CN115854863A - Focusing method applying axial chromatic aberration and optical detection framework using same - Google Patents

Abstract

The invention discloses a focusing method applying axial chromatic aberration and an optical detection framework using the method. The light of different wave bands can image the sample surface with different depths through the optical system, that is, for the sample surface with any specific depth, the light of different wave bands can present different focusing degrees through the optical imaging system. The degree of quasi-focus can be calculated by an algorithm to obtain a quantitative index, and different colors of light correspond to the respective quantitative index at each sample depth. The method includes the steps of establishing a quantization index ratio lookup table among different colored lights in advance, when the optical detection equipment performs on-line detection, acquiring one image and dividing the image into first wave band image data and second wave band image data at each detection point, acquiring a quantization index value corresponding to each detection point in the quasi-focus degree analysis, matching corresponding position parameters in the quantization index ratio lookup table according to the ratio of the two values, further acquiring an adjustment basis of the focal plane position of the optical detection equipment, and improving the detection efficiency.

Description

Focusing method applying axial chromatic aberration and optical detection framework using same

Technical Field

The present invention relates to an optical detection technology, and more particularly, to a focusing method using axial chromatic aberration and an optical detection architecture using the same.

Background

With the improvement of semiconductor processing technology, integrated Circuit (IC) chips contain more transistors per unit area, providing efficient operation with multiplexing and low power consumption. With the goal of increasing productivity, the principle of High-Volume Manufacturing (HVM) is more significant in terms of time cost. In various stages of semiconductor manufacturing processes, corresponding inspection procedures are required to be deployed to ensure the production quality of a production line and avoid the defective products from flowing into the subsequent manufacturing processes to cause waste. However, these detection procedures tend to add unnecessary time costs.

The optical detection systems are used to detect whether the surface of the semiconductor device has defects, which is caused by the limitation of the depth of field of the lens assembly of each optical detection system. The depth of field range can provide the resolution of the optical detection system in the axial direction, and the depth of field range can not provide clear resolution images for detection and analysis. Under the current process scale of semiconductor manufacturing, the working distance range (depth of field) of the lens carried by the optical detection system in the axial direction is about micron level. Therefore, when the height difference of the surface topography on the chip or the detection height error between different crystal grains caused by the translation of the detection conveyer belt is larger than the depth of field, the optical detection system often needs to search and adjust through the focal plane, so that the surface of the detection point on the chip can fall into the depth of field range of the optical detection system.

Optical inspection systems traditionally use serial motion to find the focal plane, i.e. the optical inspection system scans each inspection point of the dut axially individually. Each scan must rely on mechanical action to change the separation between the lens (objective lens) and the dut to determine whether the focal plane position at this point is in focus. The time-consuming mechanical repetitive motion procedure does not satisfy the fast detection principle required under High Volume Manufacturing (HVM). The efficiency bottleneck of the automatic optical inspection architecture with high time cost is a great technical breakthrough.

Disclosure of Invention

One of the objectives of the present invention is to save the detection time required for optical detection and improve the detection efficiency.

It is another object of the present invention to adapt an optical inspection system for High Volume Manufacturing (HVM) in-line inspection.

In order to achieve the above object, the present invention provides a focusing method using axial chromatic aberration, which enables a detector to complete focusing on a detection point in a first axial direction in a detection procedure based on a quantization index ratio lookup table established in advance, the method comprising: the inspection machine obtains an inspection image within a working interval at the inspection point. The detected image is divided into a first band image data and a second band image data of two different bands, and a quantization index value corresponding to each of the first band image data and the second band image data in the definition analysis is obtained. And obtaining a ratio between the first waveband image data and the second waveband image data according to the respective quantization index values of the first waveband image data and the second waveband image data, and searching a matched position parameter in a quantization index ratio lookup table according to the ratio. And adjusting the focal plane position of the detection machine according to the position parameter. The quantization index ratio lookup table is used for acquiring a quantization index ratio corresponding to two different wave bands and difference degree information between the current focal plane position of the detector and the reference focal plane position based on the image capture of the detector in each identification interval, wherein the difference degree information is used as a position parameter.

In an embodiment of the present invention, a denominator of the ratio is a quantization index value corresponding to the second band of image data, a numerator of the ratio is a quantization index value corresponding to the first band of image data, the first band of image data is based on a light with a first wavelength, and the second band of image data is based on a light with a second wavelength. The second wavelength may be longer than the first wavelength.

In an embodiment of the present invention, taking an expected value equal to 1 as an example, when the ratio is not equal to the expected value, in the step of adjusting the focal plane position of the detection machine, the focal plane position of the detection machine is adjusted toward the direction of the corresponding focal plane position with the quantization index ratio of 1 in the quantization index ratio lookup table.

In an embodiment of the present invention, the second wavelength may be red light, and the first wavelength may be blue light.

In an embodiment of the present invention, when the ratio is smaller than the quantization index ratio corresponding to the reference focal plane position, the control direction of the detection machine is decreased in the first axial direction in the step of adjusting the focal plane position of the detection machine, and when the ratio is larger than the quantization index ratio corresponding to the reference focal plane position, the control direction of the detection machine is increased in the first axial direction in the step of adjusting the focal plane position of the detection machine.

In order to achieve the above object, the present invention further provides an optical inspection apparatus for inspecting a device to be inspected, comprising: objective lens subassembly, light source subassembly, guide subassembly, adjusting part, sensor subassembly, and control host computer. The guiding component is coupled with the objective lens component at the lower end and is coupled with the light source component at the side edge, the guiding component guides the irradiation light generated by the light source component to irradiate towards the objective lens component, and the guiding component guides the reflection light from the device to be tested to the upper end of the guiding component. The adjusting component is coupled to the upper end of the guiding component to receive the reflected light and enable the reflected light to pass through and is used for adjusting the position of a focal plane formed by the objective lens component. The sensor assembly is coupled to the upper end of the adjusting assembly to receive the reflected light to generate image data. The control host is coupled with the adjusting component and the sensor component to receive the image data generated by the sensor component and control the adjusting component to adjust the focal plane position. The control host is used for executing the focusing method, and the control host controls the adjusting component to adjust the position of the focal plane according to a quantization index ratio lookup table stored in the control host and image data generated by the sensor component.

In an embodiment of the invention, the guiding assembly guides the irradiated light and the reflected light generated by the light source assembly through a half-reflecting mirror.

In an embodiment of the present invention, the light source assembly is configured to provide two different wavelength bands of illumination light simultaneously, wherein one wavelength band may be red light, and the other wavelength band may be blue light.

Therefore, the invention applies the characteristic of axial chromatic aberration to destroy the symmetry of the quantization index through two light waves, and further extracts key information required by matching type focusing. Different differences of light rays of different wave bands on the quantization index value corresponding to the defocusing or focusing state are utilized, so that the matching relation in the preset stage can be applied to data in actual detection, the detection machine can perform a focusing program by only shooting one image under the operation of the matching type focusing mechanism, the long and time-consuming process of sequence scanning is avoided, the method can be applied to defect detection of HVM production lines, and the efficiency is remarkably improved.

Drawings

FIG. 1 is a schematic view of the focus finding of an optical detection system during sequential detection;

FIG. 2 is a diagram of the quantization index M at a detection point in relation to the focal plane position P;

FIG. 3 is a flowchart of a focusing method according to an embodiment of the invention;

FIG. 4 is a schematic view of focusing light rays of different wavelength bands in the axial direction;

FIG. 5 is a diagram showing the relationship between the quantization index M, the ratio r of the quantization indexes and the focal plane position P of the light rays in different bands at a detection point;

FIG. 6 is a schematic diagram of an optical inspection apparatus according to an embodiment of the present invention.

Detailed Description

For a fuller understanding of the objects, features and advantages of the present invention, reference should now be made to the following detailed description taken in conjunction with the accompanying drawings in which:

in this application, the terms "a" or "an" are used to describe a unit, component, structure, device, module, system, region, or area, etc. This is done for convenience of illustration only and to provide a general sense of the scope of the invention. Thus, unless clearly indicated to the contrary, this description should be read to include one or at least one and the singular also includes the plural.

In this application, the terms "comprises, comprising, having," or any other similar term, are not intended to be limited to the particulars listed herein, but rather may include additional elements not expressly listed or inherent to such elements, components, structures, devices, modules, systems, regions or regions.

In this application, the use of the ordinal number for the description "first" or "second" is used to distinguish or refer to the same or similar elements, structures, portions or regions, and does not necessarily imply a spatial order to such elements, structures, portions or regions. It should be understood that in some cases or configurations, ordinal terms may be used interchangeably without affecting the practice of the invention.

FIG. 1 is a schematic view of focusing of an optical detection system during sequential detection. The inspection machine 300 of the optical inspection system is disposed above the carrier 100, and the device 200 to be inspected is carried on the carrier 100.

As shown in fig. 1, the detector 300 has a degree of freedom in a first axial direction X, which is parallel to the optical axis of the detector 300, and can move back and forth up and down, so that the focal plane of the detector 300 can move within a working range W. This working range W is set to cover all the variations of the surface topography of the device under test 200. The working interval W may be maintained by the movement of the susceptor 100 in the first axial direction X in a reciprocating manner up and down while the detecting machine 300 is stationary. The following description will take the operation mode of the inspection machine 300 being able to move freely and the carriage 100 being stationary as an example.

The microscopic function of the inspection machine 300 has its corresponding identification region (depth of field) with the appropriate objective lens. In the working interval W of the detecting machine 300, the identification interval will move up and down to change its position in the first axial direction X. The working interval W is for the inspection machine 300 to find a proper focal plane position (with the best resolution and the higher definition) on the surface of the dut 200, and the image of the position is used as the basis for the subsequent analysis.

Therefore, the inspection machine 300 can form multiple recognition intervals with normal resolution in the working interval W, and the length of each recognition interval corresponds to the depth of field that the inspection machine 300 can provide. For convenience of illustration, fig. 1 only illustrates four identification intervals, which are a first identification interval d1, a second identification interval d2, a third identification interval d3 and a fourth identification interval d4. After the detector 300 obtains the images of the four recognition intervals, the images are analyzed, and an optimal image is estimated by an algorithm, so that the optimal image is used as a basis for subsequent analysis.

As further illustrated in fig. 1, the dut 200 has a first portion 210 and a second portion 220, and the surface topography of the first portion 210 is lower than the surface topography of the second portion 220. First, assuming that only one inspection point is allocated to each part to be inspected, when the inspection point is the first part 210, the inspection machine 300 needs to be moved to a first position in the first axial direction X, where the first position allows the first identification interval d1 (the depth of field range of the inspection machine 300) to fall on the position shown in fig. 1, and then a first image is obtained. Then, the detector 300 is moved to a second position in the first axial direction X, where the second recognition interval d2 is located at the position shown in fig. 1, and a second image is obtained. Then, the detector 300 is moved to a third position in the first axial direction X, where the third position allows the third recognition interval d3 to fall on the position shown in fig. 1, and then a third image is obtained. Then, the inspection machine 300 is moved to a fourth position in the first axial direction X, where the fourth recognition interval d4 is located at the position shown in fig. 1, and a fourth image is obtained. Finally, the first to fourth images are analyzed to determine which of the best detected images (with the best sharpness) is at the first location 210. The best detection image is used as the basis for the optical detection system to determine the defects of the device to be detected. Each recognition section has a focus surface position of a corresponding in-focus, and clear images are formed in a certain range (i.e., the recognition section) on the upper side and the lower side of the focus surface position based on the depth of field range.

In the sequential inspection illustrated in fig. 1, it takes a lot of time to acquire the inspection images (such as the above-mentioned identification sections d1 to d 4) in the first axis direction for each inspection point in the production line, so as to select the best inspection image corresponding to each inspection point.

The focusing methods of the optical system are classified into active focusing and passive focusing. The active focusing is to actively emit a specific light wave to the object to be measured and receive the light wave reflected from the object to be measured to determine the distance between the object to be measured and the optical system, so as to achieve the purpose of focusing, for example: infrared ranging and ultrasonic ranging. The passive focusing is to determine the focusing position by analyzing the obtained image data of the object to be measured, so as to achieve the purpose of focusing. The focusing method adopted in the semiconductor manufacturing process of the present embodiment is passive focusing.

The passive focusing method usually uses the existing algorithm as the calculation basis, and then calculates the corresponding definition evaluation function value according to each image in the focusing process, and finally determines the clearest imaging position according to the position corresponding to the maximum definition evaluation function value. Accordingly, the focus-finding method further estimates the image data through an algorithm to obtain the evaluation function value of each detection point at different positions (refer to fig. 1) in the axial direction, and further finds out the image data with the maximum evaluation function value, where the position (refer to the position of the detector 300 in the first axial direction X, refer to fig. 1) corresponding to the image data is the position where the detector can image clearly at the detection point, that is, the identification interval (depth of field) provided by the detector 300 can cover the surface topography of the object to be measured at the detection point.

The basis on which existing algorithms are based includes: a gray scale function, a frequency domain function, etc. These functions are selectively used to generate the sharpness evaluation function value corresponding to each recognition interval of each detection point in the axial direction, and these functions are all suitable for the focus-finding method described in the subsequent embodiments of the present application.

The following is an introduction of several algorithms. An algorithm based on a gray scale function, such as: the Tenengrad gradient function is to use Sobel operand to extract the gradient values in horizontal and vertical directions in the image data. The Sobel operand is an existing discrete differentiation operand (discrete differentiation operand).

Operating the image gray function, wherein the approximate gradient at (x, y) satisfies the following conditions:

in formula (1): g (x, y) represents the approximate gradient value at (x, y); g _x (x, y) represents the gradient value extracted in the horizontal direction by using a Sobel operator; g _y (x， _y ) Representing the gradient value extracted in the vertical direction by using a Sobel operator; g _x (x， _y )、G _y (x， _y ) Respectively, the recorded images f (x, _y ) Convolution with different Sobel operands is calculated according to the following relation:

G _x (x，y)＝f(x，y)*S ₁ (2)

G _y (x，y)＝f(x，y)*S ₂ (3)

in the formulae (2) and (3), S ₁ 、S ₂ Is a Sobel operand template.

Finally, the calculation of the Tenengrad function can be defined as follows:

F＝∑ _x ∑ _y G ² (x，y)，|G(x，y)|＞T (4)

in equation (4), T is a threshold value for edge detection, and the sensitivity of the evaluation function can be changed by adjusting the value of T.

Accordingly, in the working interval W illustrated in fig. 1, each recognition interval can obtain a corresponding image data, and then obtain a corresponding value through the Tenengrad gradient function, and select the best value among the values, where the recognition interval corresponding to the value is the best working distance of the inspection machine 300 to the inspection point.

Another algorithm based on a gray scale function, for example: the Laplacian gradient function is basically consistent with the Tenengrad gradient function, and mainly replaces a Sobel operator with a Laplacian operator. The Laplacian gradient function is a function that operates based on gray-scale values of a target point and 4 points around the target point. The Laplacian gradient function calculates the sum of the gray level differences of a target point and 4 points around the target point, and then performs a square sum operation.

The Laplacian gradient function based calculation is as follows:

F＝∑ _x ∑ _y [4f(x，y)-f(x，y+1)-f(x，y－1)－f(x+1，y)－f(x－1，y)] ² (5)

accordingly, in the working interval W illustrated in fig. 1, each identification interval can obtain a corresponding image data, and then a corresponding numerical value can be obtained through a Laplacian (Laplacian) gradient function, and an optimal value is selected from the numerical values, where the identification interval corresponding to the numerical value is an optimal working distance of the detection machine 300 to the detection point.

Yet another algorithm based on a gray scale function, for example: the Brenner function is used as the basis for judging the definition of an image by calculating the gray scale difference between pixel points which are separated by two units in the horizontal direction.

The calculation based on the Brenner function is as follows:

F＝∑ _x ∑ _y [f(x+2，y)-f(x，y)] ² (6)

accordingly, in the working interval W illustrated in fig. 1, each recognition interval can obtain a corresponding image data, and then obtain a corresponding value through the Brenner function, and select the best value among the values, where the recognition interval corresponding to the value is the best working distance of the inspection machine 300 to the inspection point.

In the aforementioned exemplary algorithm or other algorithms for evaluating the image sharpness (whether in focus or not), it can be evaluated based on the evaluation result which one of the images obtained by the inspection machine 300 is in focus (the inspection surface can fall within the range of depth of field), and the in focus image is used as the defect judgment basis of the subsequent device under test at the inspection point.

In the above-mentioned finding method of the quasi-focal plane, for each detection point, the detector 300 must obtain the respective image of each recognition interval in the working interval W during the detection, so as to complete the sampling data required at each detection point for the subsequent determination of the quasi-focal plane. However, the optical inspection technique disclosed in the present embodiment can allow the inspection machine 300 to obtain only a single image data at each inspection point, and can determine where the inspection machine 300 should be moved at the inspection point, i.e. can directly determine the in-focus recognition interval of the inspection machine 300 at the inspection point, thereby completing the focusing operation.

Referring to fig. 1 and fig. 2, fig. 2 is a diagram illustrating a relationship between a quantization index M at a detection point and a focal plane position P. Fig. 2 shows the Brenner function as an algorithm for evaluating the sharpness of an image (whether it is in focus or not), and the curve shows the case of Lorentzian. Under other algorithms, the curve may also exhibit other types, such as Gaussian (Gaussian) distribution.

At a detection point, as the detector 300 performs image capture of each focal plane position (having a corresponding identification interval) in the first axial direction X, the image capture of each focal plane position can be evaluated by an algorithm to obtain a corresponding numerical value of a quantization index (parameter), which can represent the degree of quasi-focus. In fig. 2, the Brenner function is used as an example of an algorithm for evaluating the image sharpness (whether in-focus or not), and a higher value of the quantization index represents a higher degree of in-focus.

In fig. 2, the more the number of focal plane positions is assigned, the smoother the curve is fitted, and fig. 2 illustrates more focal plane positions (only 4 in fig. 1). The abscissa designations D10, D30, D50 in fig. 2 refer to the positions of the respective focal plane positions in the first axial direction X, and the focal plane positions D30, i.e. the positions of the targets on which the inspection machine 300 needs to be moved at the inspection point, can be obtained from fig. 2 with the best image definition at D30.

Referring to fig. 3, a flowchart of a focusing method according to an embodiment of the invention is shown. In the embodiment disclosed in the present invention, the focusing manner at each detection point is: step S100, obtaining a detection image in a working interval. Next, step S200 is to divide the detected image into a first band image data and a second band image data of two different bands, and obtain a quantization index value corresponding to each of the first band image data and the second band image data in the resolution analysis. Then, step S300 is performed to obtain a ratio between the first band image data and the second band image data according to the respective quantization index values of the first band image data and the second band image data, and find a matched location parameter in the quantization index ratio lookup table according to the ratio. And S400, adjusting the focal plane position of the detector according to the position parameter.

In step S200, the quantization indexes corresponding to the respective definitions refer to definition evaluation function values obtained by estimating the image data according to the existing algorithm. The matching in step S300 is defined as that the calculated ratio is exactly a specific ratio in the quantization index ratio lookup table, or the calculated ratio falls between two specific ratios in the quantization index ratio lookup table, which can be used to estimate the corresponding position parameter according to a ratio, for example, by interpolation.

The quantization index ratio lookup table is pre-established comparison data, according to the pre-established comparison data, the detector can be quickly matched with a target position to be adjusted, and in a detection program, the detector can be used for judging the position of a focal plane only by taking an image once for each detection point. This is because each quantization index ratio has a quantifiable correspondence with the correct focal plane position, and the embodiment of the present invention completes a high-efficiency detection procedure through this correspondence. In other words, the quantization index ratios can correspond to the current focal plane position of the detector, and accordingly, the plurality of identification intervals are assigned with respective quantization index ratios. The quantization index ratio calculated by the current focal plane position of the detection machine and the quantization index ratio calculated by the reference focal plane position correspond to the difference degree of the current focal plane position of the detection machine and the reference focal plane position on the real space.

FIG. 4 is a schematic view of the focusing points of light rays with different wave bands in the axial direction. When taking a color image, the optical system L can image the object to be measured at the focal plane position P onto the image sensor at the rear end. Each pixel of the image sensor is usually divided into three channels of red light, green light and blue light for sensing, so as to form a color representation of each pixel. When the light incident on the image sensor is divided into different wavelength bands for respective observation, the light of the different wavelength bands has respective corresponding focal plane positions (i.e., the quasi-focal planes) at the quasi-focal time due to the axial chromatic aberration of the optical system L. That is, when the detector performs image capture at a predetermined position and the image is viewed by the image data of different bands, the image data of different bands will have different definitions.

As shown in fig. 4, in the axial direction X1, when the optical system L is fixed, the quasi-focal plane positions of the blue light B and the red light R are respectively located at the upper and lower sides of the quasi-focal plane position of the green light G. That is, when the green light G is in focus, the image data of the red light R and the blue light B are respectively calculated by the algorithm to show that the quantization indexes thereof are not in focus (the definition is low). In an example of an algorithm for evaluating whether the image is in focus using the Brenner function, the quantization index values of the red R and blue B image data are lower than the quantization index values of the green G image data because they are not in focus.

Referring to fig. 1 and 5, fig. 5 is a graph showing the relationship between the quantization index M, the quantization index ratio r and the focal plane position P of light rays in different wavelength bands at a detection point. Fig. 5 illustrates the case where blue light B and red light R are present at quantization index M, quantization index ratio R, and focal plane position P. The focal plane position P on the abscissa indicates a focal plane position to which the detector 300 can be adjusted within the working range W (see fig. 1) in the first axial direction X.

It is understood that the blue light B and the red light R have different quantization index values at the same focal plane position P. Therefore, the ratio of the quantization indexes corresponding to the blue light B and the red light R at different positions in the working interval W will show different values (each position may correspond to a quantization index ratio). The focal plane position of a focus can be predetermined within the working interval W, for example: and placing a reference object to be measured on the focal plane position, so that the detector can capture images of the reference object to be measured which is always located at a fixed position in each identification interval. After the image capture of the focal plane position of the focus is finished, the quantization index ratio of the blue light B and the red light R corresponding to the position is used as a reference ratio, and the quantization index ratio required to be met during the focus calibration is further defined. Since the working area W is divided into a plurality of identification areas, each identification area can be defined with a corresponding quantization index ratio, and accordingly, the difference between the reference ratio and the corresponding quantization index ratio can be used as a basis for how much the detector 300 should be adjusted in the first axis direction X.

For example, in the preset stage, assuming that the reference ratio is defined as 1, it can be correspondingly defined that each difference of 0.1 (or other values) of the quantization index ratio requires to turn up or down the detector 300 by a preset distance, which is determined by how much distance the current focal plane position of the detector 300 is from the in-focus focal plane position in the preset stage (not in the detection procedure running on the production line), i.e. the aforementioned position parameters. Accordingly, in the detection stage (the detector 300 performs image capture in the working space W at each detection point), if the ratio of the quantization indexes obtained by actual detection is 1.2, it means that the detector 300 needs to be adjusted up or down by 2 times of the preset distance from the current position. Therefore, after a quantization index ratio lookup table of the focal plane position in the working interval W corresponding to each quantization index ratio is pre-established in a predetermined working interval W, a matching focal plane position (and a difference degree between the quantization index ratio and a reference ratio) can be found in the quantization index ratio lookup table according to the currently actually measured quantization index ratio during subsequent detection, so as to obtain a corresponding position parameter, which indicates a degree of the detector 300 being apart from the reference focal position in the first axial direction X, and the detector 300 can be correspondingly adjusted accordingly to complete focusing. Where the reference ratio value defined as 1 is an expected value, in other embodiments, the expected value may be defined by the user.

In the example with blue light B and red light R, when the measured quantization index ratio (e.g. R =0.96, 0.88, 0.8) is smaller than the reference ratio, it is representative that the control direction of the detector 300 should be decreased in the first axis X; on the other hand, when the measured quantization index ratio (e.g., r =1.19, 1.12, 1.04) is greater than the reference ratio, it represents that the control direction of the detector 300 should be raised in the first axial direction X. The distance parameter to be raised or lowered is determined according to the quantization index ratio lookup table, i.e., according to the pre-established matching data (the relationship between the difference of quantization index ratios and the position).

In an embodiment of the invention, a denominator of the ratio is a quantization index value corresponding to the second band image data, and a numerator of the ratio is a quantization index value corresponding to the first band image data. The first band of image data is based on light having a first wavelength, the second band of image data is based on light having a second wavelength, and the second wavelength is configured to be longer than the first wavelength, such as the second wavelength is red light and the first wavelength is blue light.

In the embodiment of the invention, the symmetry of the quantization index is destroyed by two light waves by applying the characteristic of axial chromatic aberration, and key information required by matching type focusing is further extracted from the characteristic. The method utilizes different differences (r values) of light rays with different wave bands corresponding to defocusing or in-focus states on quantization index values, and further can apply matching relations in a preset stage (a construction stage of a quantization index ratio lookup table) to data in actual detection, so that a detection machine can perform a focusing procedure by only shooting one image under the operation of the matching type focusing mechanism, and further shoot a second image for subsequent detection and analysis after in-focus (but if the first image is shot and is judged to be in-focus, the second image can be directly used for detection and analysis without shooting the second image again). Compared with the prior art in which a plurality of images are required to be taken to complete the focusing process during actual detection, the focusing method using axial chromatic aberration disclosed in the embodiment of the present invention has significantly improved efficiency.

The focusing method using axial chromatic aberration disclosed in the embodiment of the present disclosure can be applied to a detector (having a zoom mechanism) having a zoom element with adjustable diopter, and to a detector for moving focus in a conventional machine. In the above embodiments, the conventional mechanical focus-shifting detector is used for illustration, and in the case of the detector with the zoom mechanism, the detector can directly adjust the diopter to complete focusing (i.e. change the difference degree of the focal plane positions to the diopter adjustment range) when adjusting the focal plane position according to the matched position parameters.

FIG. 6 is a schematic view of an optical inspection apparatus according to an embodiment of the present invention. The optical detection apparatus under the optical detection architecture of the present embodiment includes: an objective lens assembly 310, a light source assembly 320, a guiding assembly 330, an adjusting assembly 340, a sensor assembly 350, and a control host 400.

The lower end of the guiding element 330 is coupled to the objective lens assembly 310, and the side of the guiding element 330 is coupled to the light source assembly 320. The guiding element 330 can guide the illumination light generated by the light source assembly 320 through the half mirror 331 so that the illumination light is emitted outward through the objective lens assembly 310 and further illuminates the device 200 to be tested carried on the carrying base 100, and in addition, the guiding element 330 can guide the reflection light from the device 200 to be tested to the upper end of the guiding element 330 through the penetrating characteristic of the half mirror 331.

The upper end of the guiding element 330 is coupled to the adjusting element 340. The adjustment assembly 340 allows the reflected light to pass through. The adjusting component 340 may be a diopter-adjustable adjusting component and controlled by the control host 400, or may be a mechanical moving component (e.g. moving in the aforementioned first axial direction) and controlled by the control host 400 to provide the focal plane position adjusting function of the optical detection device. The sensor element 350 is configured with an image sensor therein, and is coupled to the upper end of the adjustment element to receive the reflected light and generate image data through the image sensor.

The control host 400 is coupled to the adjustment assembly 340 and the sensor assembly 350 to receive the image data generated by the sensor assembly 350 and to control the adjustment assembly 340 to adjust the focal plane position formed by the objective lens assembly 310. The control host 400 may also couple the light source assembly 320 for control of the light source. The light source module 320 can be used for simultaneously providing two different wavelength bands of illumination light. For example, one of the bands may be red light, and the other band may be blue light.

In summary, the present invention uses the characteristics of axial chromatic aberration to obtain the key information required by matching focusing, and uses the different differences in the quantization index values when different wave bands of light correspond to the defocus or focus state, so as to apply the matching relationship in the preset stage to the data in the actual detection, so that the detection machine can perform the focusing procedure by only shooting one image in the actual detection procedure, thereby greatly improving the detection efficiency, and being suitable for the online detection of mass production (HVM).

While preferred embodiments of the present invention have been disclosed in the foregoing, it will be understood by those skilled in the art that the examples herein are for the purpose of illustration only and are not to be construed as limiting the scope of the invention. It should be noted that all changes and substitutions equivalent to the embodiments are understood to be included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Reference numerals

100. Bearing seat

200. Device under test

210. The first part

220. The second part

300. Detection machine

d1 The first identification interval

d2 The second identification interval

d3 The third identification interval

d4 The fourth identification interval

D10 Location of focal plane

D30 Location of focal plane

D40 Focal plane position

M quantization index

Focal plane position of P

R red light

G green light

B blue light

L-shaped optical system

r quantization index ratio

First X axial direction

S100 to S400.

Claims (9)

1. A focusing method using axial chromatic aberration, which is based on a pre-established quantization index ratio lookup table and enables a detector to finish focusing on a detection point in a first axial direction in a detection program, comprises the following steps:

making the detecting machine obtain a detecting image in a working interval on the detecting point;

dividing the detected image into a first wave band image data and a second wave band image data of two different wave bands, and obtaining a quantization index value corresponding to each of the first wave band image data and the second wave band image data in definition analysis;

obtaining a ratio between two values according to respective quantization index values of the first band image data and the second band image data, and searching a matched position parameter in the quantization index ratio lookup table according to the ratio; and

adjusting the focal plane position of the detector according to the position parameters,

the quantization index ratio lookup table is used for acquiring a quantization index ratio corresponding to the two different wavelength bands and difference degree information between the current focal plane position of the detector and the reference focal plane position based on the image of the detector in each identification interval according to a reference focal plane position in the working interval and a plurality of identification intervals distinguished from the working interval, wherein the difference degree information is used as the position parameter.

2. The focusing method of claim 1, wherein a denominator of the ratio is the quantization index value corresponding to the second band of image data, a numerator of the ratio is the quantization index value corresponding to the first band of image data, the first band of image data is based on a first wavelength of light, and the second band of image data is based on a second wavelength of light.

3. The focusing method of claim 2, wherein the second wavelength is longer than the first wavelength.

4. The focusing method of claim 3, wherein when the ratio is not equal to 1, in the step of adjusting the focal plane position of the detector, the focal plane position of the detector is adjusted toward the corresponding focal plane position in the quantization index ratio look-up table where the quantization index ratio is 1.

5. The focusing method of claim 3, wherein the second wavelength is red light and the first wavelength is blue light.

6. The focusing method of claim 5, wherein when the ratio is smaller than the quantization index ratio corresponding to the reference focal plane position, the control direction of the detector is decreased in the first axis direction in the step of adjusting the focal plane position of the detector; when the ratio is larger than the quantization index ratio corresponding to the reference focal plane position, the control direction of the detector is lifted in the first axial direction in the step of adjusting the focal plane position of the detector.

7. An optical inspection apparatus for inspecting a device under test, comprising:

an objective lens assembly;

a light source assembly;

a guiding assembly coupled to the objective lens assembly at a lower end and coupled to the light source assembly at a side edge, the guiding assembly guiding the illumination light generated by the light source assembly to illuminate toward the objective lens assembly, and the guiding assembly guiding the reflection light from the device under test to an upper end of the guiding assembly;

the adjusting component is coupled to the upper end of the guide component to receive the reflected light and allow the reflected light to pass through and is used for adjusting the position of a focal plane formed by the objective lens component;

the sensor component is coupled to the upper end of the adjusting component to receive the reflected light to generate image data; and

a control host coupled to the adjustment component and the sensor component for receiving the image data generated by the sensor component and controlling the adjustment component to adjust the focal plane position,

the control host is used for executing the focusing method applying axial chromatic aberration according to any one of claims 1 to 6, and the control host controls the adjusting component to adjust the focal plane position according to a stored quantization index ratio lookup table and image data generated by the sensor component.

8. The optical inspection apparatus of claim 7, wherein the guiding assembly guides the illumination light generated by the light source assembly and the reflected light through a half mirror.

9. The optical inspection device of claim 8 wherein the light source module is configured to provide two distinct wavelength bands of illumination light simultaneously, one wavelength band being red light and the other wavelength band being blue light.

Images