的边缘依次通过晶片位置采集结构；
    S200，CCD传感器按照预设采样频率采集来自其上方的光线，以便在所述晶片旋转一周的过程中获取第一数量的晶片边缘位置投影曲线，并将其传输至数据处理部件；
    S300，所述数据处理部件对每条所述边缘位置投影曲线与均匀光曲线相减得到第一数量的差值曲线；
    S400，对所述差值曲线进行边缘搜索，得到每条所述差值曲线对应的所述晶片的边缘像素点，并根据预设的像素点与实际长度值的对应关系确定所述边缘像素点对应的晶片边缘实际长度；
    S500，根据得到的第一数量的所述晶片边缘实际长度确定所述晶片的实际圆心坐标。
  7. 根据权利要求6所述的晶片定位方法，其特征在于，还包括以下步骤：
    S600，根据所述旋转支撑结构的圆心位置与所述晶片的实际圆心坐标确定所述晶片的偏心角和偏心距。
  8. 根据权利要求6所述的晶片定位方法，其特征在于，还包括确定所述预设的像素点与实际长度值的对应关系的步骤，具体包括以下步骤：
    用指定半径的圆盘代替所述晶片执行步骤S100～S300得到圆盘差值曲线，并继续对所述圆盘差值曲线进行边缘搜索得到第一数量的所述圆盘的边缘对应的圆盘像素点位置；
    对第一数量的所述圆盘像素点位置进行算术平均，得到所述指定半径对 应的像素点位置；
    求得预设数量的指定半径的圆盘对应的预设数量的像素点位置；
    对预设数量的像素点位置及对应的指定半径进行直线拟合，得到预设的像素点与实际长度值的对应关系；
    所述预设数量的指定半径的圆盘包括至少两个半径不同的圆盘。
  9. 根据权利要求6所述的晶片定位方法，其特征在于，步骤S400中，对所述差值曲线进行边缘搜索，得到每条所述差值曲线对应的所述晶片的边缘像素点，具体包括以下步骤：
    S410，对所述差值曲线进行滤波处理，得到滤波曲线；
    S420，根据预设的局部窗口宽度和局部窗口阈值对所述滤波曲线进行局部比较，获取所述边缘像素点所在的局部信息；
    S430，在所述局部信息中使用OSTU算法进行边缘搜索，得到所述边缘像素点。
  10. 根据权利要求9所述的晶片定位方法，其特征在于，将满足以下公式的像素点确定为所述局部信息中包含的像素点：
    Figure PCTCN2014095845-appb-100001
    其中，i表示所述滤波曲线中的第i个像素点；f(i)表示第i个像素点的光强值；l为局部窗口宽度；r为局部窗口阈值；i+2≤k≤i+l-2。
PCT/CN2014/095845 2014-12-19 2014-12-31 晶片定位装置及方法 WO2016095282A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410799981.0A CN105762101B (zh) 2014-12-19 2014-12-19 晶片定位装置及方法
CN201410799981.0 2014-12-19

Publications (1)

Publication Number Publication Date
WO2016095282A1 true WO2016095282A1 (zh) 2016-06-23

Family

ID=56125713

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/095845 WO2016095282A1 (zh) 2014-12-19 2014-12-31 晶片定位装置及方法

Country Status (2)

Country Link
CN (1) CN105762101B (zh)
WO (1) WO2016095282A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109062013A (zh) * 2018-09-06 2018-12-21 重庆科技学院 一种光刻机小工件卡具
CN116995011A (zh) * 2023-08-16 2023-11-03 泓浒(苏州)半导体科技有限公司 一种用于晶圆载体装卸的校准方法和校准系统

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107863311B (zh) * 2017-11-03 2020-02-14 上海华力微电子有限公司 一种检测及校正晶圆与腔体载物台偏移的装置及方法
CN109343037A (zh) * 2018-11-27 2019-02-15 森思泰克河北科技有限公司 光探测器安装误差检测装置、方法和终端设备
CN110440720A (zh) * 2019-08-14 2019-11-12 中国科学院长春光学精密机械与物理研究所 一种芯片与基板的平行性检测装置
CN113470108B (zh) * 2021-09-06 2022-02-22 中导光电设备股份有限公司 一种晶圆圆心偏移检测方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0745513A (ja) * 1993-07-30 1995-02-14 Canon Inc 露光装置
US20010042845A1 (en) * 2000-04-07 2001-11-22 Varian Semiconductor Equipment Associates, Inc. Wafer orientation sensor for GaAs wafers
CN2891271Y (zh) * 2006-03-17 2007-04-18 北京中科信电子装备有限公司 离子注入机晶圆定位控制系统
CN1975997A (zh) * 2005-12-02 2007-06-06 日东电工株式会社 半导体晶片的定位方法及使用它的装置
CN101127317A (zh) * 2007-09-18 2008-02-20 深圳市矽电半导体设备有限公司 晶圆片的定位装置及定位方法
CN101551593A (zh) * 2009-04-24 2009-10-07 上海微电子装备有限公司 用于光刻装置的对准系统、光刻装置及其对准方法
CN102842485A (zh) * 2011-06-23 2012-12-26 上海微电子装备有限公司 硅片处理装置及其处理方法
CN103811387A (zh) * 2012-11-08 2014-05-21 沈阳新松机器人自动化股份有限公司 晶圆预对准方法及装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3379200B2 (ja) * 1994-03-25 2003-02-17 株式会社ニコン 位置検出装置
JP4803703B2 (ja) * 2005-02-28 2011-10-26 日東電工株式会社 半導体ウエハの位置決定方法およびこれを用いた装置
JP4566798B2 (ja) * 2005-03-30 2010-10-20 東京エレクトロン株式会社 基板位置決め装置，基板位置決め方法，プログラム
JP5132904B2 (ja) * 2006-09-05 2013-01-30 東京エレクトロン株式会社 基板位置決め方法，基板位置検出方法，基板回収方法及び基板位置ずれ補正装置

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0745513A (ja) * 1993-07-30 1995-02-14 Canon Inc 露光装置
US20010042845A1 (en) * 2000-04-07 2001-11-22 Varian Semiconductor Equipment Associates, Inc. Wafer orientation sensor for GaAs wafers
CN1975997A (zh) * 2005-12-02 2007-06-06 日东电工株式会社 半导体晶片的定位方法及使用它的装置
CN2891271Y (zh) * 2006-03-17 2007-04-18 北京中科信电子装备有限公司 离子注入机晶圆定位控制系统
CN101127317A (zh) * 2007-09-18 2008-02-20 深圳市矽电半导体设备有限公司 晶圆片的定位装置及定位方法
CN101551593A (zh) * 2009-04-24 2009-10-07 上海微电子装备有限公司 用于光刻装置的对准系统、光刻装置及其对准方法
CN102842485A (zh) * 2011-06-23 2012-12-26 上海微电子装备有限公司 硅片处理装置及其处理方法
CN103811387A (zh) * 2012-11-08 2014-05-21 沈阳新松机器人自动化股份有限公司 晶圆预对准方法及装置

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109062013A (zh) * 2018-09-06 2018-12-21 重庆科技学院 一种光刻机小工件卡具
CN109062013B (zh) * 2018-09-06 2023-06-06 重庆科技学院 一种光刻机小工件卡具
CN116995011A (zh) * 2023-08-16 2023-11-03 泓浒(苏州)半导体科技有限公司 一种用于晶圆载体装卸的校准方法和校准系统
CN116995011B (zh) * 2023-08-16 2024-01-09 泓浒(苏州)半导体科技有限公司 一种用于晶圆载体装卸的校准方法和校准系统

Also Published As

Publication number Publication date
CN105762101A (zh) 2016-07-13
CN105762101B (zh) 2019-02-19

Similar Documents

Publication Publication Date Title
WO2016095282A1 (zh) 晶片定位装置及方法
TWI457685B (zh) 用以定位及檢驗基底之補償校正方法及配置
TWI431704B (zh) 用以定位基底之偏移校正技術
US10600667B2 (en) Systems and methods for wafer alignment
WO2016138744A1 (zh) 检测显示面板表面平坦度的装置及方法
TWI515785B (zh) 用以改善基板間之斜角蝕刻再現性的裝置與方法
TW201349325A (zh) 液處理裝置，液處理方法及記憶媒體
JP7398483B2 (ja) 非対称ウエハ形状特徴付けのためのメトリック
JP6598782B2 (ja) 位置敏感基板デバイス
US20210242044A1 (en) Substrate bonding device, calculation device, substrate bonding method, and calculation method
TWI390660B (zh) 用於半導體晶圓對準之方法與設備
JP2007184529A (ja) 半導体ウエハの検査装置
CN113670207A (zh) 一种将光学探头集成在硅晶圆片上的检测系统及其应用
KR101818061B1 (ko) 듀얼 오토포커싱을 적용한 영상 검출 시스템
US20220299317A1 (en) Method for positioning substrate
TW202009478A (zh) 基板檢查裝置、基板處理裝置、基板檢查方法及基板處理方法
TWI813268B (zh) 工件之雙面研磨裝置以及雙面研磨方法
JP2597754B2 (ja) 基板の回転補正方法
TWI565927B (zh) 測量系統
TW202338129A (zh) 使用基於圖像的質量分佈計量來監控樣本屬性
CN116428969A (zh) 投影画面尺寸分析方法、装置、及测量装置
CN103278242B (zh) 光辐射场均匀性测量装置及测量方法
JP2007304002A (ja) 欠陥検査方法、欠陥検査装置、欠陥検査プログラム、このプログラムを記憶した記憶媒体

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14908317

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14908317

Country of ref document: EP

Kind code of ref document: A1