KR102506283B1 - Die bonding apparatus and manufacturing method of semiconductor apparatus - Google Patents

Abstract

An object of the present invention is to provide a technique capable of improving the uniformity of imaging conditions of a plurality of imaging objects.
The die bonding device includes a conveyance path for conveying a substrate, a plurality of imaging devices fixedly arranged in a line along the width direction of the substrate above the conveyance path, and a plurality of attachment regions positioned on the substrate in a line along the width direction. is imaged with a plurality of imaging devices to acquire a plurality of images, generate a composite image based on the acquired plurality of images, and recognize an object to be captured in the attachment area based on the composite image. The imaging fields of each imaging device overlap on the substrate, and the overlapping imaging fields are larger than the attachment area.

Description

Die bonding device and manufacturing method of semiconductor device

The present disclosure relates to a die bonding device, and is applicable to, for example, a die bonder that images an attachment area with a plurality of recognition cameras.

Part of the semiconductor device manufacturing process includes a step of assembling a package by mounting a semiconductor chip (hereinafter simply referred to as a die) on a wiring board or a lead frame (hereinafter simply referred to as a board). Partly includes a process of dividing dies from a semiconductor wafer (hereinafter simply referred to as a wafer) (dicing process) and a bonding process of mounting the divided dies on a substrate. A semiconductor manufacturing device used in the bonding process is a die bonding device such as a die bonder.

A die bonder is a device for bonding (loading and adhering) a die onto a substrate or an already bonded die using resin paste, solder, gold plating, or the like as a bonding material. For example, in a die bonder that bonds a die to the surface of a substrate, a die is adsorbed and picked up from a wafer using an adsorption nozzle called a collet provided at the tip of a bonding head, and the die is placed on a predetermined position on the substrate and pressed. While applying, the operation (work) of performing bonding by heating the bonding material is repeatedly performed.

When using resin as a bonding material, resin pastes such as Ag epoxy and acrylic are used as adhesives (hereinafter referred to as paste adhesives). A pasty adhesive for adhering the die to the substrate is enclosed in a syringe, and the syringe moves up and down with respect to the substrate to inject and apply the pasty adhesive. That is, a paste-like adhesive is applied in a predetermined amount to a predetermined position by a syringe in which the paste-like adhesive is sealed, and a die is pressed and baked on the paste-like adhesive to adhere. A recognition camera is installed near the syringe, and positioning is performed by confirming the position where the pasty adhesive is applied with this recognition camera, and whether the applied pasty adhesive is applied in a predetermined amount in a predetermined shape at a predetermined position. is confirmed

Further, a recognition camera is installed near the bonding head, and positioning is performed by confirming the position of the substrate to which the die is bonded with the recognition camera, and furthermore, it is confirmed whether the bonded die is bonded to a predetermined position.

Japanese Unexamined Patent Publication No. 2013-197277

When a plurality of imaging objects are imaged with one imaging device using a non-telecentric lens, the imaging objects spaced apart from directly below the imaging device are imaged obliquely, so that the three-dimensional side view is seen.

An object of the present disclosure is to provide a technique capable of improving the uniformity of imaging conditions of a plurality of imaging objects.

Other problems and novel features will become clear from the description of this specification and the accompanying drawings.

A brief outline of representative ones of the present disclosure is as follows.

That is, the die bonding device includes a conveyance path for conveying a substrate, a plurality of imaging devices fixedly arranged in a row along the width direction of the substrate above the conveyance path, and a plurality of imaging devices positioned on the substrate along the width direction. Acquiring a plurality of images by capturing a plurality of attachment areas in a row with the plurality of imaging devices, generating a composite image based on the acquired plurality of images, and recognizing an object to be captured in the attachment region based on the composite image. and a control unit configured to do so. The imaging fields of each imaging device overlap on the substrate, and the overlapping imaging fields are larger than the attachment area.

According to the die bonding apparatus, it becomes possible to improve the uniformity of imaging conditions of a plurality of imaging objects.

FIG. 1(a) is a perspective view showing a normal field optical system, and FIG. 1(b) is a perspective view showing a wide field optical system.
FIG. 2 (a) is a conceptual diagram of a wide-field optical system using a macro lens, and FIG. 2 (b) is a conceptual diagram of a wide-field optical system using a telecentric lens.
Fig. 3(a) is a diagram showing a three-dimensional image of the paste, and Fig. 3(b) is a diagram showing bright lines on the paste when the substrate is viewed from a wide area with a macro lens.
Fig. 4 (a) is a top view explaining the diversification of the imaging device of the embodiment, and Fig. 4 (b) is a side view when viewed from the direction of arrow A in Fig. 4 (a).
5 is a diagram explaining a lighting device used in the imaging device of the embodiment.
6 is a top view illustrating overlapping of imaging fields of view of a plurality of imaging devices.
Fig. 7 is a side view when viewed from the direction of arrow A in Fig. 6;
8 is a diagram explaining the overlapping amount of the imaging field.
9 is a diagram illustrating image mosaicing and coordinate mapping.
10 is a schematic diagram illustrating image mosaicing and coordinate conversion using a calibration plate.
11 is a diagram explaining distortion.
12 is a diagram showing expressions of transformation matrices of affine transformation and projective transformation.
13 is a diagram explaining image mosaicing and coordinate transformation by a target model of a substrate.
Fig. 14 is a diagram explaining the influence of the displacement of the imaging device with the passage of time.
15 is a diagram illustrating spatial recalibration.
16 is a diagram for explaining a method of changing the height of a calibration plate.
Fig. 17 is a diagram explaining a marker provided in a chute.
Fig. 18 is a perspective view illustrating diversification of imaging devices in a modified example.
Fig. 19 is a top view schematically showing a die bonder in the embodiment.
FIG. 20 is a diagram explaining the operation of the pickup head and bonding head when viewed from the direction of arrow A in FIG. 19 .
Fig. 21 is a schematic sectional view showing a main part of the die supply section in Fig. 19;
Fig. 22 is a block diagram showing a schematic configuration of a control system of the die bonder of Fig. 19;
23 is a flowchart showing a method of manufacturing a semiconductor device.

Hereinafter, embodiments, modifications, and examples will be described using drawings. However, in the following description, the same reference numerals are assigned to the same components, and repeated explanations are omitted in some cases. In addition, in order to make explanation more clear, although the width, thickness, shape, etc. of each part may be schematically expressed compared with an actual aspect, it is only an example and does not limit the interpretation of this invention.

First, the technology examined by the present initiators will be described using FIGS. 1 to 3 . FIG. 1(a) is a perspective view showing a normal field optical system, and FIG. 1(b) is a perspective view showing a wide field optical system.

In the operation of sequentially performing positioning of the substrate S, attachment (die bonding or paste application), and inspection, as shown in FIG. 1 (a), a normal field of view optical system using a low-resolution imaging device (CML) When is used, since only an area slightly larger than one attachment area AA can be imaged, imaging and image recognition as many as the number of attachment areas AA are required. Here, as an example, the substrate S has six attachment regions AA in one row, and an example in which five rows of attachment regions are provided is shown. A low-resolution imaging device is an imaging device that can image only an area slightly larger than one attachment area with sufficient resolution, and has a resolution of about 300,000 to about 1,300,000 pixels, for example.

Therefore, as shown in Fig. 1 (a), the imaging device CML is moved in the width direction (Y-axis direction) of the substrate S, and imaging and image recognition of other attachment regions in the row are performed. It is necessary to repeat the operation of performing processing based on the image recognition. After imaging of one row, the substrate S is moved in the substrate transport direction (X-axis direction), and imaging of the next row is performed.

Therefore, a moving mechanism for moving the supporting member (not shown) of the imaging device CML in the width direction of the substrate S is required, and the supporting mechanism of the imaging device CML is complicated, enlarged, and expensive. . In addition, because of the movement time of the imaging device CML in the width direction, and because the waiting time for imaging is required until the shaking of the imaging device CML due to the vibration generated along with the movement of the supporting member converges, , the die bonder cannot be speeded up. Alternatively, a vibration prevention mechanism is required to prevent vibration, and the die bonder becomes expensive. In addition, when there is a mechanical part that reciprocates in the sky above the attachment area and below the imaging device, it is necessary to determine the timing at which the imaging field of view is not blocked by the mechanical part to perform imaging. The speed of the die bonder cannot be increased by securing this imaging timing.

On the other hand, instead of repeating it for each attachment after conveying the substrate, as shown in FIG. By imaging and recognizing in the entire attachment area (all tabs) of the substrate S or in at least one row of attachment areas in the width direction of the substrate, the efficiency of the processing time can be improved, and the above-mentioned problems can be solved. Here, the high-resolution imaging device is an imaging device that can collectively image the entire area of at least one column in the width direction of the substrate S with sufficient resolution, and has, for example, a resolution of about 5 million pixels or more. Column 1 includes a plurality of attachment areas, for example, 4 attachment areas.

However, the wide-field optical system (wide area camera) has the following problems. This problem is explained using FIGS. 2 and 3 . FIG. 2 (a) is a conceptual diagram of a wide-field optical system using a macro lens, and FIG. 2 (b) is a conceptual diagram of a wide-field optical system using a telecentric lens. Fig. 3(a) is a diagram showing a three-dimensional image of the paste, and Fig. 3(b) is a diagram showing bright lines on the paste when the substrate is viewed from a wide area with a macro lens.

(1) Securing telecentricity

A wide area is obtained simply by lowering the magnification of the macro lens, but the more it moves away from the center of the field of view, the more obliquely upward the view is. , the side is not visible, but the right side is visible in the left image pickup object OBL, and the left side is visible in the right image pickup object OBR, so that the side surface of the three-dimensional shape is seen. Moreover, magnification changes according to the height of the imaging target OBL, and when the height changes, alignment becomes difficult, for example. For example, a paste applied to a substrate or a die stacked on a substrate has a different height depending on a location.

It is possible to solve these problems by using a telecentric lens. Since the telecentric lens condenses parallel light, the sides are invisible for all imaging objects. However, in that case, as shown in (b) of FIG. 2, generally a lens with an aperture larger than the desired field of view is required. An aperture of 141 mm or more is required, and the focal length also becomes long along with it. Mounting such a large lens on a die bonding apparatus is not desirable from the viewpoint of efficiency.

(2) Irradiation of uniform illumination within the field of view

When an attempt is made to secure a wide area by using a non-telecentric lens, the end of the field of view is viewed obliquely from above. Since the directions are reversed at the right end and left end of the visual field, or at the top and bottom ends, the irradiation direction of the illumination tends to be non-uniform. Further, even when collimated light using telecentric light is irradiated, it becomes difficult to obtain a uniformly irradiated image within the imaging field. For example, as shown in Fig. 3(a), the paste adhesive PA applied to the substrate in a cross shape has a high center portion and a low tip portion. When the substrate coated with the pasty adhesive as shown in Fig. 3 (a) is viewed from a wide area with a macro lens, as shown in Fig. 3 (b), the center of the imaging visual field is directly above the object to be imaged. Although it is bright when viewed, the peripheral portion is dark because it is viewed obliquely, and since the height of the pasty adhesive is not uniform, the bright line on the pasty adhesive PA shifts. Here, the center of the field of view of the imaging device is located at the center of the image in FIG. 3(b). The substrate S has tabs TB, which are six attachment regions, in one column, and seven columns are shown. A paste-like adhesive PA is applied to the tabs TB in four rows on the right side of the substrate S.

Next, an embodiment that solves the above problems will be described using FIGS. 4 to 5 . Fig. 4 (a) is a top view explaining the diversification of the imaging device of the embodiment, and Fig. 4 (b) is a side view when viewed from the direction of arrow A in Fig. 4 (a). Fig. 5 is a diagram explaining a lighting device used in the imaging device of the embodiment, Fig. 5 (a) is a cross-sectional view of the lighting device, and Fig. 5 (b) is a side view.

As an embodiment to solve the above-mentioned problem, multiple imaging devices (stereoization) are performed. For example, as shown in FIG. 4 , a plurality of imaging devices CM1 to CM4 are arranged in a row above the substrate S in the width direction (Y-axis direction) of the substrate S and fixedly arranged. . Here, each of the imaging devices CM1 to CM4 is an imaging device having a resolution equal to or higher than that of the imaging device CML, and the entire area of one column in the width direction of the substrate S cannot be collectively imaged with sufficient resolution. It is an imaging device. In addition, each of the imaging devices CM1 to CM4 may use a telecentric lens, but it is preferable to use a non-telecentric lens such as a macro lens. These imaging devices CM1 to CM4 are spaced at predetermined intervals in the horizontal direction at the same height, and the optical axes of each imaging device CM1 to CM4 are parallel to each other and perpendicular to the substrate S. Further, the optical axes of the imaging devices CM1 to CM4 may be slightly inclined from the perpendicular to the substrate S within a range in which defocusing in the imaging visual field is permitted.

In this embodiment, each of the imaging devices CM1 to CM4 captures an image of an object in the attachment areas AA11 to AA14. In addition, the respective imaging visual fields IA1 to IA4 of the imaging devices CM1 to CM4 do not overlap. Furthermore, by moving the board|substrate S in the conveyance direction (Y-axis direction), imaging objects of the remaining rows are sequentially imaged. By diversifying the imaging devices, an image can be captured almost directly above the object to be imaged, and the uniformity of illumination within the imaging field can be improved and inspection can be performed. Further, by diversifying the imaging devices, there is no need to move the imaging devices, and the same processing efficiency as that of the wide-field optical system can be obtained.

Here, the substrate S has a rectangular shape and a flat plate shape, for example, as shown in (a) (b) of FIG. has The length of the board|substrate S in the conveyance direction (X-axis direction) is comprised longer than the length in the width direction (Y-axis direction).

It is preferable that the multiplexed imaging devices each have the same illumination system. For example, as shown in (a) of FIG. 5 , between the imaging devices CM1 to CM4 and the substrate S, a surface light source (light source) SL and a half mirror (semi-transmissive mirror) HM ) is disposed inside the lighting device LD. Illuminated light from the surface emitting illumination SL is reflected on the same optical axis as the imaging devices CM1 to CM4 by the half mirror HM, and irradiated onto the object to be imaged in the attachment areas AA11 to AA14 of the substrate S do. The scattered light irradiated to the imaging target in the attachment areas AA11 to AA14 on the same optical axis as the imaging devices CM1 to CM4 is reflected from the imaging target in the attachment areas AA11 to AA14, and the specularly reflected light among them is reflected by the half mirror ( HM) to reach the imaging devices CM1 to CM4, and form images of objects to be captured in the attachment areas AA11 to AA14. That is, the lighting device LD has a function of coaxial fall-out illumination (coaxial illumination).

As shown in (b) of FIG. 5, the Y-axis length of the surface emitting illumination SL and the half mirror HM is equal to the length of the imaging devices CM1 to CM4 in the width direction of the substrate S. It is configured slightly wider than the entire field of view, and the surface light emitting illumination SL is divided into light emitting areas SL1 to SL4 that are slightly wider than the imaging fields of each of the imaging devices CM1 to CM4, and can be individually turned on/off. . Since the light emitting areas of the coaxial lighting devices are divided, it is possible to adjust light for each of the imaging devices CM1 to CM4. This makes it possible to improve the uniformity of illumination in all areas of the imaging field of view of the imaging devices CM1 to CM4. In addition, as will be described later, when each of the imaging fields of the imaging devices CM1 to CM4 overlaps, the light emitting areas SL1 to SL4 also overlap.

Next, overlapping of imaging fields of view of a plurality of imaging devices will be described using FIGS. 6 to 8 . 6 is a top view illustrating overlapping of imaging fields of view of a plurality of imaging devices. Fig. 7 is a side view when viewed from the direction of arrow A in Fig. 6; 8 is a diagram explaining the overlapping amount of the imaging field.

In the embodiment described above, an example in which the imaging visual fields IA1 to IA4 of the imaging devices CM1 to CM4 do not overlap has been shown. However, since the pitches of various products (pitch of the attachment area) and the pitch of the imaging devices are not necessarily the same when diversifying the imaging devices, as shown in Figs. It is desirable to overlap.

6 and 7 show an example in which four imaging devices are provided and six attachment regions are provided in one column of the substrate S as an example, and the pitch of the imaging devices is larger than the pitch of the attachment regions. For this reason, imaging visual fields of adjacent imaging devices are overlapped, so that, for example, two or more imaging targets are included in the imaging field of one imaging device. Accordingly, each of the imaging devices CM1 to CM4 is an imaging device having a higher resolution than the imaging device CML. Each of the imaging devices CM1 to CM4 is an imaging device that cannot capture the entire area of the substrate S for one row in the width direction collectively with sufficient resolution, but the entire area for one column of the width direction of the substrate S is collectively It may be an imaging device capable of capturing an image with sufficient resolution. Here, the board|substrate S has a rectangular shape and flat plate shape, for example, as shown in FIG. 6, and has many attachment areas AA11 to AA16, AA21 to AA26, ... vertically and horizontally. The length of the board|substrate S in the conveyance direction (X-axis direction) is comprised longer than the length in the width direction (Y-axis direction).

The overlapping area OVL of the imaging visual field should just be the size in which the imaging target of the largest size is included (it is sufficient to maintain the maximum product size). As a result, all of the imaging objects are included in a certain imaging field. As shown in Fig. 8, the overlapping area OVL only needs to have a size where the imaging object OB2 can fit. For example, in the case of a substrate substrate, if there is an overlapping amount of the maximum tab size, all tabs are in a certain field of view, the foreground of the tab goes in

Synthesis of a plurality of captured images will be described using FIGS. 9 to 15 . 9 is a diagram illustrating image mosaicing and coordinate mapping. 10 is a schematic diagram illustrating image mosaicing and coordinate conversion using a calibration plate. 11 is a diagram explaining distortion. 12 is a diagram showing expressions of transformation matrices of affine transformation and projective transformation. 13 is a diagram explaining image mosaicing and coordinate transformation by a target model of a substrate. Fig. 14 is an image captured with a plurality of imaging devices of a state in which a paste-like adhesive is applied on tabs of a multiple lead frame, Fig. 14 (a) is a captured image when there is no shift with time, and Fig. 14 (b) ) is a captured image in the case where there is a temporal shift. Fig. 15 is a diagram explaining spatial re-correction, Fig. 15 (a) is a diagram showing a state before spatial re-correction, and Fig. 15 (b) is a diagram showing a state after spatial re-correction.

In order to suppress the amount of overlap, the control unit CNT combines images captured by a plurality of imaging devices by image mosaicking or the like. Since general image mosaicing attempts to smoothly connect a plurality of images, calibration of images may be lost in that state.

So, in the embodiment, in the synthesis of images captured by a plurality of imaging devices,

(A) At the time of shipment or adjustment of the die bonding device, image mosaicing and coordinate conversion using a calibration plate are performed,

(B) During fine adjustment or continuous operation of the die bonding apparatus, image mosaicing by the target model of the substrate and coordinate transformation by the target model disposed on the chute are performed.

The above (A) will be described using FIGS. 9 to 12 .

As shown in Fig. 9, when adjusting the die-bonding device, the control unit CNT projects a coordinate marker CMRK having a scale covering all the imaging fields of the multiplexed imaging device, and the overlapping area OVL range Coordinate transformation by projective transformation or affine transformation is performed on the image based on the same intersection point (IP) of the coordinate marker (CMRK) in get Here, the coordinate marker CMRK is used by preparing a calibration plate as an adjustment jig and marking the plate, and the coordinate marker CMRK is, for example, a lattice shape. In the case of coordinate conversion, a marker that guarantees the positional relationship of the entire space is required. If there is a coordinate marker CMRK that covers the entire synthetic visual field as shown in Fig. 9, the positional relationship of the entire space can be guaranteed during coordinate transformation such as projection transformation, and image space coordinates and real space coordinates at each intersection pitch. Matching of coordinates is possible.

As shown in Fig. 10, the control unit CNT captures an image of one large calibration plate CP having a grid-like pattern by dividing the field of view with, for example, three imaging devices. The three imaging devices have overlapping regions OV12 and OV23 in which fields of view overlap to some extent by adjacent imaging devices. A lattice-shaped intersection IP in the overlapping regions OV12 and OV23 is shown as a black dot.

First, the controller CNT transforms the pixel coordinates of an image of one of the adjacent imaging devices with respect to one of the adjacent imaging devices by affine transformation or projection transformation. Here, as an example, the reference image is described as the image IV1 and the converted image as the image IV2. Transformation is a transformation matrix of affine transformation or projection transformation so that the coordinates (sunspot coordinates) of intersections (IP) in the mutually transformed image match the coordinates (sunspot coordinates) of intersections (IP) in the corresponding reference image. Calculate each parameter of Here, in general, the transformation matrix of the affine transformation is shown in Equation (1) in FIG. 12, and the determinant of the projective transformation is shown in Equation (2) in FIG. It is generally necessary to have three coordinates to calculate the transformation matrix, but it is preferable to perform transformation for each grid pattern because more accurate transformation is possible by performing the transformation not uniquely, but for each part corresponding to the lattice pattern on the lattice. .

In the case where the reference image IV1 has distortion represented by a cylindrical shape or a threaded shape as shown in FIG. 11, all intersections IP of the calibration plate CP that fall within the field of view are used first. It is preferable to perform distortion correction by converting the original image into an image conforming to the orthogonal coordinate system to be erected (first distortion correction). In this way, other than the overlapping areas OV12 and OV23, it is possible to introduce them into the composite image by simple scaling.

Since coordinate alignment is performed in the pixel coordinate system, the coordinates after transformation of the image IV2 are required to be integer values, but the pixels transformed by affine transformation or projection transformation do not necessarily converge to that location, and the coordinates after transformation may be an intermediate value. In this case, each coordinate of the transformed image is interpolated from the grayscale value of the adjacent transformed coordinate, represented by a nearest-nearest method, a bilinear method, or a bicubic method.

The above (B) will be described using FIGS. 13 to 15.

Conversion in the adjustment work before production operation is performed because a large number of intersections (IP) of the calibration plate (CP) that indicate the reference of coordinates in the image space within the field of view are arranged at an equal pitch, so that the image space and the real coordinate space are fit. this is relatively easy In contrast, during continuous operation or simple adjustment, image synthesis and coordinate alignment are performed without using the calibration plate CP. Since image synthesis only requires knowing the positions indicating the same places, a target mark TM for positioning on the substrate or the like is used. The target mark TM for positioning is used because the registration of the template model is performed as a process for determining the position of the tab at the time of attachment (bonding or paste). As shown in Fig. 13, when a plurality of taps enter the overlapping areas OV12 and OV23 of the image, three or more points are secured using target marks TM for positioning of adjacent taps. When only one tap enters the overlapping area, three positioning target marks TM are previously registered as a template image model in one tap.

In this method, synthesis by matching coordinates between images is possible, but alignment with real space is impossible. Even if any image is to be matched as a reference, if it is synthesized as it is without performing fitting with the coordinate space first, as shown in FIG. 11, an image IV2 adjacent to the image IV1 as a reference image and an image ( The image IV3 adjacent to the image IV2) is affected by the distortion of the image IV1. When the displacement due to distortion is successively adjacent to the image IV1, the image IV3 at the farthest position is most amplified. In order to suppress this amount of amplification, markers SMRK are placed at known coordinates on the suit SCT, and the coordinates of the markers SMRK are measured again from the image, thereby collectively returning the synthesized image. Since the distortion of this image depends on the lens of the imaging device, image synthesis is performed after the transformation (first distortion correction) at the time of the first transformation, and the markers SMRK on the suit SCT are placed to match the overall coordinates. It is preferable to use That is, after correcting the real space and the image space using the coordinate marker CMRK, the control unit CNT grasps the coordinates of the marker SMRK separately provided on the chute SCT as a transport path. Here, the chute SCT is located on the outer side of both ends of the width direction of the board|substrate S.

As described above, the coordinate markers CMRK are aligned as an adjustment operation before the production run by the die bonding apparatus. However, as the groundbreaking progresses, the imaging field of each imaging device slightly changes as shown in FIG. It may have displacement with time. Here, the image within the imaging visual field IA2 of the imaging device CM2 is shifted to the right with respect to the image within the imaging visual field IA1 of the imaging device CM1. That is, the imaging device CM2 is displaced to the right. In addition, the location surrounded by the rectangular frame indicates the vicinity of the lower left corner of each tab. This shift cannot be ignored in a die bonding apparatus that requires micrometer-level precision. Even in the middle of construction, it is necessary to compensate for these subtle discrepancies. A method of correcting the slight shift will be described below with reference to FIG. 15 .

As described above, when a shift between imaging devices occurs during continuous construction, a difference may occur in the pitch between tabs of the substrate S having a known pitch. The controller CNT detects the inherent displacement due to factors such as heat of the imaging device by periodically measuring this known pitch between taps. In addition, the control unit CNT also detects the inherent displacement due to factors such as heat of the imaging device by periodically measuring the distance between the markers SMRK on the known chute SCT.

When this displacement is detected, the control unit CNT recalculates a projection transformation matrix or an affine transformation matrix for synthesizing and transforming images using feature markers such as target marks TM for positioning on the substrate S. At this time, the obtained projective transformation matrix or affine transformation matrix can be connected to images, but as shown in FIG. 15(a), the image space and the real space may not match. Therefore, the control unit CNT uses the marker SMRK on the chute SCT measured for the first time and performs reconversion based on the coordinates. In this way, a composite image as shown in Fig. 15(b) is obtained.

Moreover, the magnification changes depending on the thickness of the substrate P (the height of the upper surface of the substrate P), and alignment becomes difficult when the height changes. A method for reducing the influence of the thickness of the substrate will be described using FIGS. 16 and 17 .

16 is a diagram explaining a method of changing the height of the calibration plate, FIG. 16 (a) is a cross-sectional view showing a state in which a substrate is transported and loaded onto an attachment stage, and FIG. 16 (b) is a view of the calibration plate It is a cross-sectional view showing vertical movement. Fig. 17 is a view explaining markers provided in the chute, Fig. 17 (a) is a cross-sectional view showing a state in which substrates are transported and loaded onto an attachment stage, and Fig. 17 (b) is a chute provided with markers. It is a top view, and FIG. 17(b) is a cross-sectional view of a chute provided with a marker in another example.

In order to respond to the height displacement, the control unit CNT finely moves the calibration plate CP up and down to obtain a projection transformation matrix for each height. The control unit CNT calculates the expected height of the alignment pattern position or inspection field position from the known substrate thickness, paste height, die thickness, etc., and automatically selects which projection transformation matrix maintained for each height to be used. The controller CNT recognizes the same point on the substrate in the overlapping area between adjacent imaging devices, and measures the height. The control unit CNT automatically selects a projection transformation matrix to be used from the measurement value.

Specifically, at the time of correction by the first calibration plate CP, as shown in Fig. 16(b), the attachment stage AS having the calibration plate CP is moved up and down to The coordinates of the intersections of (CP) are measured for each height of the attachment stage AS. The vertical movement mechanism of the attachment stage is configured to be capable of vertical movement in μm units.

Moreover, it is preferable to set the marker SMRK installed on the chute SCT at the same height as the height of the upper surface of the board|substrate P. As shown in Fig. 17 (a) (b), holes (holes) having different depths and diameters are formed in the chute SCT, for example. The hole representing the marker SMRK may be individually provided for each depth as shown in FIG. 17(c). The marker SMRK may not be a hole as long as it is flush with the upper surface of the substrate P.

According to the embodiment, it has one or more of the following effects.

(a) Since each imaging target can be viewed from almost directly above, it is possible to prevent an inclination in the height direction of an image in a deep field of view that occurs in a simple low-magnification optical system.

(b) Since the inherent displacement due to factors such as heat of the imaging device can be detected, and correction between the imaging devices can be performed during the driving operation of the die bonder, it is possible to reduce the influence of the displacement of the imaging device over time. .

(c) Since it can cope with height displacement and is not affected by the change in thickness depending on the type of substrate, it is possible to reduce the influence of the difference in type.

(d) Stabilization of positioning accuracy and stabilization of inspection are possible by at least any one of the above (a) to (c).

<Example of modification>

Hereinafter, representative modifications of the embodiments will be illustrated. In the description of the modified example below, it is assumed that the same reference numerals as those in the above-described embodiment can be used for parts having the same configuration and function as those described in the above-described embodiment. And regarding the explanation of these parts, it is assumed that the explanation in the above-mentioned embodiment can be appropriately used within the range which is not technically contradictory. In addition, all or some of the above-described embodiments and modified examples can be appropriately applied in combination within a range that is not technically contradictory.

Fig. 18 is a perspective view illustrating diversification of imaging devices in a modified example. In the embodiment, an example has been described in which imaging objects in a row in the width direction of the substrate S are imaged with a plurality of imaging devices, but a plurality of imaging devices are also disposed in the longitudinal direction of the substrate S, that is, a plurality of imaging devices. The devices may be arranged in a lattice form to capture images of a plurality of columns of objects to be imaged.

For example, as shown in FIG. 18 , four rows of imaging device groups CM10 to CM40 are arranged, and each imaging device group has four imaging devices CM1 to CM4 of the embodiment arranged in a row, respectively. , 16 imaging devices are arranged in a grid pattern. Here, on the substrate S, 6 imaging target units per column are arranged in 5 rows, and the imaging fields of adjacent imaging devices overlap.

As shown in FIG. 18, not only all the imaging objects in the first row of the substrate S, but also all the imaging objects in the second and subsequent rows are captured in parallel, and the synthesized image can be recognized. It is possible to make the frequency|count of imaging by moving the board|substrate S smaller than embodiment.

[Example]

Examples to which the above-described embodiments are applied will be described below. Fig. 19 is a top view schematically showing a die bonder in the embodiment. FIG. 20 is a diagram explaining the operation of the pickup head and bonding head when viewed from the direction of arrow A in FIG. 19 .

The die bonder 10 is largely divided into a die supply unit 1 for supplying a die D to be mounted on a substrate S, a pickup unit 2, an intermediate stage unit 3, and a preform unit ( 9), a bonding unit 4, a transport unit 5, a substrate supply unit 6, a substrate carrying unit 7, and a control unit 8 that monitors and controls the operation of each unit. The Y-axis direction is the front-back direction of the die bonder 10, and the X-axis direction is the left-right direction. The die supply unit 1 is disposed on the front side of the die bonder 10, and the bonding unit 4 is disposed on the inner side. Here, a plurality of product areas (hereinafter referred to as package areas P), which constitute one final package, are formed on the substrate S. For example, when the substrate S is a lead frame, the package area P has a tab on which the die D is loaded.

First, the die supply unit 1 supplies the die D to be mounted in the package area P of the substrate S. The die supply unit 1 has a wafer holder 12 for holding the wafer 11 and a peeling unit 13 indicated by dotted lines for pushing the die D from the wafer 11 . The die supply unit 1 is moved in the XY-axis direction by a drive unit (not shown) to move the die D to be picked up to the position of the peeling unit 13 .

The pick-up unit 2 lifts a pick-up head 21 that picks up the die D, a Y drive part 23 of the pick-up head that moves the pick-up head 21 in the Y-axis direction, and a collet 22, It has each drive part (not shown) which rotates and moves in the X-axis direction, and the wafer recognition camera 24 which grasps the pickup position of the die D to be picked up from the wafer 11. The pick-up head 21 has a collet 22 (see also FIG. 14 ) that adsorbs and holds the pushed-up die D at its tip, picks up the die D from the die supply unit 1, and performs an intermediate stage Load in (31). The pick-up head 21 has respective drive units (not shown) that lift, rotate, and move the collet 22 in the X-axis direction.

The intermediate stage unit 3 has an intermediate stage 31 for temporarily loading the die D, and a stage recognition camera 32 for recognizing the die D on the intermediate stage 31 .

The preform unit 9 includes a syringe 91, a driving unit 93 that moves the syringe 91 in the Y and Z directions, and an adhesive recognition camera 94 as an imaging device for grasping the application position of the syringe 91 and the like. have Here, the adhesive recognition camera 94 is, for example, the diversified imaging devices CM1 to CM4 of the embodiment, and the imaging devices CM1 to CM4 are configured to capture images using the lighting device LD, respectively. . In the preform section 9, a paste-like adhesive such as epoxy resin is applied with a syringe 91 to the substrate S conveyed by the conveyance section 5. The syringe 91 is configured so that a pasty adhesive is sealed therein, and the pasty adhesive is extruded and applied to the substrate S from the tip of the nozzle by air pressure. In the case where the substrate S is a multi-lead frame in which a plurality of unit lead frames are arranged horizontally and continuously installed in series, for example, a paste adhesive is applied to each tab of the unit lead frame.

The bonding unit 4 picks up the die D from the intermediate stage 31 and bonds it onto the package area P coated with the pasty adhesive of the transported substrate S. The bonding unit 4 includes a bonding head 41 provided with a collet 42 (see also FIG. 20 ) for adsorbing and holding the die D at the tip, similarly to the pick-up head 21, and the bonding head 41 It has a Y drive unit 43 that moves in the Y-axis direction, and a substrate recognition camera 44 that captures an image of a position recognition mark (not shown) of the package area P of the substrate S and recognizes the bonding position. . Here, the substrate recognition camera 44 is, for example, the diversified imaging devices CM1 to CM4 of the embodiment, and the imaging devices CM1 to CM4 are configured to capture images using the lighting device LD, respectively. . With this configuration, the bonding head 41 corrects the pick-up position and posture based on the image data of the stage recognition camera 32, picks up the die D from the intermediate stage 31, and Based on the imaging data of 44), the die D is bonded to the substrate.

The conveyance unit 5 has substrate conveyance claws 51 that grip and convey the substrate S, and conveyance lanes 52 through which the substrate S moves. The substrate S is moved by driving the not-shown nut of the substrate transport claw 51 provided in the transport lane 52 with a ball screw (not shown) provided along the transport lane 52 . With this configuration, the substrate S moves from the substrate supply unit 6 to the bonding position along the transfer lane 52, and after bonding, moves to the substrate carrying out unit 7 and carries the substrate to the substrate carrying out unit 7. convey (S).

Next, the configuration of the die supply section 1 will be described using FIG. 21 . Fig. 21 is a schematic sectional view showing a main part of the die supply section in Fig. 19;

The die supply unit 1 includes a wafer holding table 12 that moves in the horizontal direction (XY-axis direction) and a separation unit 13 that moves in the vertical direction. The wafer holding table 12 includes an expander ring 15 holding the wafer ring 14 and a support ring 17 horizontally positioning the dicing tape 16 fixed to the wafer ring 14. have A die D diced into a mesh shape in the wafer 11 is adhesively fixed to the dicing tape 16 . The peeling unit 13 is disposed inside the support ring 17 .

The die supply unit 1 lowers the expand ring 15 holding the wafer ring 14 when the die D is pushed up. As a result, the dicing tape 16 held by the wafer ring 14 is extended, the gap between the dies D is increased, and the dicing tape 16 is removed from below the die D by the peeling unit 13. ) is pushed up or horizontally moved to improve the pick-up of the die D.

As shown in FIG. 22, the control system 80 includes a control unit 8, a drive unit 86, a signal unit 87, and an optical system 88. The control unit 8 is largely divided into a control/arithmetic unit 81 mainly composed of a CPU (Central Processor Unit), a storage unit 82, an input/output unit 83, a bus line 84, It has a power supply unit (85). The storage device 82 includes a main storage device 82a composed of RAM storing processing programs and the like, and an auxiliary storage device 82b composed of a HDD storing control data and image data necessary for control. have The input/output device 83 includes a monitor 83a for displaying device status and information, a touch panel 83b for inputting operator instructions, a mouse 83c for operating the monitor, and an optical system 88. It has an image introduction device 83d that introduces image data. In addition, the input/output device 83 includes a motor control device 83e that controls a drive unit 86 such as an XY table (not shown) of the die supply unit 1 and a ZY drive shaft of a bonding head table, and various sensor signals and an I/O signal control device 83f for introducing or controlling signals from a signal section 87 such as a switch or the like of a lighting device or the like. The optical system 88 includes a wafer recognition camera 24, an adhesive recognition camera 94, a stage recognition camera 32, and a substrate recognition camera 44 shown in FIG. 20 . The control/arithmetic unit 81 takes in necessary data via the bus line 84, calculates it, and sends information to the control of the bonding head 41 or the like or to the monitor 83a or the like.

The control unit 8 stores image data captured by the optical system 88 via the image introduction device 83d in the storage device 82. The positioning of the die D and the substrate S is determined using the control/arithmetic unit 81 by the software programmed based on the stored image data, the inspection of the application pattern of the pasty adhesive, and the die D and The surface inspection of the board|substrate S is performed. Based on the positions of the die D and the substrate S calculated by the control/arithmetic unit 81, the driver 86 is operated via the motor control unit 83e by software. By this process, the positioning of the die D on the wafer 11 is performed, and the die D is bonded onto the substrate S by operating the drive unit of the die supply unit 1 and the die bonding unit 4. A recognition camera used in the optical system 88 is a gray scale or color camera, and digitizes the light intensity.

Next, a semiconductor device manufacturing method using the die bonder according to the embodiment will be described with reference to FIG. 23 . FIG. 23 is a flowchart showing a method of manufacturing a semiconductor device using the die bonder of FIG. 19 .

(Step S51: Wafer/substrate loading process)

The wafer ring 14 holding the dicing tape 16 to which the die D divided from the wafer 11 is attached is stored in a wafer cassette (not shown) and carried into the die bonder 10 . The controller 8 supplies the wafer ring 14 to the die supply unit 1 from the wafer cassette filled with the wafer ring 14 . Further, the substrate S is prepared and carried into the die bonder 10 . The control unit 8 installs the substrate S on the substrate transport claw 51 in the substrate supply unit 6 .

(Step S52: pick-up process)

The control unit 8 moves the wafer ring 14 so that the desired die D can be picked up from the wafer ring 14 by the wafer holder 12, and the data captured by the wafer recognition camera 24 Based on this, positioning and surface inspection are performed. The controller 8 peels the positioned die D from the dicing tape 16 by the peeling unit 13 .

The controller 8 picks up the peeled die D from the wafer 11 by the pick-up head 21 . In this way, the die D peeled from the dicing tape 16 is adsorbed and held by the collet 22 of the pick-up head 21, and conveyed to the next step (step BS13). Then, when the collet 22 that transported the die D to the next step returns to the die supply unit 1, the next die D is peeled from the dicing tape 16 according to the above procedure, and thereafter the same According to the procedure of, the dies D are peeled from the dicing tape 16 one by one.

(Step S53: Bonding process)

The control unit 8 acquires an image of the surface of the substrate S before application by means of the adhesive recognition camera 94, and confirms the surface on which the paste adhesive is to be applied. If there is no problem with the surface to be applied, the controller 8 applies the paste adhesive from the syringe 91 to the substrate S transported by the transport unit 5. When the substrate S is a multi-lead frame, a paste-like adhesive is applied to all tabs. After the application, the control unit 8 checks again with the adhesive recognition camera 94 whether the pasty adhesive is applied correctly, and inspects the applied pasty adhesive. If there is no problem in the coating, the control unit 8 conveys the substrate S to the bonding stage BS by the conveying unit, and performs positioning based on image data captured by the substrate recognizing camera 44.

The controller 8 loads the die D picked up from the wafer 11 by the pick-up head 21 on the intermediate stage 31, and transfers the die D from the intermediate stage 31 to the bonding head 41 again. is picked up and bonded to the positioned substrate (S). The controller 8 inspects whether or not the die D is bonded to a desired position based on the image data captured by the substrate recognition camera 44 .

(Step S54: Substrate unloading process)

The control unit 8 takes out the substrate S to which the die D is bonded from the substrate carrying claw 51 in the substrate carrying out unit 7 . The substrate S is carried out from the die bonder 10 .

In the above, the invention made by the present inventors has been specifically described based on the embodiments, modifications, and examples, but the present disclosure is not limited to the above embodiments, modifications, and examples, and various changes are possible. needless to say

For example, in the embodiment, an example in which a paste adhesive is applied to a substrate in the preform unit has been described, but the adhesive for adhering the die to the substrate is applied to the wafer 11 and the wafer 11 instead of the paste adhesive applied by the syringe 91. A film-like adhesive material called a die attach film (DAF) that is affixed between the dicing tapes 16 may be used. The DAF is suitable for a stacked package configured by stacking several dies on a die on a substrate (S).

Further, in the embodiment, an intermediate stage unit 3 is provided between the die supply unit 1 and the bonding unit 4, and the die D picked up from the die supply unit 1 by the pickup head 21 is placed on an intermediate stage ( 31), the bonding head 41 picks up the die D again from the intermediate stage 31, and bonds it to the transported substrate S. The die D picked up from (1) may be bonded to the substrate S.

10: die bonder (die bonding device)
AA: attachment area
CM1 to CM4: imaging device
CNT: control part
S: substrate
SCT: chute (return route)

Claims (23)

A conveyance path for conveying the substrate;
a plurality of imaging devices fixedly arranged in a line above the conveyance path along the width direction of the substrate;
A plurality of attachment regions in a row along the width direction located on the substrate are imaged with the plurality of imaging devices to acquire a plurality of images, a composite image is generated based on the plurality of acquired images, and the composite image A control unit configured to recognize an object to be captured in the attachment area based on
to provide,
imaging fields of each imaging device overlap on the substrate, and the overlapping imaging fields are larger than the attachment area;
The transport path has a plurality of fiducial markers on the outside of both ends of the substrate in the width direction, respectively;
The substrate has a plurality of tabs disposed at predetermined intervals,
The die bonding device, wherein the control unit is configured to measure an interval of the tab or an interval of the fiducial marker with the imaging device to detect a displacement between the imaging devices. The die bonding apparatus according to claim 1, wherein the control unit is configured to generate the composite image based on coordinate markers located in overlapping imaging fields of view. 3. The method of claim 2, wherein the coordinate markers are scales on a grid covering all fields of view of the imaging device,
Wherein the control unit illuminates the coordinate markers, projects images based on the same intersection of the coordinate markers entering the overlapping imaging fields of view, and connects images between respective imaging devices to generate the synthesized image. die bonding device. delete The method of claim 1, wherein the substrate further has feature markers,
wherein the control unit is configured to recalculate a projection transformation matrix for synthesizing and transforming an image based on the feature marker, when displacement between the imaging devices is detected. The die bonding apparatus according to claim 5, wherein the control unit is configured to recalculate the projective transformation matrix based on coordinates of the reference marker measured in advance. The method of claim 6, wherein the control unit finely moves the substrate up and down to obtain a projection transformation matrix for each height,
A die configured to calculate an expected height of an alignment pattern position or an inspection field position from a thickness of the substrate or a paste height or a die thickness, and select one of the projection transformation matrices maintained for each height based on the calculated expected height. bonding device. 8. The projection transformation according to claim 7 , wherein the control unit measures a height by recognizing the same point on the substrate in overlapping imaging fields of view between adjacent imaging devices, and maintains the projection transformation for each height based on the measured height. A die bonding apparatus configured to select a matrix. The method according to any one of claims 1 to 3 and 5 to 8, further comprising a plurality of lighting devices provided corresponding to each of the plurality of imaging devices,
wherein the controller is configured to independently dim a plurality of the lighting devices. The method according to any one of claims 1 to 3 and 5 to 8, wherein the control unit conveys the substrate in the longitudinal direction of the substrate to transfer a plurality of attachment regions in a next row to the plurality of imaging devices. A die bonding device configured to take an image. The die bonding device according to any one of claims 1 to 3 and 5 to 8, wherein the imaging target is a paste adhesive applied to the substrate. 12. The die bonding device according to claim 11, wherein the control unit is configured to inspect the appearance of the paste adhesive applied to the substrate by the imaging device. The die bonding apparatus according to any one of claims 1 to 3 and 5 to 8, wherein the imaging target is a die bonded onto the substrate or an already bonded die. A conveyance path for conveying substrates having a plurality of tabs disposed at predetermined intervals, a plurality of imaging devices fixedly arranged in a row along the width direction of the substrate above the conveyance path, and each of the plurality of imaging devices A plurality of lighting devices are provided corresponding to the image pickup devices, the imaging fields of each imaging device overlap on the substrate, the overlapping imaging fields are configured to be larger than the tab, and the transport path is configured to have a width direction of the substrate A step of carrying the substrate into a die bonding device having a plurality of fiducial markers on the outside of both ends, respectively;
measuring the distance between the tabs or the fiducial marker with the imaging device and detecting displacement between the imaging devices;
A plurality of tabs in a row along the width direction located on the substrate are imaged with the plurality of imaging devices to acquire a plurality of images, a composite image is generated based on the plurality of acquired images, and the composite image a process of recognizing an imaging target of the tab based on;
a step of conveying the substrate in the longitudinal direction of the substrate and imaging a plurality of tabs in a next row with the plurality of imaging devices;
A method of manufacturing a semiconductor device comprising: 15. The method of manufacturing a semiconductor device according to claim 14, wherein the synthesized image is generated based on coordinate markers located in overlapping imaging fields of view. 16. The method of claim 15, wherein the coordinate markers are scales on a grid covering all fields of view of the imaging device,
The method of manufacturing a semiconductor device comprising generating the synthesized image by illuminating the coordinate markers, subjecting images to projective transformation based on the same intersection of the coordinate markers entering the overlapping imaging fields of view, and linking images between image pickup devices. delete 15. The method of claim 14, wherein the substrate further has feature markers;
and recalculating a projective transformation matrix for synthesizing and transforming an image based on the feature markers when displacement between the imaging devices is detected. 19 . The method of claim 18 , wherein the projection transformation matrix is recalculated based on coordinates of the previously measured fiducial marker. The method of claim 19, wherein the substrate is moved up and down to obtain a projection transformation matrix for each height,
A semiconductor device comprising: calculating an expected height of an alignment pattern position or an inspection visual field position from the substrate thickness, paste height, or die thickness, and selecting one of the projection transformation matrices held for each height based on the calculated expected height. manufacturing method. 21. The method of claim 20, wherein a height is measured by recognizing the same point on the substrate in overlapping imaging fields of view between adjacent imaging devices, and the projection transformation matrix maintained for each height is selected based on the measured height. , A method of manufacturing a semiconductor device. The method according to any one of claims 14 to 16 and 18 to 21, further comprising a step of applying a paste adhesive to the substrate,
The method of manufacturing a semiconductor device, wherein the imaging target is the applied paste adhesive. The method according to any one of claims 14 to 16 and 18 to 21, further comprising a step of bonding a die onto the substrate or an already bonded die,
The method of manufacturing a semiconductor device, wherein the imaging target is the bonded die.

Images