JP7510064B2 - Crack detection apparatus and method - Google Patents

Description

The present invention relates to a crack detection device and method, and to a crack detection device and method for non-destructively detecting the depth position of a crack formed inside a workpiece.

Conventionally, there is known a laser processing device (also called a laser dicing device) that irradiates a laser beam along a planned dividing line by focusing the beam inside a workpiece such as a semiconductor wafer, and forms a laser processing area that serves as the starting point for cutting inside the workpiece along the planned dividing line. The workpiece with the laser processing area formed therein is then divided along the planned dividing line by a dividing process such as expanding or breaking, and separated into individual chips (see, for example, Patent Document 1).

However, when a laser processing area is formed in a workpiece using a laser processing device, a crack propagates from the laser processing area in the thickness direction of the workpiece. The crack formed inside the workpiece becomes the starting point when the workpiece is divided, so the degree to which the crack propagates affects the quality of the chips after the workpiece is divided.

Therefore, after the laser processing area is formed by the laser processing device, by detecting the depth (depth position) of the crack formed inside the workpiece before the cutting process, it is possible to predict whether the workpiece will be successfully cut into chips during the cutting process.

Patent Document 1 discloses a crack detection device that detects the depth of a crack formed inside a workpiece by illuminating the workpiece with a detection light in an oblique manner and receiving the light reflected from the workpiece.

JP 2017-133997 A

When a crack is formed, the processing line may meander due to vibrations of the laser processing device itself, causing the crack to be inclined relative to the thickness direction of the workpiece. If the crack that is the starting point of the division is greatly inclined, the workpiece will be divided along the crack at a large inclination relative to the thickness direction. As a result, there will be partial deviations in the thickness of the resulting chips, which may cause problems such as partial chipping of the chips. Therefore, in order to divide the workpiece well, it is desirable for the crack that is the starting point of the division to be formed in a straight line along the thickness direction of the workpiece.

However, although the crack detection device described in Patent Document 1 can detect the depth (depth position) of a crack, it does not take into consideration the inclination of the crack formed inside the workpiece. Therefore, even if a crack formed inside the workpiece is inclined, it is difficult to grasp the crack condition of the crack. This causes a problem in that it is not possible to accurately predict whether the crack will be successfully cut into chips during the cutting process.

The present invention has been made in consideration of the above circumstances, and aims to provide a crack detection device and method capable of detecting the inclination state of a crack in the thickness direction of a workpiece, and ultimately to make it possible to more accurately predict whether or not the workpiece will be successfully divided into chips during the cutting process.

In order to solve the above problems, the crack detection device according to the first aspect of the present invention includes an oblique illumination means for irradiating a workpiece having a crack formed therein with a first detection light and a second detection light from different oblique directions, a first crack detection means for detecting a first crack length, which is the length of the crack in the thickness direction of the workpiece, based on a first reflected light of the first detection light from the workpiece, a second crack detection means for detecting a second crack length, which is the length of the crack in the thickness direction of the workpiece, based on a second reflected light of the second detection light from the workpiece, and a crack inclination state detection means for detecting the inclination state of the crack with respect to the thickness direction of the workpiece, based on the first crack length and the second crack length.

The length of the crack is detected using the first reflected light and the second reflected light of the first detection light and the second detection light irradiated from different oblique directions on the workpiece, and the first crack length and the second crack length are obtained, respectively. At this time, if the crack is inclined toward the irradiation direction of either the first detection light or the second detection light, a difference occurs between the detected first crack length and the second crack length. Using this, the crack detection device according to the first aspect can detect the inclination state of the crack relative to the thickness direction of the workpiece based on the first crack length and the second crack length. In turn, it becomes possible to more accurately predict the success or failure of the separation into chips in the fracture process.

Preferably, the crack inclination state detection means compares the first crack length with the second crack length and determines the inclination direction of the crack based on the comparison result. For example, if the crack is inclined toward the irradiation direction of the first detection light out of the first and second detection lights, the detected first crack length will be shorter than the detected second crack length. Using this, the crack inclination state detection means can determine the inclination direction of the crack based on the comparison result between the first crack length and the second crack length.

Preferably, the crack inclination state detection means determines that the crack is inclined when the difference between the first crack length and the second crack length is greater than a first threshold value. Preferably, the crack inclination state detection means determines that the crack is not inclined when the difference between the first crack length and the second crack length is less than a second threshold value.

Preferably, when the incident angle of the first detection light is α degrees, the incident angle of the second detection light is β degrees, the first crack length is KL(B), and the second crack length is KL(C), the crack inclination state detection means calculates the inclination angle θ of the crack using the following formula (1).

Preferably, when the incidence angle of the first detection light and the incidence angle of the second detection light are the same angle α degrees, the inclination state detection means calculates the inclination angle θ of the crack using the following formula (2).

In this way, when the incidence angle of the first detection light and the incidence angle of the second detection light are known, the crack inclination state detection means can calculate the inclination angle θ of the crack from equation (1) or equation (2) using the first crack length and the second crack length. This in turn makes it possible to more accurately predict whether the chip will be successfully cut in the cutting process.

In order to solve the above problem, the crack detection method according to the second aspect of the present invention includes an oblique illumination step of irradiating a workpiece having a crack formed therein with a first detection light and a second detection light from different oblique directions, a first crack detection step of detecting a first crack length, which is the length of the crack in the thickness direction of the workpiece, based on a first reflected light of the first detection light from the workpiece, a second crack detection step of detecting a second crack length, which is the length of the crack in the thickness direction of the workpiece, based on a second reflected light of the second detection light from the workpiece, and a crack inclination state detection step of detecting the inclination state of the crack with respect to the thickness direction of the workpiece, based on the first crack length and the second crack length. The crack detection method according to the second aspect can also achieve the same effect as the crack detection device according to the first aspect.

The present invention makes it possible to detect the inclination state of the crack, making it possible to more accurately predict whether the chips will be successfully separated during the cutting process.

FIG. 1 is a block diagram showing a crack detection device according to an embodiment of the present invention. FIG. 2 is an explanatory diagram showing a state in which the workpiece is subjected to oblique illumination with detection light. FIG. 3 is an explanatory diagram showing a state in which the workpiece is subjected to oblique illumination with detection light. FIG. 4 is an explanatory diagram showing a state in which the workpiece is subjected to oblique illumination with detection light. FIG. 5 is a diagram showing the state of reflected light received by a photodetector (corresponding to FIG. 2). FIG. 6 is a diagram showing the state of reflected light received by a photodetector (corresponding to FIG. 3). FIG. 7 is a diagram showing the state of reflected light received by a photodetector (corresponding to FIG. 4). FIG. 8 is a diagram for explaining the path along which reflected light from the workpiece reaches the objective lens pupil. FIG. 9 is a functional block diagram of the control unit. FIG. 10 is a diagram for explaining the principle of detecting the inclination angle of a crack. FIG. 11 is a diagram for explaining the principle of detecting the inclination angle of a crack. FIG. 12 is a diagram showing the relationship between the length of a crack detected in the first region, the length of a crack detected in the second region, and the inclination angle of the crack. FIG. 13 is a flow chart showing the flow of a method for detecting the inclination state of a crack K formed inside a wafer W. FIG. 14 is a graph showing an example of the results of actual crack detection in the cases where the crack K is not inclined with respect to the thickness direction of the wafer W and where it is inclined.

Below, an embodiment of a crack detection device and method according to the present invention will be described with reference to the attached drawings. First, an overview of the crack detection device will be described.

[Crack detection device]
FIG. 1 is a block diagram showing a crack detection device according to an embodiment of the present invention.

The crack detection device 10 is a device used in combination with a laser processing device (not shown) that forms a laser processing area inside the workpiece, the wafer W, and is, for example, provided so as to be movable integrally with the processing head of the laser processing device. In the following explanation, the components related to the crack detection device 10 will be explained, and the configuration of the laser processing device will not be explained.

The crack detection device 10 according to this embodiment detects the crack depth of a crack K formed inside a wafer W, such as a silicon wafer, by irradiating a detection light L1 onto the wafer W and detecting reflected light L2 from the wafer W. In the following description, a three-dimensional orthogonal coordinate system is used in which the stage 510 on which the wafer W is placed is a plane parallel to the XY plane, and the Z direction is the thickness direction of the wafer W. In addition, in FIG. 1, the wafer W is placed on the stage 510 so that the lower surface Wa (the surface in contact with the stage 510) of the wafer W becomes the front surface (device surface) of the wafer W, and the upper surface Wb of the wafer W becomes the back surface (the surface opposite to the device surface) of the wafer W.

As shown in FIG. 1, the crack detection device 10 according to this embodiment includes a light source unit 100, an illumination optical system 200, an interface detection optical system 300, a crack detection optical system 400, a control unit 500, a focal point position moving mechanism 502, an objective lens 504, an operation unit 506, and a display unit 508.

The light source unit 100 emits detection light L1. The detection light L1 is used to detect the interface position of the wafer W and to detect cracks K formed inside the wafer W. Here, when the wafer W is a silicon wafer, the detection light L1 is light that is transparent to the wafer W, for example, infrared light with a wavelength of 1,000 nm or more.

The light source unit 100 includes light sources 102A, 102B, and 102C and a half mirror 104. The light sources 102A, 102B, and 102C and the half mirror 104 are arranged along a main optical axis AX that is coaxial with the lens optical axis of the objective lens 504.

The light sources 102A, 102B, and 102C emit detection light L1 along the main optical axis AX. For example, a laser light source (infrared laser light source, laser diode) or an LED (Light Emitting Diode) light source can be used as the light sources 102A, 102B, and 102C.

Light source 102A has a laser aperture capable of illuminating substantially the entire surface of objective lens pupil 504a of objective lens 504. Light source 102A is used for interface detection, which will be described later.

Light sources 102B and 102C each have a laser aperture capable of illuminating only a portion of the objective lens pupil 504a of the objective lens 504 that is decentered from the main optical axis AX (lens optical axis). Light sources 102B and 102C are used for crack detection, which will be described later.

In this embodiment, the opening for interface detection (light source 102A) and the opening for crack detection (light sources 102B and 102C) are provided separately, but the present invention is not limited to this. For example, a single opening may be used for both purposes, and a light blocking means may be used to switch between the opening for interface detection and the opening for crack detection.

The half mirror 104 reflects the detection light L1 emitted from the light source 102A for interface detection and transmits the detection light L1 emitted from the light sources 102B and 102C for crack detection. Although not shown in the figures, the detection light L1 emitted from the light sources 102A, 102B, and 102C will be referred to as L1(A), L1(B), and L1(C), respectively.

In this embodiment, a total reflection mirror or a dichroic mirror can be used instead of the half mirror 104. In this case, a mirror is inserted in the position of the half mirror 104 on the optical path when detecting an interface, and the mirror is removed from the optical path when detecting a crack.

Light sources 102A, 102B, and 102C are each connected to a control unit 500, which controls the emission of light sources 102A, 102B, and 102C.

The control unit 500 is realized by, for example, a personal computer, a workstation, a PLC (Programmable Logic Controller), etc. The control unit 500 includes a CPU (Central Processing Unit) that controls the operation of each part of the crack detection device 10, a ROM (Read Only Memory), a storage device (e.g., an HDD (Hard Disk Drive) or an SSD (Solid State Drive)) that stores a control program, and an SDRAM (Synchronous Dynamic Random Access Memory) that can be used as a working area for the CPU. The control unit 500 accepts operation input by an operator via the operation unit 506, and transmits control signals corresponding to the operation input to each part of the crack detection device 10 to control the operation of each part.

The operation unit 506 is a means for accepting operation input by an operator, and includes, for example, a keyboard, a mouse, or a touch panel.

The display unit 508 is a device that displays an operation GUI (Graphical User Interface) for operating the crack detection device 10 and images (e.g., crack detection results, etc.). For example, a liquid crystal display can be used as the display unit 508.

The illumination optical system 200 guides the detection light L1 emitted from the light source unit 100 to the objective lens 504. The illumination optical system 200 includes relay lenses 202 and 206 and a mirror 204 (e.g., a total reflection mirror).

The detection light L1 emitted from the light source unit 100 passes through the relay lens 202 and is reflected by the mirror 204, causing the optical path to be bent. The detection light L1 reflected by the mirror 204 passes through the relay lens 206, and is then reflected in sequence by the half mirror 304 and the half mirror 302, and is emitted toward the objective lens 504.

The return light (observation light) reflected by the wafer W and passing through the half mirror 302 can be observed using the observation optical system 600 (e.g., a photodetector, etc.). If the observation optical system 600 is not used, a dichroic mirror or a total reflection mirror can be used instead of the half mirror 302.

The objective lens 504 collects (focuses) the detection light L1 emitted from the illumination optical system 200 onto the wafer W. The objective lens 504 is disposed at a position facing the wafer W and is coaxial with the main optical axis AX.

The focal point position moving mechanism 502 changes the position of the focal point of the detection light L1 in the Z direction (the optical axis direction of the objective lens 504). The focal point position moving mechanism 502 includes an actuator (not shown) that moves the objective lens 504 in the Z direction. Here, a case where a piezo actuator is used will be described as an example, but any actuator can be used. The focal point position moving mechanism 502 moves the objective lens 504 in the Z direction by driving the piezo actuator under the control of the control unit 500. This changes the relative distance between the objective lens 504 and the wafer W in the Z direction, thereby adjusting (fine-tuning) the position of the focal point of the detection light L1 in the Z direction.

The focal point position moving mechanism 502 may also include a Z drive mechanism that moves the crack detection device 10 in the Z direction relative to the stage 510. The Z drive mechanism moves the crack detection device 10 in the Z direction to perform Z-direction alignment (coarse adjustment) between the objective lens 504 and the wafer W with a larger adjustment range than the piezoelectric actuator.

As described above, when the position adjustment (coarse adjustment) of the focal point by the Z drive mechanism is combined with the position adjustment (fine adjustment) of the focal point by the piezoelectric actuator, the degree of freedom (adjustment range) of adjustment of the Z-direction position of the focal point of the detection light L1 is increased compared to the case of using only a piezoelectric actuator. This makes it possible to detect cracks, etc., on wafers W of various thicknesses.

The Z drive mechanism may be a mechanism for driving the stage 510 in the Z direction, or a mechanism for driving both the crack detection device 10 and the stage 510 in the Z direction. The Z drive mechanism may also serve as a drive mechanism for moving the processing head of the laser processing device.

The reflected light L2 collected by the objective lens 504 and reflected by the wafer W is guided to the interface detection optical system 300 and the crack detection optical system 400, and is used to detect the interface and cracks of the wafer W, respectively.

[Crack detection procedure]
In this embodiment, an example is described in which the interface position of the underside Wa of the wafer W (the surface in contact with the stage 510, for example, the front surface (device surface) of the wafer W) is detected, and then the crack depth is detected based on the interface position of the underside Wa of the wafer W.

In this embodiment, the crack depth is detected based on the bottom surface Wa of the wafer W, but the present invention is not limited to this. For example, the crack depth may be detected based on the top surface Wb of the wafer W (the back surface of the wafer W), or the average value of the crack depths detected based on the interface positions of both the bottom surface Wa and the top surface Wb of the wafer W may be taken.

[Interface detection optical system]
First, an outline of the interface detection of the wafer W using the confocal focus mechanism will be described, which will be described in more detail later.

The interface detection optical system 300 is an optical system for detecting the interface (lower surface Wa or upper surface Wb) of the wafer W, and includes a half mirror 302, a half mirror 304, a relay lens 306, a half mirror 308, and a photodetector 310.

When detecting the underside Wa of the wafer W as the interface of the wafer W, the control unit 500 causes the light source 102A to emit light and irradiate the wafer W with detection light L1(A). Here, the control unit 500 and the interface detection optical system 300 each function as part of the interface detection means.

The detection light L1(A) (first detection light) from the light source 102A is a laser light having an aperture of approximately the same size as the objective lens pupil 504a of the objective lens 504, and is reflected by the half mirror 304 and the half mirror 302 in sequence and guided to the objective lens 504.
A) is irradiated onto substantially the entire surface of the objective lens pupil 504 a of the objective lens 504 .

Here, the reflected light of the detection light L1(A) reflected by the wafer W is referred to as L2(A) (first reflected light). The reflected light L2(A) is reflected by the half mirror 302, passes through the half mirror 304, and is then guided to the relay lens 306. The reflected light L2(A) that passes through the relay lens 306 is reflected by the half mirror 308 and is then guided to the photodetector 310.

The photodetector 310 is a device for receiving reflected light L2(A) from the wafer W and detecting the interface of the wafer W, and includes a detector body 310A and a pinhole panel 310B.

The detector body 310A may be a photodetector (e.g., a photodiode) that converts received light into an electrical signal and outputs it to the control unit 500, or an infrared camera, etc.

A pinhole is formed in the pinhole panel 310B to transmit a portion of the incident light. The pinhole panel 310B is disposed upstream of the light receiving surface of the detector body 310A, and is disposed so that the pinhole of the pinhole panel 310B is located on the optical axis of the reflected light L2(A). The position of the pinhole of the pinhole panel 310B is optically conjugate with the focal point (front focal position) of the objective lens 504 (confocal pinhole). The size of the pinhole of the pinhole panel 310B is adjusted to approximately the diffraction limit of the objective lens 504.

The reflected light L2(A) reflected by the wafer W is focused at the pinhole position of the pinhole panel 310B, which is optically conjugate with the focusing point of the objective lens 504. When the focusing point of the objective lens 504 coincides with the lower surface Wa of the wafer W, which serves as the reflecting surface, the light beam of the detection light L1(A) is reflected by the lower surface Wa of the wafer W, becomes a parallel light beam, and passes through the objective lens 504 and returns. Therefore, the signal S output from the detector body 310A has a sharp peak when the focusing point of the objective lens 504 coincides with the position of the lower surface Wa of the wafer W, which serves as the reflecting surface.

The control unit 500 changes the relative distance between the objective lens 504 and the wafer W with the focus point position moving mechanism 502 while irradiating the wafer W with the detection light L1(A) from the light source 102A, and moves the position of the focus point of the detection light L1(A) (i.e., the front focal position of the objective lens 504) in the Z direction. This causes the focus point of the detection light L1(A) to be scanned in the Z direction. The control unit 500 detects the reflected light L2(A) from the wafer W when the focus point of the detection light L1(A) is scanned in the Z direction with the photodetector 310, and detects the interface position Z(0) of the underside Wa of the wafer W by detecting the peak of the signal from this photodetector 310.

In this embodiment, the interface of the wafer W is detected using the confocal method, but the present invention is not limited to this. For example, other focus detection methods such as the astigmatism method and the white light interferometry may be used.

[Optical system for crack detection]
Next, a description will be given of detection of a crack K formed inside the wafer W. For ease of understanding, the description will be given without taking into consideration the inclination of the crack K with respect to the thickness direction of the wafer W.

The crack detection optical system 400 includes a relay lens 402 and photodetectors 404 and 406 .

When detecting a crack K formed inside the wafer W, the control unit 500 causes the light sources 102B and 102C to emit light, and irradiates the wafer W with detection light L1(B) and L1(C) (second detection light). Here, the control unit 500 and the crack detection optical system 400 each function as part of the crack detection means. The light sources 102B and 102C each have a laser aperture at a position offset from the main optical axis AX. As a result, the detection light L1(B) and L1(C) that is eccentric with respect to the main optical axis AX is irradiated onto the wafer W. The light source unit 100, the illumination optical system 200, and the objective lens 504 are an example of a tilted illumination means.

The reflected lights L2(B) and L2(C) (second reflected lights) obtained by reflecting the detection lights L1(B) and L1(C) by the wafer W are reflected by the half mirror 302, and then pass through the half mirror 304, the relay lens 306, and the half mirror 308 in sequence to enter the relay lens 402. The reflected lights L2(B) and L2(C) that pass through the relay lens 402 are received by the photodetectors 404 and 406.

In the interface detection optical system 300, a total reflection mirror or a dichroic mirror can be used instead of the half mirror 308. In this case, a mirror is inserted in the optical path at the position of the half mirror 308 during interface detection, and the mirror is removed from the optical path during crack detection.

The optical detectors 404 and 406 are devices for receiving reflected light L2(B) and L2(C) from the wafer W and detecting cracks K inside the wafer W. The optical detectors 404 and 406 can be photodetectors (e.g., photodiodes) that convert the received light into electrical signals and output them to the control unit 500, or infrared cameras, etc.

The photodetectors 404 and 406 are positioned conjugate with the objective lens pupil 504a, and are further positioned offset from the optical axis of the objective lens 504 so as to receive the detection light L1(B) and L1(C).

Figures 2 to 4 are explanatory diagrams showing the state when the detection light L1 is obliquely illuminated onto the wafer W. Figure 2 shows the case where a crack K exists at the focal point of the objective lens 504, Figure 3 shows the case where a crack K does not exist at the focal point of the objective lens 504, and Figure 4 shows the case where the focal point of the objective lens 504 coincides with the crack depth (bottom end position of the crack) of the crack K.

Figures 5 to 7 show the reflected light L2 received by the photodetectors 404 and 406, and correspond to the cases shown in Figures 2 to 4, respectively.

Figure 8 is a diagram for explaining the path that reflected light L2 from the wafer W takes to reach the objective lens pupil 504a. Note that here, a case is explained in which detection light L1 passes through a first region G1 on one side (the right side in Figure 8) of the objective lens pupil 504a and oblique illumination is performed on the wafer W.

2, when a crack K is present at the focal point of the objective lens 504, the detection light L1 is totally reflected by the crack K, and the reflected light L2 travels along a path on the same side as the optical path of the detection light L1 with respect to the main optical axis AX, and becomes a component that reaches an area on the same side of the objective lens pupil 504a as the detection light L1. That is, as shown in Fig. 8, when the path of the detection light L1 when the detection light L1 from the light source unit 100 is irradiated onto the wafer W via the objective lens 504 is R1, the reflected light L2 totally reflected by the crack K inside the wafer W travels along a path R2 on the same side (right side in Fig. 8) as the path R1 of the detection light L1 with respect to the main optical axis AX and passes through a first area G1 of the objective lens pupil 504a.

As shown in FIG. 3, when no crack K is present at the focal point of the objective lens 504, the detection light L1 is reflected by the underside Wa of the wafer W, and the reflected light L2 is a component that reaches the area of the objective lens pupil 504a on the opposite side to the detection light L1. That is, as shown in FIG. 8, the reflected light L2 reflected by the underside Wa of the wafer W passes through the second area G2 of the objective lens pupil 504a by following a path R3 on the opposite side (left side in FIG. 8) of the main optical axis AX from the path R1 of the detection light L1.

4, when the focal point of the objective lens 504 coincides with the lower end position of the crack K, the detection light L1 is split into a reflected light component L2a and a non-reflected light component L2b. The reflected light component L2a is totally reflected by the crack K, then reflected by the lower surface Wa and reaches an area on the same side of the objective lens pupil 504a as the detection light L1, while the non-reflected light component L2b is not totally reflected by the crack K, but is reflected by the lower surface Wa of the wafer W and reaches an area on the opposite side of the objective lens pupil 504a to the detection light L1. That is, as shown in FIG. 8, of the reflected light L2, the reflected light component L2a that is totally reflected by the crack K inside the wafer W passes through a first region G1 of the objective lens pupil 504a, following a path R2 on the same side of the main optical axis AX as the path R1 of the detection light L1 (the right side in FIG. 8), while the non-reflected light component L2b that is not totally reflected by the crack K and is reflected by the underside Wa of the wafer W passes through a second region G2 of the objective lens pupil 504a, following a path R3 on the opposite side of the main optical axis AX from the path R1 of the detection light L1 (the left side in FIG. 8).

The photodetectors 404 and 406 are arranged so that they are optically conjugate with the first region G1 and the second region G2 of the objective lens pupil 504a, respectively. This allows the photodetectors 404 and 406 to selectively receive light that has passed through the first region G1 and the second region G2 of the objective lens pupil 504a, respectively.

Here, in the example shown in FIG. 2 (where crack K is present at the focal point of objective lens 504), reflected light L2 is incident on light receiving surface 404C of light detector 404, out of light detectors 404 and 406. Therefore, as shown in FIG. 5, the level of the detection signal output from light receiving surface 404C of light detector 404 is higher than the level of the detection signal output from light receiving surface 406C of light detector 406.

On the other hand, in the example shown in FIG. 3 (where crack K is not present at the focal point of objective lens 504), the reflected light is incident on light receiving surface 406C of light detector 406, out of light detectors 404 and 406. Therefore, as shown in FIG. 6, the level of the detection signal output from light receiving surface 406C of light detector 406 is higher than the level of the detection signal output from light receiving surface 404C of light detector 404.

In the example shown in FIG. 4 (where the focal point of the objective lens 504 coincides with the lower end position of the crack K), the components L2a and L2b of the reflected light L2 are incident on the light receiving surfaces 404C and 406C of the photodetectors 404 and 406, respectively. Therefore, as shown in FIG. 7, the levels of the detection signals output from the light receiving surfaces 404C and 406C of the photodetectors 404 and 406 are approximately equal.

In this way, the amount of light received by the light receiving surfaces 404C and 406C of the photodetectors 404 and 406 changes depending on whether or not a crack K is present at the focal point of the objective lens 504. In this embodiment, this property is utilized to detect the crack depth (bottom end position or top end position) of a crack K formed inside the wafer W.

Specifically, when the detection signal outputs from the light receiving surfaces 404C and 406C of the photodetectors 404 and 406 are D1 and D2, respectively, the evaluation value S for determining the presence of a crack K at the focal point of the objective lens 504 can be expressed by the following formula.

S = (D1 - D2) / (D1 + D2) ... (1)
In equation (1), when the condition S = 0 is satisfied, that is, when the amount of light received by the light receiving surfaces 404C and 406C of the photodetectors 404 and 406 is the same, this indicates a state in which the focal point of the objective lens 504 and the bottom end position of the crack (or the top end position of the crack) are coincident.

The control unit 500 (see FIG. 1) controls the focal point position moving mechanism 502 (focus point changing means) to move the focal point of the detection light L1 in the Z direction, sequentially changing it from the interface position of the underside Wa of the wafer W in the thickness direction (Z direction) of the wafer W, sequentially acquiring detection signals output from the light receiving surfaces 404C and 406C of the photodetectors 404 and 406, and calculating the evaluation value S shown in equation (1) based on this detection signal. The crack depth (crack bottom end position or crack top end position) of the crack K can be detected by evaluating this evaluation value S and the focal point position information.

The crack depth may be detected based on either surface of the wafer W. For example, the crack depth may be detected based on the top surface Wb of the wafer W, or the bottom surface Wa of the wafer W. It is also possible to take the average value of the crack depths detected based on the interface positions of both the top surface Wb and the bottom surface Wa of the wafer W.

[Functional configuration of the control unit]
Fig. 9 is a functional block diagram of the control unit 500. As shown in Fig. 9, the control unit 500 includes an interface detection unit 520, a first crack detection unit 530, a second crack detection unit 540, and a crack inclination state detection unit 550.

The interface detection unit 520 detects a peak in the detection signal level from the detection signal (focus signal; light intensity signal of reflected light L2(A) from the wafer W) output from the photodetector 310 when the focal point of the detection light L1(A) is scanned in the Z direction, and detects the position of the objective lens 504 corresponding to the peak value as the position of the interface of the wafer W.

The first crack detection unit 530 detects the positions of the upper and lower ends (both ends) Kt and Kb of the crack K based on the change in the level of the detection signal output from the photodetector 404. This detection result corresponds to the positions of the upper and lower ends Kt and Kb of the crack K detected on the first region G1 side. Furthermore, the first crack detection unit 530 detects the length of the crack K in the thickness direction of the wafer W (first crack length) based on the positions of the upper and lower ends Kt and Kb of the crack K detected on the first region G1 side.

The second crack detection unit 540 detects the positions of the upper and lower ends Kt and Kb of the crack K based on the change in the level of the detection signal output from the photodetector 406. This detection result corresponds to the positions of the upper and lower ends Kt and Kb of the crack K detected on the second region G2 side. Furthermore, the second crack detection unit 540 detects the length of the crack K in the thickness direction of the wafer W (second crack length) based on the positions of the upper and lower ends Kt and Kb of the crack K detected on the second region G2 side.

The crack inclination state detection unit 550 detects the inclination state of the crack K relative to the thickness direction of the wafer W based on the first crack length detected by the first crack detection unit 530 and the second crack length detected by the second crack detection unit 540.

More specifically, the crack inclination state detection unit 550 includes an inclination direction determination unit 552 that compares the first crack length detected by the first crack detection unit 530 with the second crack length detected by the second crack detection unit 540, and determines the inclination direction of the crack K relative to the thickness direction of the wafer W based on the comparison result.

Furthermore, the crack inclination state detection unit 550 includes an inclination angle calculation unit 554 that calculates the inclination angle of the crack K relative to the thickness direction of the wafer W based on the incident angle of the detection light L1 (L1(B) and L1(C)) relative to the main optical axis AX, the length of the crack K detected on the first region G1 side, and the length of the crack K detected on the second region G2 side.

The inclination state (inclination direction and inclination angle) of the crack K detected by the crack inclination state detection unit 540 (inclination direction determination unit 552 and inclination angle calculation unit 554) is output to the display unit 508.

[Detection principle of crack inclination angle]
Next, the principle of detecting the inclination angle and inclination direction of the crack K performed by the crack inclination state detection unit 540 and the method of calculating the inclination angle will be described in detail. First, the principle of detecting the inclination angle of the crack will be outlined with reference to Fig. 10 and Fig. 11. Fig. 10 is a diagram showing the geometric positional relationship between the crack K and the detection light L1(B) and L1(B) when the crack K is not inclined with respect to the thickness direction of the wafer W (Z direction, parallel to the main optical axis AX).

In FIG. 10, reference numeral 10A indicates the overall positional relationship of the crack K, the detection light L1(B), and L1(B) relative to the objective lens 504, reference numeral 10B indicates in more detail the positional relationship of the crack K and the detection light L1(B) relative to the objective lens 504 (the positional relationship relative to the first region G1 side of the objective lens pupil 504a), and reference numeral 10C indicates in more detail the positional relationship of the crack K and the detection light L1(C) relative to the objective lens 504 (the positional relationship relative to the second region G2 side of the objective lens pupil 504a). It is assumed that the detection light L1(B) is irradiated at a known angle of incidence α (degrees) relative to the main optical axis AX (which overlaps with the optical axis of the objective lens 504), and the detection light L1(C) is irradiated at a known angle of incidence β (degrees) relative to the main optical axis AX.

As explained using Figures 2 to 4, in crack detection, the objective lens 504 and the wafer W are aligned so that the crack K is on the main optical axis AX, and detection light L1 (L1(B) and L1(C)) that is off-centered with respect to the main optical axis AX is irradiated onto the wafer W. Then, the control unit 500 moves the objective lens 504 in the thickness direction (Z direction) of the wafer W, and detects reflected light L2 (L2(B) and L2(C)) at the position of the focal point of each objective lens 504.

As shown by symbols 10B and 10C in FIG. 10, when the crack K is not inclined with respect to the thickness direction of the wafer W, the upper and lower ends Kt and Kb of the crack K formed inside the wafer W exist on the main optical axis AX, and the entire crack K almost completely overlaps with the main optical axis AX. Therefore, as shown by symbols 10B and 10C, the positions Pt(B) and Pb(B) of the upper and lower ends of the crack K detected on the first region G1 side coincide with the positions Pt(C) and Pb(C) of the upper and lower ends of the crack K detected on the second region G2 side.

Therefore, the distance KL(B) between the positions Pt(B) and Pb(B) of the upper and lower ends of the crack K detected on the first region G1 side (which is the length of the crack K detected on the first region G1 side and corresponds to the first crack length) is the same as the distance KL(C) between the positions Pt(C) and Pb(C) of the upper and lower ends of the crack K detected on the second region G2 side (which is the length of the crack K detected on the second region G2 side and corresponds to the second crack length).

Figure 11 shows the geometric positional relationship of the crack K, detection light L1(B), and L1(B) with respect to the objective lens 504 when the crack K is inclined toward the first region G1 at an inclination angle θ (degrees) with respect to the thickness direction of the wafer W. As in Figure 10, in Figure 11, reference numeral 11A indicates the overall positional relationship of the crack K, detection light L1(B), and L1(B) with respect to the objective lens 504, reference numeral 11B indicates the positional relationship with respect to the first region G1 side of the objective lens pupil 504a, and reference numeral 11C indicates the positional relationship with respect to the second region G2 side of the objective lens pupil 504a.

In FIG. 11, the geometrical positional relationship between the objective lens 504, the detection light L1(B), and the detection light L1(C) is the same as in FIG. 10, so the description will be omitted. In FIG. 11, it is assumed that the objective lens 504 and the wafer W are aligned so that the bottom end Kb of the crack K is on the main optical axis AX.

First, the first region G1 side will be described using the reference symbol 11B. When the crack K is inclined with respect to the thickness direction of the wafer W, as shown by reference symbol 11B in FIG. 11, if the position of the intersection of the detection light L1(B) passing through the upper end Kt of the crack K and the main optical axis AX (the optical axis of the objective lens 504) is Pt(B), when crack detection is performed in this embodiment, the position Pt(B) on the main optical axis AX is detected as the position of the upper end Kt of the crack K detected using the detection light L1(B) on the first region G1 side. The detected position Pt(B) at this time is lower than the actual position of the upper end Kt of the crack K.

As shown by the reference symbol 11C, in the case where the crack K is inclined with respect to the thickness direction of the wafer W on the second region G2 side as on the first region G1 side, if the position of the intersection of the detection light L1(C) passing through the upper end Kt of the crack K and the main optical axis AX is Pt(C), when crack detection is performed in this embodiment, the position Pt(C) on the main optical axis AX is detected as the position of the upper end Kt of the crack K detected using the detection light L1(C) on the second region G2 side. The detected position Pt(C) at this time is higher than the actual position of the upper end Kt of the crack K.

In addition, since the lower end Kb of the crack K is aligned on the main optical axis AX, the Z-direction position Pb(B) of the lower end Kb of the crack K detected on the first region G1 side is approximately the same as the position Pb(C) of the lower end Kb of the crack K detected on the second region G2 side.

Therefore, when the crack K is inclined toward the first region G1 with respect to the thickness direction of the wafer W, the distance KL (B) between the positions Pt (B) and Pb (B) of the upper and lower ends of the crack K detected on the first region G1 side is shorter than the distance KL (C) between the positions Pt (C) and Pb (C) of the upper and lower ends of the crack K detected on the second region G2 side.

Note that, conversely, if the crack K is inclined toward the second region G2 with respect to the thickness direction of the wafer W, KL(B) will be longer than KL(C). This is self-evident, so a description of this case will be omitted.

In this way, when the crack K is inclined relative to the thickness direction of the wafer W, a difference occurs between the crack length KL(B) detected on the first region G1 side and the crack length KL(C) detected on the second region G2 side. The control unit 500 can use this difference to detect the inclination state of the crack K (inclination direction and inclination angle).

[Calculation method of crack inclination angle]
Next, a method for calculating the inclination angle of the crack K will be specifically described with reference to Fig. 12. When the crack K and the detection lights L1(B) and L1(B) have the geometric positional relationship shown in Fig. 11 with respect to the objective lens 504, it is possible to create a triangle as shown in Fig. 12 from the lengths KL(B) and KL(C) of the detected crack and the angles α, β, and θ.

Based on the triangle shown in Figure 12, the inclination angle θ of the crack K can be expressed by the following equation (2) using the sine theorem.

Therefore, the inclination angle calculation unit 554 (see FIG. 9) of the crack inclination state detection unit 550 can calculate the inclination angle θ by substituting the known incidence angles α and β, and the detected crack lengths KL(C) and KL(B) into equation (2). Also, as described above, the inclination direction determination unit 552 (see FIG. 9) of the crack inclination state detection unit 550 can detect whether the crack K is inclined toward the first region G1 side or the second region G2 side by determining which of the crack lengths KL(B) and KL(C) is longer.

Here, if the incident angle α of the detection light L1(B) is the same as the incident angle β of L1(C) (α = β), equation (2) can be simplified to the following equation (3):

From this equation (3), it can be seen that when the length of the crack K detected on the first region G1 side and the length of the crack K detected on the second region G2 side are the same, θ = 0 (zero), so it can be confirmed that the crack K is not inclined in the thickness direction.

[Method for detecting the inclination state of a crack]
Next, a method (one example of a crack detection method) for detecting the inclination state of a crack K formed inside the wafer W using the crack detection device 10 of this embodiment will be described. Fig. 13 is a flowchart showing the flow of the method for detecting the inclination state of a crack K formed inside the wafer W.

First, as described with reference to Figures 2 to 8, the interface detection unit 520 detects the position of the interface that serves as the reference for the height position, for example, the position of the underside Wa of the wafer W, based on the detection signal output from the photodetector 310 when the focal point of the detection light L1(A) is scanned in the Z direction (step S10).

Next, detection light L1(B) and L1(C), which are eccentric with respect to the main optical axis AX, are irradiated onto the wafer W from light sources 102B and 102C, respectively (step S12).

Next, the first crack detection unit 520 detects the reflected light L2(B) of the detection light L1(B) reflected by the wafer W with each photodetector 404 of the crack detection optical system 400 while sequentially moving the focal point of the detection light L1(B) relative to the position of the lower surface Wa in the thickness direction of the wafer W. At this time, the first crack detection unit 520 detects the positions Pt(B) and Pb(B) of the upper and lower ends of the crack K as the movement amount (piezo movement amount, objective lens movement amount) from the wafer lower surface, which is the reference, based on the change in the level of the detection signal. Furthermore, the first crack detection unit 520 detects the crack length KL(B), which is the length of the crack K in the thickness direction of the wafer W from the positions Pt(B) and Pb(B) (step S14).

Similarly, the second crack detection unit 530 detects the reflected light L2(C) of the detection light L1(C) reflected by the wafer W with each photodetector 406 of the crack detection optical system 400 while sequentially moving the focal point of the detection light L1(C) relatively in the thickness direction of the wafer W with reference to the position of the lower surface Wa. Then, the second crack detection unit 530 detects the positions Pt(C) and Pt(C) of the upper and lower ends of the crack K based on the level of the detection signal, and further detects the crack length KL(C), which is the length of the crack K in the thickness direction of the wafer W from the positions Pt(C) and Pb(C) (step S16).

Furthermore, the crack inclination state detection unit 550 detects the inclination state of the crack K relative to the thickness direction of the wafer W based on the crack length KL(B) detected by the first crack detection unit 520 and the crack length KL(C) detected by the second crack detection unit 530 (step S18).

More specifically, the crack inclination state detection unit 550 determines that the crack K is inclined in the thickness direction when the difference between the crack length KL(B) and the crack length KL(C) is greater than a predetermined threshold value (first threshold value).

Alternatively, the crack inclination state detection unit 550 may determine that the crack K is not inclined in the thickness direction when the difference between the crack length KL(B) and the crack length KL(C) is smaller than a predetermined threshold value (second threshold value).

In addition, according to the above-described aspect of determining the inclination state of the crack K using the first threshold value and the second threshold value, it is possible to perform an accurate determination even for a state in which the inclination of the crack K is so small that it can be considered that there is essentially no inclination (i.e., a level that does not affect the quality of the chip after division) compared to an aspect of simply comparing whether or not there is a difference between the crack length KL(B) and the crack length KL(C). In addition, the first threshold value and the second threshold value may be the same value or different values.

In addition, the inclination direction determination unit 552 of the crack inclination state detection unit 550 compares the crack length KL(B) with the crack length KL(C) and determines the inclination direction of the crack K, i.e., whether the crack K is inclined toward the first region G1 side or the second region G2 side, based on the comparison result.

Furthermore, if the crack K is inclined, the inclination angle calculation unit 554 of the crack inclination state detection unit 550 can calculate the inclination angle θ of the crack K by substituting the lengths KL(B) and KL(C) of the detected crack K into equation (2) or equation (3).

When the crack K is inclined, the position Pt(B) of the upper end Kt of the crack K detected on the first region G1 side does not coincide with the position Pt(C) of the upper end Kt of the crack K detected on the second region G2 side. However, the crack inclination state detection unit 550 can easily calculate the actual position of the upper end Kt of the crack K using trigonometric functions from the detected length KL(B) or KL(C) of the crack K and the inclination angle θ (see the triangle shown in Figure 12). Therefore, even when the crack K is inclined, the actual depth position and length of the crack K can be accurately calculated.

[Concrete example]
14, a specific example of crack detection in the cases where the crack K is not inclined and is inclined with respect to the thickness direction of the wafer W will be described. In the graphs indicated by reference numerals 14A and 14B in FIG. 14, the horizontal axis represents the movement amount of the objective lens 504 (the piezoelectric movement amount of the focus point position moving mechanism 502) indicating the position of the focus point (the focus point of the objective lens 504) of the detection lights L1(B) and L1(C), and the vertical axis represents the level (intensity) of the detection signals of the photodetectors 404 and 406.

The solid line indicates the detection signal by the photodetector 404, which corresponds to the detection result obtained on the first region G1 side, and the dotted line indicates the detection signal by the photodetector 406, which corresponds to the detection result obtained on the second region G2 side.

In crack detection, the detection light L1(B) and L1(C) are half-blocked by the upper and lower ends Kt and Kb of the crack K, so the position of the objective lens 504 when the level of the detection signal of the photodetectors 404 and 406 is halved (i.e., the difference between the maximum signal level and the minimum signal level is halved) corresponds to the positions of the upper and lower ends Kt and Kb of the crack.

Reference numeral 14A in FIG. 14 is a graph showing the results of crack detection when the crack K is not inclined relative to the thickness direction of the wafer W as shown in FIG. 10. As shown by reference numeral 14A, when the crack K is not inclined relative to the thickness direction of the wafer W, the positions Pt(B) and Pb(B) of the upper and lower ends of the crack K detected on the first region G1 side are approximately the same as the positions Pt(C) and Pb(C) of the upper and lower ends of the crack K detected on the second region G2 side.

Therefore, the length KL(B) of the crack K detected on the first region G1 side is approximately equal to the length KL(C) of the crack K detected on the second region G2 side. As a result, the crack inclination state detection unit 550 can detect that the crack K is not inclined with respect to the thickness direction of the wafer W (the inclination angle θ is approximately 0 degrees).

Reference numeral 14B in FIG. 14 is a graph showing the results of crack detection when the crack K is inclined toward the first region G1 side at an inclination angle θ with respect to the thickness direction of the wafer W as shown in FIG. 11. Even when the crack K is inclined with respect to the thickness direction of the wafer W, the lower end Kb of the crack K is aligned on the main optical axis AX (the optical axis of the objective lens 504), so that the Z-direction position Pb(B) of the lower end Kb of the crack K detected on the first region G1 side, as shown by reference numeral 14B, substantially coincides with the position Pb(C) of the lower end Kb of the crack K detected on the second region G2 side.

However, the position Pt(B) of the upper end Kt of the crack K detected on the first region G1 side does not match the position Pt(C) of the upper end Kt of the crack K detected on the second region G2 side. Therefore, the length KL(B) of the crack K detected on the first region G1 side does not match the length KL(C) of the crack K detected on the second region G2 side. As a result, the crack inclination state detection unit 550 can determine that the crack K is inclined with respect to the thickness direction of the wafer W, for example, because the difference between the lengths KL(B) and KL(C) is greater than a predetermined value (first threshold value).

In addition, the length KL (B) of the crack K detected on the first region G1 side is shorter than the length KL (C) of the crack K detected on the second region G2 side. Therefore, the inclination direction determination unit 552 of the crack inclination state detection unit 550 can detect that the crack K is inclined toward the first region G1 side.

Furthermore, the inclination angle calculation unit 554 of the crack inclination state detection unit 550 can calculate the inclination angle θ of the crack K by substituting the lengths KL(B) and KL(C) of the detected crack K into equation (2) or equation (3).

When the crack K is inclined, the position Pt(B) of the upper end Kt of the crack K detected on the first region G1 side does not coincide with the position Pt(C) of the upper end Kt of the crack K detected on the second region G2 side, but the crack inclination state detection unit 550 can easily calculate the actual position of the upper end Kt of the crack K using trigonometric functions from the detected length KL(B) or KL(C) of the crack K and the inclination angle θ.

The control unit 500 outputs the depth position, inclination direction, and inclination angle of the detected crack K to the display unit 508. The user can determine whether or not the crack K has been successfully formed based on the information output to the display unit 508. For example, the user can accurately predict whether or not the chip will be successfully divided in the fracturing process by taking into account the inclination angle θ in addition to the depth position of the crack K.

For example, the user can predict the position and shape of the cut surface when the wafer W is cut from the depth position, inclination direction, and inclination angle θ of the crack K. Also, for example, the user can predict that even if the depth position of the crack K is appropriate, the wafer W cannot be cut well because the inclination angle θ is too large.

[Modification]
In the above explanation, it is assumed that crack detection is performed with the lower end Wb of the crack K aligned with the main optical axis AX when the crack K is inclined with respect to the thickness direction, but crack detection may also be performed with the upper end Wt of the crack K aligned with the main optical axis AX.

[Effect of the invention]
Thus, according to this embodiment, when a crack is detected, not only the depth position of the crack K but also the inclination state of the crack K with respect to the thickness direction of the wafer W can be detected. Particularly, in this embodiment, it is possible to detect the inclination direction and inclination angle of the crack K as the inclination state of the crack K. This makes it possible to accurately predict whether the chip will be successfully divided in the fracture process.

According to this embodiment, it is possible to detect the inclination state of the crack K using the crack detection optical system 400 without adding a new detector, so it is possible to realize the crack detection device and method according to this embodiment at low cost.

According to this embodiment, the inclination direction and inclination angle θ of the crack K can be detected non-destructively and non-contact using the crack detection optical system 400, making it possible to efficiently evaluate the processing quality.

When the angles of incidence of the detection lights L1(B) and L1(C) are set to the same angle, it is possible to calculate the inclination angle θ of the crack K using formula (3). Even when the angles of incidence of the detection lights L1(B) and L1(C) are set to different angles, it is possible to calculate the inclination angle θ of the crack K using formula (2).

Although the present invention has been described above as an example, it goes without saying that the present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the spirit of the present invention.

10...crack detection device, 100...light source unit, 102A, 102B, 102C...light source, 104...half mirror, 200...illumination optical system, 202...relay lens, 204...mirror, 206...relay lens, 300...interface detection optical system, 302...half mirror, 304...half mirror, 306...relay lens, 308...half mirror, 310...photodetector, 400...crack detection optical system, 402...relay lens, 404, 406...photodetector, 500...control unit, 502...focus point position moving mechanism, 504... Objective lens, 506...operation unit, 508...display unit, 510...stage, 520...interface detection unit, 530...first crack detection unit, 540...second crack detection unit, 550...crack inclination state detection unit, 552...inclination direction detection unit, 554...inclination angle calculation unit, K...crack, Kt...upper end of crack, Kb...lower end of crack, KL(B), KL(C)...detected crack length, Pt(B), Pt(C)...detected position of upper end, Pb(B), Pb(C)...detected position of lower end, Wb...upper surface, Wa...lower surface, α, β, θ...angles

Claims (7)

an oblique illumination means for irradiating a workpiece having a crack formed therein with a first detection light and a second detection light from different oblique directions;
a first crack detection means for detecting a first crack length, which is a length of the crack in a thickness direction of the workpiece, based on a first reflected light of the first detection light from the workpiece;
a second crack detection means for detecting a second crack length, which is a length of the crack in a thickness direction of the workpiece, based on a second reflected light of the second detection light from the workpiece;
a crack inclination state detection means for detecting an inclination state of the crack with respect to a thickness direction of the workpiece based on the first crack length and the second crack length;
A crack detection device comprising: The crack inclination state detection means compares the first crack length with the second crack length, and determines the inclination direction of the crack according to a result of the comparison.
The crack detection device of claim 1 . The crack inclination state detection means determines that the crack is inclined when a difference between the first crack length and the second crack length is greater than a first threshold value.
3. The crack detection device of claim 2. The crack inclination state detection means determines that the crack is not inclined when a difference between the first crack length and the second crack length is smaller than a second threshold value.
A crack detection device according to claim 2 or 3. When the incident angle of the first detection light is α degrees, the incident angle of the second detection light is β degrees, the first crack length is KL(B), and the second crack length is KL(C),
The crack detection device according to claim 1 , wherein the crack inclination state detection means calculates the inclination angle θ of the crack by the following formula (1):
When the incidence angle of the first detection light and the incidence angle of the second detection light are the same angle α degrees,
The crack inclination state detection means calculates the inclination angle θ of the crack by the following formula (2):
A crack detection device according to any one of claims 1 to 4.
an oblique illumination step of irradiating a workpiece having a crack formed therein with a first detection light and a second detection light from different oblique directions;
a first crack detection step of detecting a first crack length, which is a length of the crack in a thickness direction of the workpiece, based on a first reflected light of the first detection light from the workpiece;
a second crack detection step of detecting a second crack length, which is a length of the crack in a thickness direction of the workpiece, based on a second reflected light of the second detection light from the workpiece;
a crack inclination state detection step of detecting an inclination state of the crack with respect to a thickness direction of the workpiece based on the first crack length and the second crack length;
A crack detection method comprising:

Images