JP2020131335A - Dicing device and dicing method - Google Patents

Abstract

To provide a dicing device and a dicing method capable of stabilizing the dicing quality without being affected by the waviness of a work table surface and the variation in the thickness of a dicing tape.SOLUTION: A dicing device (10, 10A) for dicing by relatively moving a blade rotated by a spindle and a work table, with a workpiece held on the work table via a dicing tape, includes: a map creation unit that creates a calibration map indicative of a surface shape of the work table from measurement results of a surface shape of a calibration work mounted on the work table; a blade shape calculation unit that acquires an image of a groove formed on a surface of a blade determination work using the blade, and that calculates a shape of a blade tip from the image; and a control unit that controls the height of the blade on the basis of the calibration map and the shape of the blade tip.SELECTED DRAWING: Figure 1

Description

The present invention relates to a dicing apparatus and a dicing method, and more particularly to a dicing apparatus and a dicing method for dividing a workpiece such as a wafer on which a semiconductor device or an electronic component is formed into individual chips.

In a dicing device that divides a workpiece such as a wafer on which a semiconductor device or an electronic component is formed into individual chips, a blade that is rotated at high speed by a spindle, a work table that attracts and holds the workpiece, and a work table and a blade are used. It includes X, Y, Z, and θ drive units that change their relative positions. In this dicing device, dicing (cutting) is performed by cutting the work with the blade while relatively moving the blade and the work by each drive unit.

In a dicing device, it is an important factor to match the cutting amount of the blade with the set value, and in order to match the cutting amount of the blade with the set value, positioning of the Z axis, which is the cutting direction to the work, is repeated with high accuracy. It is necessary to detect and correct the wear of the blade.

For example, as described in Patent Document 1, conventionally, in positioning the Z-axis, the blade is first brought into contact with the upper surface of the work table to detect electrical continuity, and the center position of the blade at this time is used as a reference position for the Z-axis. Is in control. The operation of bringing this blade into contact with the upper surface of the work table to detect electrical continuity is called a cutter set. Further, in the blade wear correction, the blade and the work table are brought into contact with each other every time the work is machined for a set number of lines, and the reference position is corrected. However, this method brings the blade into contact with the work table, which can damage the blade.

In order to solve the problem of the contact type cutter set, for example, Patent Document 2 proposes an optical cutter set device. In this optical cutter set device, the blade is moved in the Z-axis direction orthogonal to the optical axis between the light projecting means and the light receiving means, and the detected light is gradually blocked by the blade to reach a preset light receiving amount. The blade is positioned with reference to the center position of the blade when the blade is closed.

Japanese Unexamined Patent Publication No. 2003-21135 Japanese Unexamined Patent Publication No. 2003-234309

By the way, there are three blade dicing methods: half-cut, semi-full-cut, and full-cut. In recent years, as the diameter of the workpiece has increased, the full-cut method for completely cutting the workpiece has become the mainstream. In the full-cut method, the work is placed on the work table with the work attached to the dicing tape, and the work is cut from the surface side of the work to the dicing tape by the blade while moving the blade and the work table relatively. Is divided into individual chips.

However, the above-mentioned conventional technique has a problem that it cannot sufficiently cope with dicing for performing a full cut of a work.

That is, in the case of the full cut method, it is necessary to control the cutting depth (cutting amount) of the blade so that the blade sufficiently penetrates the work and does not reach the work table.

For example, as shown in FIG. 18, when the work W is fully cut, the height H of the blade 90 (height from the surface of the work table to the center position of the blade 90) is the depth of cut D into the work W. Is positioned to be larger than the thickness M of the work W. Then, the work table (not shown) is cut and fed in the X direction with respect to the blade 90 rotated at high speed, so that a cutting groove 92 for one line is formed in the work W.

However, when the surface of the work table is wavy and the thickness of the dicing tape T varies, for example, as shown in FIG. 19, the dicing tape thickness K1 is the dicing tape thickness K shown in FIG. In the portion thinner than the above, the cutting depth D1 into the work W is smaller than the cutting depth D shown in FIG. That is, the blade 90 is shallowly cut with respect to the work W. In this case, the blade 90 is not sufficiently cut into the work W, which causes a cutting defect.

On the other hand, as shown in FIG. 20, in the portion where the dicing tape thickness K2 is thicker than the dicing tape thickness K shown in FIG. 18, the cutting depth D2 into the work W is the cutting depth D shown in FIG. Will be larger than. That is, the blade 90 is deeply cut with respect to the work W. In this case, the blade 90 cuts into the dicing tape T more than necessary. Therefore, the blade 90 is liable to be clogged by the adhesive on the surface of the dicing tape, which causes the blade 90 to become dull. Alternatively, it may cause tape breakage or the like in the expanding process performed after the dicing process.

In this way, even when the blade is set to a predetermined height, if there is a thickness variation in the dicing tape, the depth at which the blade cuts into the dicing tape varies, so that the processing quality is stable. There may be problems that do not occur. It should be noted that the same problem occurs not only due to the variation in the thickness of the dicing tape but also due to the waviness of the surface of the work table and the wear of the blade.

Particularly in recent years, the blade width of the blade tends to become thinner as the number of chips per wafer increases. As the blade width becomes thinner, the abrasive grains of the blade also decrease, so that the blade is likely to be clogged by the adhesive on the surface of the dicing tape, and the above-mentioned problem becomes more remarkable.

The present invention has been made in view of such circumstances, and is a dicing apparatus capable of stabilizing the processing quality without being affected by the waviness of the surface of the work table and the variation in the thickness of the dicing tape. The purpose is to provide a dicing method.

The following inventions are provided in order to achieve the above object.

The dicing apparatus according to the first aspect of the present invention is a dicing device for performing dicing by relatively moving a blade rotated by a spindle and a work table while holding a work to be machined on a work table via a dicing tape. A map creation unit that creates a calibration map showing the shape of the surface of the work table from the measurement results of the surface shape of the calibration work held on the work table, and a blade determination work using the blade. The blade shape determination unit that acquires the image of the groove formed on the surface of the blade and determines the tip shape of the blade from the image, the calibration map, and the blade tip shape when processing the work to be machined. It includes a control unit that controls the height.

In the dicing apparatus according to the second aspect of the present invention, in the first aspect, the control unit determines the state of wear of the blade from the shape of the tip of the blade, and the height of the blade is increased as the wear progresses. make low.

In the first or second aspect, the dicing apparatus according to the third aspect of the present invention further includes a cutting mark detecting portion for detecting cutting mark information formed in the surface region of the dicing tape to which the work is not attached. The control unit controls the height of the blade so that the depth of cut into the dicing tape is constant, based on the cutting mark information detected by the cutting mark detecting unit.

In the dicing apparatus according to the fourth aspect of the present invention, in the third aspect, the cutting mark detection unit calculates the cutting mark formation rate in the surface region of the dicing tape based on the cutting mark information, and the control unit cuts. Based on the cutting mark formation rate calculated by the mark detection unit, the height of the blade is controlled so that the cutting mark formation rate is within a certain range.

In the dicing method according to the fifth aspect of the present invention, a dicing process is performed by relatively moving a blade rotated by a spindle and a work table while holding a work to be machined on a work table via a dicing tape. This is a method, which is a calibration map creation process for creating a calibration map showing the surface shape of the work table from the measurement result of the surface shape of the calibration work held on the work table, and a blade determination using a blade. When processing the workpiece to be machined based on the blade shape determination process that acquires an image of the groove formed on the surface of the work piece and determines the tip shape of the blade from the image, the calibration map, and the tip shape of the blade. It is provided with a control process for controlling the height of the blade.

In the dicing method according to the sixth aspect of the present invention, the calibration work has a uniform thickness in the fifth aspect.

In the dicing method according to the seventh aspect of the present invention, in the fifth or sixth aspect, the blade determination work is a mirror-finished surface or a glass plate.

According to the present invention, it is possible to stabilize the processing quality without being affected by the waviness of the surface of the work table and the variation in the thickness of the dicing tape.

Schematic diagram showing the configuration of the dicing apparatus according to the present embodiment Floor plan showing the work Diagram showing an example of a calibration map Plan view showing the plan shape of the calf A table showing an example of the correction amount of the blade height for each judgment result of the tip shape of the blade. Schematic diagram showing how dicing is performed on the work Schematic diagram showing how the image pickup device images the dicing tape area. Schematic diagram showing how the line sensor camera images the dicing tape area The figure which showed an example of the correction table The figure which showed the concrete operation example in this embodiment The figure which shows the locus of the tip part (-Z side end part) of a blade 18. Flowchart showing the flow of blade height correction operation Flowchart showing the calibration map creation process Flow chart showing the blade shape determination process Flowchart showing dicing process Schematic diagram showing the configuration of the dicing apparatus according to another embodiment The figure which showed an example of the height graph generated by the cutting mark detection part. Diagram to explain conventional problems Diagram to explain conventional problems Diagram to explain conventional problems

Hereinafter, the dicing apparatus and dicing method according to the embodiment of the present invention will be described with reference to the accompanying drawings.

[Configuration of dicing device 10]
FIG. 1 is a schematic view showing the configuration of the dicing apparatus 10 according to the present embodiment.

As shown in FIG. 1, the dying device 10 includes a work table 12, a sub-table 13, a θ table 14, an X table 16, a blade 18, a spindle 20, a Y table (not shown), and a Z table. (Not shown), an image pickup device 22, a displacement sensor 25, and a control device 50 are provided.

In the present embodiment, a calibration map reflecting the waviness of the surface of the work table 12 is created before the dicing process. Further, before the dicing process, the tip shape of the blade 18 is determined. Then, when dicing is performed, the height of the blade 18 is set so that the depth of cut into the dicing tape T is constant based on the calibration map, the determination result of the tip shape of the blade 18, and the cutting mark formation rate described later. To control.

The X table 16 is provided on the upper surface of an X base (not shown). The X table 16 is configured to be movable in the X direction by an X drive unit (not shown) including a motor, a ball screw, and the like. A θ table 14 is placed on the X table 16, and a work table 12 is attached to the θ table 14. The θ table 14 is configured to be rotatable in the θ direction (rotation direction centered on the Z axis) by a rotation drive unit (not shown) including a motor and the like.

The Y table is provided on the side surface of the Y base (not shown). The Y table is configured to be movable in the Y direction by a Y drive unit (not shown) including a motor and a ball screw. A Z table (not shown) is attached to the Y table. The Z table is configured to be movable in the Z direction by a Z drive unit (not shown) including a motor and a ball screw. A spindle 20 with a built-in high-frequency motor, to which a blade 18 is attached to the tip, is fixed to the Z table.

With this configuration, the blade 18 is index-fed in the Y direction and cut-fed in the Z direction. Further, the work table 12 is rotated in the θ direction and cut and fed in the X direction.

The spindle 20 is rotated at high speed, for example, at 30,000 rpm to 60,000 rpm.

The blade 18 is a cutting blade configured in a thin disk shape. As the blade 18, a diamond abrasive grain, an electrodeposition blade obtained by electrodepositing CBN (Cubic form of Boron Nitride) abrasive grains with nickel, a resin blade bonded with a resin, or the like is used. The dimensions of the blade 18 are variously selected depending on the processing content, but when dicing a normal semiconductor wafer as a work W, those having a diameter of about 50 mm and a thickness of about 30 μm are used.

When creating the calibration map, the work table 12 attracts and holds the calibration work. As the calibration work, it is preferable to use a work having a uniform thickness (for example, a silicon wafer) having a plane shape substantially the same as or larger than the work W to be machined. .. The calibration work may be adsorbed and held on the surface of the work table 12 in a state of being attached to the same dicing tape T used for dicing of the work W to be processed.

The displacement sensor 25 measures the shape (unevenness) of each position on the surface of the calibration work while the calibration work is attracted and held on the work table 12. As the displacement sensor 25, for example, a laser displacement sensor, an optical or contact type displacement sensor, a TOF (Time of Flight) camera, or the like can be used. The displacement sensor 25 is movable relative to the work table 12. The displacement sensor 25 may be attached to the Y table together with the spindle 20, or may be separately provided with a drive mechanism for moving the displacement sensor 25 in the XYZ direction.

The data of the surface shape of the calibration work measured by the displacement sensor 25 is input to the map creation unit 58 of the control device 50. For example, the map creation unit 58 calculates the Z coordinates (measured values) of each position on the surface of the work table 12 when the surface (design value) of the work table 12 is the XY plane (Z = 0). The data of the calibration map is created and stored in the storage unit 54.

Next, determination of the tip shape of the blade 18 will be described. When determining the tip shape of the blade 18, the blade 18 forms a groove with respect to the blade determination work W0 that is attracted and held on the sub-table 13. Then, the image of the groove is imaged by the imaging device 22, and the tip shape of the blade 18 is determined based on the planar shape of the groove.

The sub-table 13 is fixed in the vicinity of the work table 12 by a holding member (arm), and can be moved integrally with the work table 12. The sub-table 13 attracts and holds the blade determination work W0.

The image pickup apparatus 22 is arranged at a position facing the work table 12. The image pickup apparatus 22 images the surface of the work W in order to determine the tip shape of the blade 18, align the work W, and evaluate the processing state (calf check). The image pickup device 22 is an example of the image pickup device (alignment camera) of the present invention.

The image pickup apparatus 22 is composed of a microscope, a camera, or the like, and images the surface of the work W at a high magnification (for example, 8.0 times) or a low magnification (for example, 1.0 times) by a method such as switching the lens of the microscope. It is possible. As the camera, an area sensor camera is used.

The image pickup apparatus 22 is fixed to the spindle 20 via the holding member 24, and can move in the Y direction and the Z direction integrally with the spindle 20. Further, two stages may be provided to individually control the positions of the microscope and the spindle 20 of the imaging device 22. In this case, a drive shaft of the microscope is provided separately from the drive shaft of the spindle 20, and the spindle 20 and the microscope are individually moved along the respective drive shafts.

When determining the tip shape of the blade 18, the Z table moves the blade 18 in the −Z direction with respect to the blade determination work W0 that is attracted and held by the sub-table 13. Then, the tip end portion (blade edge) of the blade 18 is brought into contact with the blade determination work W0 substantially perpendicularly to form a groove (calf) having a predetermined depth that does not penetrate the blade determination work W0 (chop cut or Chop processing). After that, the blade 18 is moved in the + Z direction by the Z table. During this time, the relative position of the sub-table 13 with respect to the blade 18 remains unchanged. Note that the relative position may be corrected so that the relative position of the sub-table 13 with respect to the blade 18 is always constant according to the temperature characteristics of both the blade 18 and the sub-table 13.

The image pickup device 22 includes a light projecting unit that irradiates the image pickup target with light, irradiates the blade determination work W0 with light, receives the reflected light from the blade determination work W0, and obtains a flat image of the calf. Take an image. The plane image of the calf is input to the control unit 56, and the control unit 56 determines the tip shape of the blade 18. Here, as the blade determination work W0, one having a mirror-finished surface (mirror work) or one having a high surface reflectance such as a glass plate is used. As a result, the contour of the calf becomes clear due to the difference in reflectance between the surface of the blade determination work W0 and the calf, so that the tip shape of the blade 18 can be accurately determined. The control unit 56 and the image pickup device 22 are examples of the blade shape determination unit of the present invention.

At the time of dicing, the work table 12 attracts and holds the work W to be processed. The work W is attached to the frame F via a dicing tape Т having an adhesive on the surface, and is adsorbed and held on the work table 12. The frame F to which the dicing tape T is attached is held by the frame holding means (not shown) arranged on the work table 12.

FIG. 2 is a plan view showing the work W to be machined. As shown in FIG. 2, processing lines (scheduled division lines) S are formed in a grid pattern on the surface of the work W, and devices are formed in a plurality of regions (device formation regions) C partitioned by these processing lines S. Is formed.

The control device 50 controls the operation of each part of the dicing device 10. The control device 50 is realized by a general-purpose computer such as a personal computer or a microcomputer.

The control device 50 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a hard disk, and the like. In the control device 50, various programs such as a control program stored in the ROM are expanded in the RAM, and the programs expanded in the RAM are executed by the CPU, so that each part shown in the control device 50 in FIG. 1 The function is realized.

The control device 50 functions as a cutting mark detection unit 52, a storage unit 54, a map creation unit 58, and a control unit 56.

The control unit 56 controls the operation of each unit of the control device 50. Specifically, the control unit 56 controls the cutting feed of the work table 12 in the X direction via the X drive unit and the rotation of the work table 12 in the θ direction via the θ drive unit. Further, the control unit 56 controls the index feed of the spindle 20 in the Y direction via the Y drive unit and the cut feed of the spindle 20 in the Z direction via the Z drive unit. Further, the control unit 56 controls the rotation operation of the spindle 20 and the imaging operation of the imaging device 22.

The storage unit 54 stores various data necessary for the operation of the dicing device 10. The various data stored in the storage unit 54 include data related to the work W, data related to alignment, and the thickness of the dicing tape T. For example, the data related to the work W includes a product type number, a material, an external dimension, a thickness, a chip size, and the like. Further, the various data stored in the storage unit 54 include data of the calibration map and data of the determination result of the tip shape of the blade 18. Further, the various data stored in the storage unit 54 include a cutting mark formation rate Q and a correction table (see FIG. 9), which will be described later.

The cutting mark detection unit 52 receives the image data captured by the image pickup apparatus 22 during the dicing process, and detects the cutting mark information described later by image processing or the like.

[Action of dicing device 10]
Next, the operation of the dicing device 10 configured in this way will be described.

First, the work W to be processed, which is attached to the frame F via the dicing tape Т, is conveyed by a conveying means (not shown) and placed on the work table 12. The work W placed on the work table 12 is imaged by the image pickup apparatus 22, and an alignment operation for relatively aligning the work W and the blade 18 is started.

When the alignment operation is completed, the spindle 20 is activated to rotate the blade 18, and cutting water and cooling water are supplied from various nozzles (not shown) provided on the wheel cover (not shown) covering the blade 18. In this state, the work table 12 is cut and fed in the X direction, the spindle 20 is cut and fed in the Z direction, and the work W is diced along the machining line S. Every time one line is machined, the spindle 20 is indexed in the Y direction, and when the machining in one direction is completed, the work table 12 is rotated 90 degrees and the work W is diced in a grid pattern.

In the present embodiment, the calibration map is created and the tip shape of the blade 18 is determined before the dicing process for the work W to be processed. Then, during the dicing process of the work W, a cutting mark detection operation and a blade height correction operation, which will be described later, are performed for each processing line S.

In the present embodiment, the cutting mark detection operation and the blade height correction operation, which will be described later, are performed for each processing line S, but the present invention is not limited to this, and each processing line number specified by the user (for example, for example). It may be carried out every 5 processing lines). Further, it may be performed on a machining line S designated by the user (for example, a machining line S such as the first line, the fifth line, the seventh line, etc.).

(Creation of calibration map)
Next, the creation of the calibration map will be described.

In the present embodiment, the shape (unevenness) of the surface of the work table 12 on which the work W to be processed is placed is measured in advance to create a calibration map. Then, the height of the blade 18 is controlled using this calibration map. As a result, the depth of cut of the blade 18 into the dicing tape T can be made constant.

At the time of creating the calibration map, the shape (unevenness) of each position on the surface of the calibration work attracted and held by the work table 12 is measured while scanning the displacement sensor 25 in the XY directions. At this time, the displacement sensor 25 is scanned in the XY directions by the X table 16 and the Y table.

The map creation unit 58 obtains the height position (Z coordinate) of the surface of the calibration work from the detection signal output from the displacement sensor 25. Then, the map creation unit 58 subtracts the thickness of the calibration work from the height position (Z coordinate) of the surface of the calibration work to calculate the height position (Z coordinate) of each position on the surface of the work table 12. Then, a calibration map showing the shape of each position on the surface of the work table 12 is created.

FIG. 3 is a diagram showing an example of a calibration map. In the example shown in FIG. 3, the shape of each position on the surface of the work table 12 is represented as the Z coordinate of each position on the surface of the work table 12 when the work table 12 is in the XY plane (Z = 0). ..

In the present embodiment, a calibration work having substantially the same shape as the work W to be machined or a calibration work larger than the work W to be machined is used, and the calibration work is scanned while scanning the displacement sensor 25 in the XY directions. Create a calibration map of the entire surface that is attracted and held. Then, the height of the blade 18 with respect to the work W to be machined is controlled by using this calibration map (see FIG. 11). As a result, the height of the blade 18 can be controlled according to the shape of the surface of the work table 12 on which the work W to be processed is attracted and held.

(Judgment of blade tip shape)
Next, determination of the tip shape of the blade 18 will be described.

As the wear of the blade 18 progresses, the work W to be machined is cut, so that the cutting edge portion formed in an annular shape on the blade 18 becomes thinner and the diameter of the blade 18 becomes smaller. Then, the amount of cutting on the back surface (the surface to which the dicing tape T is attached) of the work W to be processed decreases. Therefore, if the same height control as that of the blade 18 before wear is performed, the cutting amount of the work W on the back surface side is reduced, so that the cutting of the work W becomes insufficient and the back surface side of the work W is used. Is chipped, and cracks (chipping) are likely to occur. Therefore, in the present embodiment, the shape of the tip of the blade 18, that is, the shape of the outer peripheral portion including the outer peripheral end of the cutting edge portion of the blade 18, is determined before the dicing process. Then, in addition to the calibration map, the height of the blade 18 during dicing is controlled based on the determination result of the tip shape of the blade 18. This makes it possible to prevent chipping due to wear of the blade 18.

To determine the tip shape of the blade 18, a groove (calf) is formed on the surface of the blade determination work W0 which is attracted and held by the sub-table 13 by using the blade 18. Then, the control unit 56 takes an image of the calf with the image pickup device 22, and determines the tip shape of the blade 18 based on the plane shape of the imaged calf.

Specifically, when the blade 18 is not worn, the variation in the width of the calf is small with respect to the length direction of the calf, and the shape of both ends of the calf becomes substantially rectangular (see calf C1 in FIG. 4). ). On the other hand, as the wear of the blade 18 progresses, the thickness of the cutting edge portion of the blade 18 becomes thin, so that the width of the calf becomes narrower and becomes triangular as it approaches both ends of the calf. Then, as the blade 18 wears, the width of both ends of the calf becomes narrower and the angles of both ends of the calf become smaller (see calfskin C2 and C3 in FIG. 4).

Further, the tip shape of the blade 18 may be determined by measuring the depth of the calf using a displacement sensor 25 or the like. When the blade 18 is not worn, the variation in the depth of the calf is small in the width direction of the calf. On the other hand, as the wear of the blade 18 progresses, the thickness of the cutting edge portion of the blade 18 becomes thinner or the diameter of the blade 18 decreases, so that the variation in the depth of the calf increases in the width direction of the calf. Also, the calf becomes shallower.

In the present embodiment, the tip shape of the blade 18 is determined by utilizing the characteristics of the calf formed on the blade determination work W0. For the determination of the tip shape of the blade 18, for example, the blade diagnostic method described in JP-A-2017-164843 can be applied. Further, the method for determining the tip shape of the blade 18 is not limited to the one using the calf image. For example, it is possible to determine the tip shape of the blade 18 by photographing the tip portion including the outer peripheral end of the cutting edge portion of the blade 18 from a direction parallel to the blade 18 using a camera and analyzing this image. is there. For example, the image of the tip of the blade 18 may be accumulated, and the shape of the tip of the blade 18 may be determined from the time course of the image of the tip of the blade 18.

FIG. 4 is a plan view showing the plan shape of the calf. FIG. 5 is a table showing an example of the correction amount of the height of the blade 18 for each determination result of the tip shape of the blade 18.

The calf shown by reference numeral C1 in FIG. 4 has a substantially rectangular planar shape at both ends, and the amount of wear of the blade 18 is minimized. In this case, as shown in FIG. 5, the amount of cut of the blade 18 into the dicing tape T is added by 5 μm.

On the other hand, the calf shown by the reference numeral C3 in FIG. 4 has a substantially triangular shape with acute angles at both ends, and the amount of wear of the blade 18 is maximum. In this case, as shown in FIG. 5, the amount of cut of the blade 18 into the dicing tape T is added by 15 μm.

Further, the calf shown by reference numeral C2 in FIG. 4 has a substantially triangular shape with acute angles at both ends, but the angles at both ends are smaller than C3, and the amount of wear of the blade 18 is between the examples of C1 and C3. is there. In this case, as shown in FIG. 5, the amount of cut of the blade 18 into the dicing tape T is added by 7.5 μm.

In the present embodiment, the larger the amount of wear of the blade 18, the lower the height of the blade 18 during dicing processing and the larger the amount of cut into the dicing tape T. This makes it possible to prevent the occurrence of chipping due to the wear of the blade 18.

It is not necessary to create the calibration map and determine the tip shape of the blade 18 every time the dicing process is performed. For example, the dicing process may be performed every time the dicing process is performed a predetermined number of times. Further, since it is considered that the wear of the blade 18 is more likely to occur than the change of the surface shape of the work table 12, the calibration map is created less frequently than the determination of the tip shape of the blade 18. May be good.

(Cutting mark detection operation)
Next, the cutting mark detection operation will be described.

FIG. 6 is a schematic view showing how the dicing process is performed on the work W.

In the present embodiment, as shown in FIG. 6, the blade 18 is dicing while moving relative to the work W from the cutting start position P1 on one side to the cutting end position P4 on the other side with the work W in between. Processing is performed. At this time, a cutting groove 26 is formed in the work W, and the cutting is performed by the blade 18 in the surface region R (hereinafter, referred to as “dicing tape region R”) of the dicing tape T to which the work W is not attached. Traces 28 are formed.

For example, when the cutting mark 28 in the dicing tape region R is short and shredded as a whole (see FIG. 6), or when the cutting mark 28 does not exist at all, the blade 18 is shallowly cut. It has become. In this case, the blade 18 does not sufficiently cut into the work W, which causes a cutting defect.

On the other hand, when the cutting marks 28 in the dicing tape region R are connected for a long time as a whole, the blade 18 is deeply cut. In this case, the sharpness may deteriorate due to clogging of the blade 18.

As a result of repeated studies, the present inventor may correct the height (position in the Z direction) of the blade 18 based on the cutting mark formation rate Q in the dicing tape region R in order to stabilize the processing quality. I found a good thing. Here, the cutting mark formation rate Q is the ratio of the region (length in the cutting feed direction) in which the cutting marks 28 are formed in the dicing tape region R to the whole (total length in the cutting feed direction of the dicing tape region R). To say.

The cutting mark detection operation in the present embodiment is performed by imaging the dicing tape region R with the image pickup device 22.

FIG. 7 is a schematic view showing how the image pickup apparatus 22 images the dicing tape region R. As shown in FIG. 7, the image pickup apparatus 22 images the dicing tape area R at a plurality of locations at a high magnification while shifting the position with respect to the dicing tape area R in the X direction, and divides the dicing tape area R into a plurality of divided images. Take an image of the whole. The image data captured by the image pickup device 22 is output to the control device 50. The image pickup apparatus 22 may image a wide range of the dicing tape region R at a time with a low magnification.

In the present embodiment, the area sensor camera is used as the camera constituting the image pickup apparatus 22, but the present invention is not limited to this, and for example, a line sensor camera may be used.

FIG. 8 is a schematic view showing how the line sensor camera 30 images the dicing tape region R. As shown in FIG. 8, in the line sensor camera 30, a plurality of light receiving elements (not shown) are arranged in a row in the Y direction, and the entire dicing tape region R is imaged while scanning in the X direction. It has become.

The cutting mark detecting unit 52 detects the length (length in the cutting feed direction) of the cutting mark 28 formed in the dicing tape region R by performing image processing on the image data captured by the imaging device 22 by a known method. To do. Further, the cutting mark detection unit 52 calculates the cutting mark formation rate Q in the dicing tape region R based on the length of the detected cutting mark 28.

Here, a method of calculating the cutting mark formation rate Q in the dicing tape region R will be described.

In FIG. 6, when the blade 18 moves relative to the work W from the cutting start position P1 to the cutting end position P4, the position where the blade 18 starts cutting into the work W is set as the work entry position P2, and the blade 18 works. The position at which the cutting into W is completed is defined as the work exit position P3.

At this time, the cutting marks 28 are formed in the dicing tape area R1 between the cutting start position P1 and the work entry position P2, and in the dicing tape area R2 between the work exit position P3 and the cutting end position P4, respectively.

Further, the total length of the dicing tape areas R1 and R2 in the cutting feed direction (X direction) is L1 and L2, respectively, and the total length of the cutting marks 28 formed in the dicing tape areas R1 and R2 is l1. Let l2. For example, as shown in FIG. 6, when a plurality of cutting marks 28 are formed in the dicing tape region R1, the total length of each cutting mark 28 is set to l1. The same applies to the dicing tape area R2.

The cutting mark detection unit 52 calculates the cutting mark formation rate Q by the following formula (1). The cutting mark formation rate Q is stored in the storage unit 54 in association with the machining line information (machining line number or the like) when the cutting mark 28 is formed. The unit of the cutting mark formation rate Q is%.

Q = {(l1 + l2) / (L1 + L2)} × 100 ... (1)
(Blade height correction operation)
Next, the blade height correction operation will be described.

In the blade height correction operation, the control unit 56 acquires the cutting mark formation rate Q of the reference line from the storage unit 54. The reference line is a machining line S in which dicing is performed immediately before the current machining line S, and the cutting mark formation rate Q has already been calculated by the above-mentioned cutting mark detection operation.

The control unit 56 further determines the blade height correction amount G corresponding to the cutting mark formation rate Q of the reference line according to the correction table (see FIG. 9) stored in the storage unit 54. Then, the control unit 56 controls the height (position in the Z direction) of the blade 18 based on the determined blade height correction amount G.

(Correction table)
FIG. 9 is a diagram showing an example of a correction table. As shown in FIG. 9, the correction table shows the correspondence between the cutting mark formation rate Q and the blade height correction amount G. This correction table also includes a target cutting mark formation rate (target formation rate) and an allowable range thereof (target formation rate allowable range). Each numerical value in the correction table can be appropriately set by the user via an operation unit (not shown) connected to the control device 50. Further, the blade height correction amount G indicates a correction in the direction in which the depth of cut from the set value to the work W becomes deeper when the value is positive, and the set value when the value is negative. The correction is shown in the direction in which the depth of cut from the work W to the work W becomes shallower.

(Specific operation example)
FIG. 10 is a diagram showing a specific operation example in the present embodiment. FIG. 11 is a diagram showing a locus of the tip end portion (−Z side end portion) of the blade 18. Here, in order to simplify the description, a case where the four machining lines S1 to S4 along the cutting feed direction (X direction) are diced in the order of the line numbers will be described.

As shown in FIG. 10, first, dicing processing is performed on the first processing line S1. In this case, since the reference line does not exist, the control unit 56 sets the blade height correction amount G to 0 and leaves the height of the blade 18 as the set value without correcting it. Then, the dicing process is started from the cut side of the work W to be processed.

When the blade 18 reaches the work W to be machined, the control unit 56 performs dicing while correcting the height of the blade 18 based on the calibration map and the determination result of the tip shape of the blade 18. Reference numeral 18T in FIG. 11 indicates a locus of the tip portion of the blade 18. Specifically, as shown in FIG. 11, the control unit 56 detects that the cutting edge of the blade 18 has come into contact with the upper end of the work W to be machined based on a change in torque applied to the spindle 20 or the like. .. In the control unit 56, the cutting edge of the blade 18 contacts the upper end of the work W to be machined based on the positional relationship between the work W to be machined and the blade 18 detected by the imaging device 22 or a position sensor or the like. You may try to detect what you have done. Then, the blade 18 is gradually pushed down to the −Z side while scanning in the X direction. Then, the height correction amount of the blade 18 based on the calibration map and the correction amount based on the determination result of the tip shape of the blade 18 (see FIG. 5) are added to the blade height correction amount G to obtain the height position. As described above, the dicing process is performed while controlling the position of the blade 18.

Next, when the control unit 56 detects that the blade 18 has separated (cut through) from the work W to be machined, the control unit 56 gradually pulls the blade 18 toward the + Z side, and the blade height correction amount G (first machining line) In S1, the height control is shifted to only G = 0). Here, it is possible to detect the cutout based on the change in the torque applied to the spindle 20 or the positional relationship between the work W to be machined and the blade 18.

The cutting mark detection unit 52 calculates the cutting mark formation rate (23% in this example) of the first processing line S1 based on the image data captured by the imaging device 22 on the cut-out side of the work W to be processed. It is stored in the storage unit 54.

Next, a dicing process is performed on the second processing line S2. In this case, the control unit 56 acquires the cutting mark formation rate Q (23%) of the first machining line S1, which is the reference line, from the storage unit 54. Then, the control unit 56 determines the blade height correction amount G to +0.006 mm, controls the height of the blade 18 according to the blade height correction amount G, and dices from the cut side of the work W to be machined. To start.

When the blade 18 reaches the work W to be machined, the blade 18 is gradually pushed down to the −Z side while scanning in the X direction. Then, the height correction amount of the blade 18 based on the calibration map and the correction amount based on the determination result of the tip shape of the blade 18 (see FIG. 5) are added to the blade height correction amount G to obtain the height position. As described above, the dicing process is performed while controlling the position of the blade 18. When the control unit 56 detects that the blade 18 is separated (cut through) from the work W to be machined, the blade 18 is gradually pulled up to the + Z side to shift to height control using only the blade height correction amount G. ..

The cutting mark detection unit 52 determines the cutting mark formation rate (78% in this example) of the second processing line S2 on the cut-out side (+ X side) of the work W to be processed based on the image data captured by the imaging device 22. It is calculated and stored in the storage unit 54.

Next, dicing processing is performed on the third processing line S3. In this case, the control unit 56 acquires the cutting mark formation rate (78%) of the second processing line S2, which is the reference line, from the storage unit 54. At this time, since the cutting mark formation rate (78%) exceeds the allowable range (± 5%) of the target ratio (70%), the control unit 56 determines the blade height correction amount G to be −0.001 mm. Then, the height of the blade 18 is controlled according to the blade height correction amount G, and the dicing process is started from the cut side of the work W to be machined.

When the blade 18 reaches the work W to be machined, the correction amount of the height of the blade 18 based on the calibration map and the correction amount based on the determination result of the tip shape of the blade 18 (see FIG. 5) as in the above case. ) Is added to the blade height correction amount G to obtain a height position, and the dicing process is performed while controlling the position of the blade 18. When the control unit 56 detects that the blade 18 is separated (cut through) from the work W to be machined, the control unit 56 shifts to height control using only the blade height correction amount G.

The cutting mark detection unit 52 determines the cutting mark formation rate (73% in this example) of the third processing line S3 on the cut-out side (+ X side) of the work W to be processed based on the image data captured by the imaging device 22. It is calculated and stored in the storage unit 54.

Next, dicing processing is performed on the fourth processing line S4. In this case, the control unit 56 acquires the cutting mark formation rate (73%) of the third processing line S3, which is the reference line, from the storage unit 54. At this time, since the cutting mark formation rate (73%) is within the allowable range (± 5%) of the target ratio (70%), the control unit 56 determines that the blade height correction amount G is 0 mm, and the blade thereof. The height of the blade 18 is controlled according to the height correction amount G. Since the subsequent control is the same as that of the first to third machining lines S1 to S3, the description thereof will be omitted.

In the present embodiment, the blade height correction amount of the blade 18 is controlled based on the calibration map, the tip shape of the blade 18, and the cutting mark formation rate, but the present invention is limited to this. is not. For example, the blade 18 may be offset to the −Z side by a predetermined amount in order to reliably perform the dicing process on the work W regardless of the wear condition of the blade 18.

(flowchart)
FIG. 12 is a flowchart showing the flow of the blade height correction operation. As shown in FIG. 12, in the present embodiment, prior to the dicing process (step S30), the calibration map is created (step S10) and the tip shape of the blade 18 is determined (step S20).

(Proofreading map creation process)
FIG. 13 is a flowchart showing a calibration map creation process.

First, as shown in FIG. 13, the calibration work is attracted and held (installed) on the surface of the work table 12 via the dicing tape T (step S100), and the shape of the surface of the calibration work is measured by the displacement sensor 25. ..

Next, the map creation unit 58 calculates the surface shape of the work table 12 based on the measurement result of the surface shape of the calibration work (step S102). Then, the map creation unit 58 creates the data of the calibration map showing the shape of the surface of the work table 12 and stores it in the storage unit 54 (step S104).

(Blade shape determination process)
FIG. 14 is a flowchart showing a blade shape determination process.

First, the blade determination work W0 is sucked and held on the sub-table 13. Then, the blade 18 is brought into contact with the blade determination work W0 to form a calf having a predetermined depth on the surface of the blade determination work W0 (step S200).

Next, the image pickup device 22 captures a flat image of the calf (step S202). The cutting mark detection unit 52 extracts the contour of the calf from the plan image of the calf and measures the dimensions of the calf (step S204). Then, the control unit 56 calculates the shape of the tip portion of the blade 18 based on the dimensions of the calf (step S206).

(Dicing process)
FIG. 15 is a flowchart showing a dicing process.

First, the work W to be processed is sucked and held on the surface of the work table 12. Then, cutting is started from the first machining line S (i) (i = 1) (step S300).

In cutting the machining line S (i), the blade height correction amount G is determined on the cutting side of the work W based on the cutting mark formation rate, and the machining line S (i) is controlled while controlling the height of the blade 18. ) Is cut (step S302).

When it is detected that the blade 18 has reached the work W to be machined (Yes in step S304), the control unit 56 adds a calibration map and a calibration map in addition to the blade height correction amount G determined based on the cutting mark formation rate. The machining line S (i) is cut while controlling the height of the blade 18 based on the tip shape of the blade 18 (step S306: control step). Then, when it is detected that the work W has been cut through (Yes in step S308), the control unit 56 shifts to the height control of the blade 18 based on the cutting mark formation rate (step S310).

Next, the cutting mark detection unit 52 detects the cutting mark information on the cut-out side of the machining line S (i) (Yes in step S312, step S314). Then, the cutting mark detection unit 52 calculates the cutting mark formation rate Q in the dicing tape region R based on the detected cutting mark information. Then, the cutting mark detection unit 52 stores the calculated cutting mark formation rate Q in the storage unit 54 in association with the information (line information) related to the machining line S.

Next, with i = i + 1, the user moves to the next machining line S (i) (step S316) and performs cutting (steps S302 to S312). Then, steps S302 to S312 are repeated, and when the cutting of all the machining lines S (i) (see FIG. 2) in the XY directions is completed (Yes in step S312), the dicing process is completed.

[Action and effect of this embodiment]
Next, the action and effect of this embodiment will be described.

According to the present embodiment, the height of the blade 18 is controlled based on the shape of the surface of the work table 12 on which the work W to be processed is attracted and held, and the progress of wear of the blade 18 is taken into consideration. The depth of cut of the blade 18 is adjusted. Further, according to the present embodiment, the cutting depth into the dicing tape T is constant based on the cutting mark information in the dicing tape region R (the surface region of the dicing tape T to which the work W is not attached). The height of the blade 18 is controlled. As a result, the depth at which the blade 18 cuts into the dicing tape T can be made relatively shallow and constant without being affected by the waviness of the work table 12 and the thickness variation of the dicing tape T. The quality can be stabilized. Further, since the cutting depth of the blade 18 is adjusted in consideration of the progress of wear of the blade 18, it is possible to prevent the occurrence of chipping on the back surface side of the work W.

Further, according to the present embodiment, the cutting mark detecting unit 52 detects the cutting mark information in the dicing tape region R based on the image data captured by the imaging device 22. The image pickup apparatus 22 is composed of an alignment camera and is originally provided in the dicing apparatus 10, so that there is no problem that the apparatus configuration is complicated and the cost is increased.

In the present embodiment, the cutting mark information is detected based on the image data captured by the imaging device 22, but the present invention is not limited to this, and for example, a distance measuring device 32 (see FIG. 16) described later is used. It may be used to detect cutting mark information. Further, the cutting mark information may be detected visually.

Further, when the number of machining lines (for example, 2067 lines) is large, the amount of blade wear until the dicing of all the machining lines S is completed cannot be ignored. Therefore, conventionally, it has been necessary to perform the cutter set many times during the dicing process to measure the current amount of blade wear and adjust the height of the blade 18. On the other hand, in the present embodiment, during the dicing process of the work W, a cutting mark detection operation is performed for each processing line S. Therefore, since the current blade wear amount can be grasped in real time, it is not necessary to frequently perform the conventional blade wear amount measurement operation, and there is an advantage that the time until the dicing process is completed can be reduced. is there.

Further, in the present embodiment, when dicing the current machining line S, the blade height correction amount G is determined using the cutting mark formation rate Q with the immediately preceding machining line S as a reference line. However, the present invention is not limited to this, and a dicing line S temporally or spatially adjacent to or close to the current dicing line S may be used as a reference line. The adjacent machining line S means a machining line that is temporally or spatially adjacent to the current machining line S. Further, the adjacent machining line S means a machining line S within a range of several lines (for example, 1 to 5 lines) temporally or spatially from the current machining line. Further, the reference line is not limited to one processing line S, and may be a plurality of processing lines S. In this case, for example, the blade height correction amount G may be determined based on the average value of the cutting mark formation rates Q corresponding to each of the plurality of machining lines S.

Further, in the present embodiment, there is a possibility that the cutting quality of the dicing processing lines S (first 1 to 3 lines) may deteriorate until the height of the blade 18 becomes stable. This problem can be solved by performing empty cutting on the dicing tape T several times before actually starting the dicing process, adjusting the height of the blade 18 in advance, and then performing the dicing process.

Further, in the present embodiment, in the cutting mark detection operation, the lengths of the cutting marks 28 of both the dicing tape area R1 on the cutting side and the dicing tape area R2 on the cutting side are measured. Depending on the situation, the following operations can be considered.

(For workpieces with a high cutting load)
Since the blade 18 wears more during the dicing process, only the length of the cutting mark 28 existing in the dicing tape region R2 on the cut-out side is measured. This is because the blade 18 is severely worn during the dicing process, and even if the length of the cutting mark 28 in the dicing tape region R1 on the cutting side is measured, it cannot be used as a reference for correcting the height of the blade 18.

(If you want to improve the overall throughput)
Only the cutting marks 28 of the dicing tape area R1 on the cut-out side or the dicing tape area R2 on the cut-out side are measured, and the time required for the cutting mark detection operation is shortened as compared with the case of measuring both sides. In this case, since it is not necessary to leave the cutting mark 28 on the dicing tape T on the side where the cutting mark 28 is not measured, the work W can be cut at the position where it can be cut, and the cutting time can be further shortened.

(When considering the rigidity of the dicing tape T)
The rigidity of the dicing tape T may change depending on the temperature. Specifically, when the temperature of the cutting environment is high, the rigidity of the dicing tape T becomes low. When the rigidity of the dicing tape T becomes low, when cutting the work W using the blade 18, the work W may be pushed toward the −Z side by the blade 18 and the cutting may not be sufficiently performed. Therefore, in addition to the above conditions, the pushing down amount of the blade 18 may be adjusted in consideration of the temperature dependence of the rigidity of the dicing tape T.

Specifically, the temperature dependence of the rigidity of the dicing tape T (for example, the relationship between the temperature and the amount of the work W pushed in) is experimentally obtained in advance, and the work W is subjected to the temperature of the environment in which the cutting is performed. The amount of pushing may be estimated, and the amount of pushing down of the blade 18 may be added. For example, when the temperature of the cutting environment is high, the rigidity of the dicing tape T is low. Therefore, the higher the temperature of the cutting environment, the larger the addition value of the pushing down amount of the blade 18. On the other hand, the lower the temperature of the cutting environment, the smaller the added value of the pushing amount of the blade 18.

[Other Embodiments]
FIG. 16 is a schematic view showing the configuration of the dicing apparatus 10A according to another embodiment. In FIG. 16, components common to those in FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted.

As shown in FIG. 16, the dicing device 10A according to another embodiment includes a distance measuring device 32 in addition to the configuration of the dicing device 10 according to the present embodiment.

The distance measuring device 32 is arranged at a position facing the work table 12. The distance measuring device 32 measures the distance to the surface of the dicing tape region R, which is the object to be measured, and is composed of, for example, a laser displacement meter or an interference microscope. The measurement result of the distance measuring device 32 is output to the cutting mark detection unit 52 as distance data.

Further, the distance measuring device 32 is fixed to the side surface of the image pickup device 22, and can move in the Y direction and the Z direction integrally with the spindle 20 and the image pickup device 22.

With this configuration, as a cutting mark detection operation performed while the work W is dicing, the distance measuring device 32 measures the distance to the dicing tape area R while shifting the position with respect to the dicing tape area R in the X direction. The distance data, which is the measurement result of the distance measuring device 32, is output to the cutting mark detection unit 52.

The cutting mark detection unit 52 generates a height graph showing a change in height (concavo-convex state) in the dicing tape region R based on the distance data acquired from the distance measuring device 32.

FIG. 17 is a diagram showing an example of a height graph generated by the cutting mark detection unit 52. As shown in FIG. 17, the cutting mark detecting unit 52 sets a region lower than a predetermined threshold height (height shown by a broken line in FIG. 17) in the generated height graph as a cutting mark forming region K (cutting mark). It is detected as a region where 28 is formed). Then, the cutting mark detection unit 52 calculates the cutting mark formation rate Q based on the detected cutting mark forming region K. Subsequent processing is the same as in this embodiment.

According to another embodiment, the cutting mark information in the dicing tape region R (the surface region of the dicing tape T to which the work W is not attached) can be detected based on the measurement result of the distance measuring device 32. It is possible to stabilize the processing quality in the same manner as in the present embodiment described above.

Although an example of the present invention has been described in detail, the present invention is not limited to this, and it goes without saying that various improvements and modifications may be made without departing from the gist of the present invention.

10 ... Dicing device, 12 ... Work table, 13 ... Sub table, 14 ... θ table, 16 ... X table, 18 ... Blade, 20 ... Spindle, 22 ... Imaging device, 24 ... Holding member, 26 ... Cutting groove, 28 ... Cutting mark, 30 ... Line sensor camera, 32 ... Distance measuring device, 50 ... Control device, 52 ... Cutting mark detection unit, 54 ... Storage unit, 56 ... Control unit, 58 ... Map creation unit, 90 ... Blade, 92 ... Cutting Groove, W ... Work, W0 ... Blade judgment work, T ... Dicing tape, Q ... Cutting mark formation rate, G ... Blade height correction amount

Claims (7)

A dicing device that performs dicing by relatively moving a blade that is rotated by a spindle and the work table while holding the work to be machined on the work table via dicing tape.
A map creation unit that creates a calibration map showing the shape of the surface of the work table from the measurement results of the surface shape of the calibration work held on the work table.
A blade shape determination unit that acquires an image of a groove formed on the surface of the blade determination work using the blade and determines the tip shape of the blade from the image.
A control unit that controls the height of the blade when the workpiece to be machined is machined based on the calibration map and the tip shape of the blade.
A dicing device equipped with. The control unit determines the state of wear of the blade from the shape of the tip of the blade, and lowers the height of the blade as the wear progresses.
The dicing apparatus according to claim 1. Further provided with a cutting mark detecting portion for detecting cutting mark information formed in the surface region of the dicing tape to which the work is not attached.
The control unit controls the height of the blade so that the depth of cut into the dicing tape is constant based on the cutting mark information detected by the cutting mark detecting unit.
The dicing apparatus according to claim 1 or 2. The cutting mark detection unit calculates the cutting mark formation rate in the surface region of the dicing tape based on the cutting mark information.
The control unit controls the height of the blade so that the cutting mark formation rate is within a certain range based on the cutting mark formation rate calculated by the cutting mark detection unit.
The dicing apparatus according to claim 3. This is a dicing method in which a blade rotated by a spindle and the work table are relatively moved to perform dicing while the work to be machined is held on the work table via a dicing tape.
A calibration map creation process for creating a calibration map showing the surface shape of the work table from the measurement results of the surface shape of the calibration work held on the work table.
A blade shape determination step of acquiring an image of a groove formed on the surface of a blade determination work using the blade and calculating the tip shape of the blade from the image.
A control step for controlling the height of the blade when machining the workpiece to be machined based on the calibration map and the tip shape of the blade.
A dicing method that includes. The calibration work has a uniform thickness.
The dicing method according to claim 5. The blade determination work has a mirror-finished surface or a glass plate.
The dicing method according to claim 5 or 6.

Images