JP7722866B2 - Wafer processing method - Google Patents

Description

The present invention relates to a wafer processing method for thinning a circular wafer having a chamfered outer periphery to a predetermined thickness.

The device chip manufacturing process uses a disk-shaped wafer on which devices such as ICs (Integrated Circuits) and LSIs (Large Scale Integration) are formed in multiple areas defined by multiple intersecting planned division lines (streets). By dividing this wafer along the planned division lines, multiple device chips, each equipped with a device, are obtained. The device chips are installed in a variety of electronic devices, such as mobile phones and personal computers.

In recent years, there has been a remarkable trend toward miniaturization of electronic devices, and the demand for thinner device chips is also increasing. Therefore, before dividing the wafer, it is ground from the back side using a grinding device to thin it to a specified finished thickness, and the thinned wafer is then divided. In this case, thin device chips are ultimately obtained.

A chamfered portion is formed on the outer periphery of a disc-shaped wafer, where the corners have been removed. As a result, when the wafer is thinned, a sharp knife-edge-like shape appears on the outer periphery, making the wafer susceptible to damage. Therefore, before grinding the wafer from the back side, an edge trimming process is performed in which the outer periphery of the wafer is cut from the front side to a depth greater than the finishing thickness, thereby partially removing the chamfered portion (see Patent Document 1).

Japanese Patent Application Laid-Open No. 2000-173961

However, when the wafer is thinned by grinding it from the backside after edge trimming, the scrap metal that makes up the annular terrace that protrudes outward, formed by edge trimming, breaks off just before the grinding wheel reaches it, causing relatively large scrap metal to fall. The fallen scrap metal accumulates in the drainage ditch used to drain the grinding water used in grinding, requiring the grinding machine to be stopped and the drainage ditch to be cleaned frequently.

The present invention was made in consideration of these problems, and aims to provide a wafer processing method that can prevent relatively large scrap pieces from falling when grinding edge-trimmed wafers.

According to one aspect of the present invention, there is provided a wafer processing method for thinning a disk-shaped wafer having a chamfered portion on its outer periphery by grinding it from its backside to a finishing thickness, the method comprising the steps of: a trimming step in which a first cutting blade is inserted into the chamfered portion from the front side of the wafer to a first depth position that is equal to or greater than the finishing thickness, thereby cutting the wafer along the outer periphery to form an annular terrace portion; a protective member adhering step in which a protective member is adhered to the front side of the wafer after the trimming step; and a protective member adhering step in which the wafer is ground from its backside to the finishing thickness after the protective member adhering step. and a grinding step of thinning the first cutting blade to a thickness of 1000 mm, and further comprising, before the grinding step, a first ultrasonic vibration cutting step of applying ultrasonic vibration to a second cutting blade having a blade thickness thinner than that of the first cutting blade, causing the second cutting blade to cut into the terrace portion from the front surface side of the wafer along the outer peripheral edge to a second depth position whose depth from the surface exceeds the first depth position, thereby forming a first annular cutting groove in the terrace portion (excluding the case where the first annular cutting groove is formed on the inner peripheral edge of the terrace portion) .

According to another aspect of the present invention, there is provided a wafer processing method for grinding a disk-shaped wafer having a chamfered portion on its outer periphery from its backside to thin it to a finishing thickness, the method comprising: a trimming step in which a first cutting blade is inserted into the chamfered portion from the front side of the wafer to a first depth position that is equal to or greater than the finishing thickness, thereby cutting the wafer along the outer periphery to form an annular terrace portion; a protective member adhering step in which a protective member is adhered to the front side of the wafer after the trimming step; and a grinding step in which the wafer is ground from its backside to thin it to the finishing thickness after the protective member adhering step, wherein before the grinding step, ultrasonic vibrations are applied to a second cutting blade having a blade thickness thinner than that of the first cutting blade. a first ultrasonic vibration cutting step of cutting the second cutting blade into the terrace portion from the front surface side of the wafer along the outer peripheral edge to a second depth position where the depth from the surface exceeds the first depth position, thereby forming a first annular cutting groove in the terrace portion; and a second ultrasonic vibration cutting step of cutting the second cutting blade into the terrace portion from the front surface side of the wafer along the outer peripheral edge to a third depth position where the depth from the surface exceeds the first depth position, while applying ultrasonic vibration to the second cutting blade, before performing the grinding step, thereby forming a second annular cutting groove in the terrace portion , wherein the first annular cutting groove and the second annular cutting groove are formed concentrically.

More preferably, the first annular cutting groove is formed more inward than the second annular cutting groove, and the second depth position is closer to the surface than the third depth position.

Also, preferably, the first ultrasonic vibration cutting step is performed after the trimming step.

In one aspect of the wafer processing method of the present invention, before grinding the wafer from the backside, the wafer is cut along the outer periphery with a first cutting blade to form an annular terrace portion, and separately, a first annular cutting groove is formed by cutting with a second cutting blade. When forming the first annular cutting groove, ultrasonic vibrations are applied to the second cutting blade, causing the wafer to break into fine particles near the first annular cutting groove, forming a fractured layer.

When a wafer with a terraced portion formed on its outer periphery and a first annular cutting groove is ground from the backside, the terraced portion is subdivided using the first annular cutting groove and the fractured layer as starting points for division, producing fine scraps. Since the fine scraps are easily washed away when they fall into the drain and are not likely to accumulate in the drain, there is no need to frequently clean the drain.

Therefore, one aspect of the present invention provides a wafer processing method that can prevent relatively large scrap pieces from falling when grinding edge-trimmed wafers.

FIG. 2 is a perspective view showing a wafer. FIG. 2 is a perspective view schematically showing a cutting device. FIG. 2 is a cross-sectional view schematically showing a wafer being cut by a first cutting blade. FIG. 10 is a cross-sectional view schematically showing a wafer being cut by a second cutting blade. FIG. 2 is an exploded perspective view schematically showing a second cutting unit. FIG. 4 is a cross-sectional view schematically showing a second cutting unit. 1 is an enlarged cross-sectional view schematically showing a portion of a wafer in which a plurality of annular cut grooves are formed in a terrace portion. FIG. FIG. 2 is a perspective view schematically showing a grinding device. FIG. 2 is a cross-sectional view schematically showing a wafer to which a protective member is attached. FIG. 2 is a cross-sectional view schematically showing a wafer being ground. FIG. 10 is a cross-sectional view schematically showing the wafer being further ground. 3 is a flowchart showing the flow of each step of a wafer processing method according to an embodiment. 10 is a photograph showing scraps recovered by carrying out a wafer processing method according to an embodiment of the present invention and scraps recovered by carrying out a wafer processing method according to a comparative example, side by side.

Embodiments of the present invention will now be described in detail with reference to the accompanying drawings. First, a wafer processed using the wafer processing method according to this embodiment will be described. Figure 1 is a perspective view schematically showing wafer 1.

The wafer 1 is a disk-shaped member made of a material such as silicon, and has a front surface 1a and a back surface 1b that are generally parallel to each other. A plurality of planned division lines 3 are set on the front surface 1a of the wafer 1, arranged in a grid pattern so that they intersect with each other. Devices 5, such as ICs and LSIs, are formed in each of the areas defined by the planned division lines 3 on the front surface 1a of the wafer 1. The area on the front surface 1a of the wafer 1 where the devices 5 are formed is called the device area 7. The outer area surrounding the device area 7 is called the peripheral excess area 9.

There are no restrictions on the material, structure, size, etc. of the wafer 1. For example, the wafer 1 may be a substrate made of a semiconductor other than silicon (GaAs, InP, GaN, SiC, etc.), sapphire, glass (quartz glass, borosilicate glass, etc.), etc. Furthermore, there are no restrictions on the type, number, shape, structure, size, arrangement, etc. of the devices 5, and the wafer 1 does not necessarily have to have any devices 5 formed on it.

If the wafer 1 has corners on its periphery, they are likely to chip or crack when the wafer 1 receives an impact at those corners. If these chips or cracks progress from the peripheral excess region 9 to the device region 7, the devices 5 will be damaged. Therefore, the outer periphery 1c of the wafer 1 is chamfered in advance to remove the corners and form a rounded chamfer. Figure 3 includes a cross-sectional view showing the chamfered portion of the wafer 1.

When wafer 1 is ground from the back surface 1b side to thin it, and then divided along planned division lines 3, multiple thin chips (device chips) each equipped with a device 5 are manufactured. To divide wafer 1, for example, a cutting device equipped with an annular cutting blade or a laser processing device equipped with a laser processing unit that irradiates wafer 1 with a laser beam is used. Device chips manufactured by these processing devices are installed in various electronic devices such as mobile phones and personal computers.

Next, we will explain the cutting device that performs edge trimming and the formation of annular cut grooves with fractured layers in the wafer processing method according to this embodiment. Figure 2 is a perspective view that schematically shows cutting device 2. However, edge trimming and the formation of annular cut grooves with fractured layers do not need to be performed using the same cutting device.

The wafer 1 to be cut by the cutting device 2 is brought into the cutting device 2, for example, integrated with an annular frame (not shown) having an opening with a diameter larger than that of the wafer 1, and tape attached to the annular frame to cover the opening. By attaching tape to the front side of the wafer 1 and integrating the wafer 1, tape, and annular frame to form a frame unit, the wafer 1 can be handled via the annular frame, making it easier to handle the wafer 1.

The cutting device 2 includes a base 4 that supports each component. An opening 4a is formed in the front corner of the base 4, and within this opening 4a is a cassette support table 8 that is raised and lowered by an elevation mechanism (not shown). Cassettes 10, each containing multiple wafers 1 that form part of a frame unit, are mounted on the top surface of the cassette support table 8. For ease of explanation, only the outline of the cassette 10 is shown in Figure 1.

A rectangular opening 4b is formed on the side of the cassette support base 8, with its longitudinal direction aligned with the X-axis direction (front-to-back direction, processing feed direction). A ball-screw type X-axis movement mechanism (not shown) is located within the opening 4b, along with a table cover 14 and a dust-proof/water-resistant cover 16 that cover the top of the X-axis movement mechanism. The X-axis movement mechanism is equipped with an X-axis movement table (not shown) covered by the table cover 14, and moves this X-axis movement table in the X-axis direction.

A holding table 18 is disposed on the top surface of the X-axis moving table so as to be exposed from the table cover 14. The holding table 18 has the function of suction-holding the wafer 1 placed on the holding surface 18a exposed upward. The holding table 18 is connected to a rotational drive source (not shown) such as a motor, and rotates around an axis of rotation roughly parallel to the Z-axis direction (vertical direction).

The holding table 18 comprises a porous member 18c with the same diameter as the wafer 1 and a frame that covers the porous member 18c. A suction path (not shown) is formed inside the holding table 18, one end of which is connected to a suction source (not shown) such as an ejector provided outside the holding table 18. The other end of the suction path reaches the porous member 18c.

The upper surface of the porous member 18c is exposed on the holding surface 18a of the holding table 18. The upper surface of the porous member 18c has the same diameter as the wafer 1 and is formed roughly parallel to the X-axis and Y-axis directions. Furthermore, multiple clamps 18b are provided around the periphery of the holding table 18 to secure an annular frame that supports the wafer 1.

When holding the wafer 1 on the holding table 18, first, the frame unit including the wafer 1 is placed on the holding surface 18a of the holding table 18. Then, the suction source is connected to the porous member 18c via the suction path, and negative pressure is applied to the wafer 1 via the tape attached to the back surface 1b of the wafer 1.

The cutting device 2 is equipped with a transport unit (not shown) in an area adjacent to the opening 4b that transports the wafer 1 to the holding table 18 or the like. A temporary placement mechanism for temporarily placing the wafer 1 is provided in a position adjacent to the side of the cassette support table 8. The temporary placement mechanism includes, for example, a pair of guide rails 12 that move toward and away from each other while maintaining a parallel state in the Y-axis direction (indexing feed direction). The pair of guide rails 12 sandwich the wafer 1 pulled out of the cassette 10 by the transport unit along the X-axis direction and align it to a predetermined position.

Once the wafer 1 is aligned to the specified position, it is lifted by the transport unit and transported to the holding table 18. At this time, the pair of guide rails 12 are moved apart, and the wafer 1 is passed between the pair of guide rails 12.

Above the holding table 18, there are provided a first cutting unit 24a and a second cutting unit 24b, which cut the wafer 1 using an annular cutting blade. A gate-shaped support structure 20 for supporting the first cutting unit 24a and the second cutting unit 24b is positioned on the top surface of the base 4, straddling the opening 4b.

A first moving unit 22a that moves the first cutting unit 24a in the Y-axis and Z-axis directions, and a second moving unit 22b that moves the second cutting unit 24b in the Y-axis and Z-axis directions are provided on the upper front surface of the support structure 20. The first moving unit 22a has a Y-axis moving plate 28a, and the second moving unit 22b has a Y-axis moving plate 28b. The two Y-axis moving plates 28a, 28b are slidably mounted on a pair of Y-axis guide rails 26 that are arranged along the Y-axis direction on the front surface of the support structure 20.

A nut portion (not shown) is provided on the back side (rear side) of the Y-axis moving plate 28a, and a Y-axis ball screw 30a that is generally parallel to the Y-axis guide rail 26 is threaded into this nut portion. Furthermore, a nut portion (not shown) is provided on the back side (rear side) of the Y-axis moving plate 28b, and a Y-axis ball screw 30b that is generally parallel to the Y-axis guide rail 26 is threaded into this nut portion.

A Y-axis pulse motor 32a is connected to one end of the Y-axis ball screw 30a. By rotating the Y-axis ball screw 30a with the Y-axis pulse motor 32a, the Y-axis moving plate 28a moves in the Y-axis direction along the Y-axis guide rail 26. Furthermore, a Y-axis pulse motor (not shown) is connected to one end of the Y-axis ball screw 30b. By rotating the Y-axis ball screw 30b with the Y-axis pulse motor, the Y-axis moving plate 28b moves in the Y-axis direction along the Y-axis guide rail 26.

A pair of Z-axis guide rails 34a are provided on the surface (front) side of Y-axis moving plate 28a along the Z-axis direction, and a pair of Z-axis guide rails 34b are provided on the surface (front) side of Y-axis moving plate 28b along the Z-axis direction. A Z-axis moving plate 36a is slidably attached to the pair of Z-axis guide rails 34a, and a Z-axis moving plate 36b is slidably attached to the pair of Z-axis guide rails 34b.

A nut (not shown) is provided on the back side (rear side) of the Z-axis moving plate 36a, and a Z-axis ball screw 38a, which is provided in a direction generally parallel to the Z-axis guide rail 34a, is threaded into this nut. A Z-axis pulse motor 40a is connected to one end of the Z-axis ball screw 38a, and by rotating the Z-axis ball screw 38a with this Z-axis pulse motor 40a, the Z-axis moving plate 36a moves in the Z-axis direction along the Z-axis guide rail 34a.

A nut (not shown) is provided on the back side (rear side) of the Z-axis moving plate 36b, and a Z-axis ball screw 38b, which is provided in a direction generally parallel to the Z-axis guide rail 34b, is threaded into this nut. A Z-axis pulse motor 40b is connected to one end of the Z-axis ball screw 38b, and by rotating the Z-axis ball screw 38b with this Z-axis pulse motor 40b, the Z-axis moving plate 36b moves in the Z-axis direction along the Z-axis guide rail 34b.

A first cutting unit 24a is provided below the Z-axis moving plate 36a. A camera unit 46a is provided adjacent to the first cutting unit 24a to photograph the wafer 1 held by suction on the holding table 18. Furthermore, a second cutting unit 24b is provided below the Z-axis moving plate 36b. A camera unit 46b is provided adjacent to the second cutting unit 24b to photograph the wafer 1 held by suction on the holding table 18.

The first moving unit 22a controls the positions of the first cutting unit 24a and camera unit 46a in the Y-axis and Z-axis directions, and the second moving unit 22b controls the positions of the second cutting unit 24b and camera unit 46b in the Y-axis and Z-axis directions. The positions of the first cutting unit 24a and the second cutting unit 24b are controlled independently.

Opening 4c is formed at a position opposite opening 4b from opening 4a. A cleaning unit 48 for cleaning wafers 1 is disposed within opening 4c, and wafers 1 cut on holding table 18 are cleaned by cleaning unit 48. Wafers 1 cleaned by cleaning unit 48 are stored back into cassette 10. Each cutting unit 24a and 24b will be described in more detail below.

Figure 3 is a cross-sectional view showing a wafer 1 being cut by the first cutting unit 24a. The first cutting unit 24a is used for trimming, which cuts the wafer 1 along the outer periphery 1c to remove part of the chamfer. Note that Figure 3 does not include the clamp 18b of the holding table 18, the camera unit 46a, the annular frame, tape, etc.

The first cutting unit 24a includes a spindle 50a extending in the Y-axis direction and a rotational drive source (not shown), such as a motor, connected to the base end of the spindle 50a. A circular first cutting blade 56a is fixed to the tip of the spindle 50a via a flange mechanism 52a.

The first cutting blade 56a is a so-called hub-type cutting blade that has an annular base 58a made of a material such as aluminum with a through-hole in the center, and a grinding wheel portion 60a fixed to the outer periphery of the base 58a. However, the first cutting blade 56a is not limited to being a hub-type.

A grinding wheel 60a is fixed to the outer periphery of the base 58a. The grinding wheel 60a contains countless abrasive grains and a bonding material (bond) that disperses and fixes the abrasive grains. For example, the abrasive grains may be made of materials such as diamond or cubic boron nitride (cBN), and the bonding material may be a nickel plating layer, a resin bond, a vitrified bond, or a metal bond. The thickness of the first cutting blade 56a used in edge trimming is determined, for example, according to the width of the terrace portion formed on the wafer 1. However, the method for determining the blade thickness is not limited to this. The blade thickness of the first cutting blade 56a is preferably, for example, 2 mm or more. However, the blade thickness is not limited to this.

The boss portion (not shown) of the flange mechanism 52a protruding in the Y-axis direction is inserted into the insertion hole of the base 58a, the first cutting blade 56a is brought into contact with the flange surface of the flange mechanism 52a, and the fixing nut 54a is tightened onto the tip of the boss portion. This secures the first cutting blade 56a to the tip of the spindle 50a. The first cutting unit 24a also includes a pair of cutting fluid supply nozzles 62a arranged to sandwich the lower portion of the first cutting blade 56a secured to the tip of the spindle 50a.

Activating the rotary drive source to rotate the spindle 50a rotates the first cutting blade 56a, and the rotating first cutting blade 56a cuts into the wafer 1 held by suction on the holding table 18, cutting the wafer 1. At this time, cutting fluid such as pure water is sprayed from the cutting fluid supply nozzle 62a onto the wafer 1 and the first cutting blade 56a, removing any processing debris and frictional heat generated by cutting.

Figure 4 is a cross-sectional view schematically showing a wafer 1 being cut by the second cutting unit 24b. Note that the clamp 18b of the holding table 18, the camera unit 46b, the annular frame, the tape, etc. are omitted from Figure 4. Also, Figure 5 is an exploded perspective view schematically showing the second cutting unit 24b, and Figure 6 is a cross-sectional view schematically showing the second cutting unit 24b. The second cutting unit 24b is used to cut the wafer 1 along its outer periphery to form an annular cutting groove in the outer periphery excess region 9.

The second cutting unit 24b includes a spindle 50b extending in the Y-axis direction and a housing 51b that rotatably accommodates the base end of the spindle 50b. A rotary drive source (not shown), such as a motor, connected to the base end of the spindle 50b is housed inside the housing 51b. A circular second cutting blade 56b is fixed to the tip of the spindle 50b via a flange mechanism 52b. The second cutting unit 24b includes a pair of cutting fluid supply nozzles 62b that are arranged to sandwich the lower part of the second cutting blade 56b fixed to the tip of the spindle 50b.

The second cutting blade 56b is a so-called washer-type cutting blade that consists of a grinding stone portion and does not have an annular base. However, the second cutting blade 56b is not limited to a washer-type. The grinding stone portion contains countless abrasive grains such as diamond and a binder that disperses and fixes the abrasive grains. For example, the abrasive grains may be made of diamond, cubic boron nitride, or the like, and the binder may be made of metal, ceramic, resin, or the like.

The blade thickness of the second cutting blade 56b used to form the annular cutting groove in the peripheral excess region 9 is preferably determined according to the width of the annular cutting groove to be formed, and is preferably set to, for example, 300 μm. However, the blade thickness of the second cutting blade 56b is not limited to this.

The second cutting unit 24b includes an ultrasonic vibration imparting unit that imparts ultrasonic vibrations to the second cutting blade 56b along its radially outward direction. The configuration of the second cutting unit 24b, which imparts ultrasonic vibrations while rotating the second cutting blade 56b, will now be described in further detail.

Figure 5 is an exploded perspective view showing the second cutting unit 24b. The second cutting blade 56b attached to the second cutting unit 24b has a first surface 57a and a second surface 57b that are generally parallel to each other, and a circular opening 57c is provided in the center of the second cutting blade 56b, penetrating the second cutting blade 56b in the thickness direction.

The tip (one end) of the spindle 50b is exposed from the housing 51b, and a threaded portion (external thread) 53 is formed on the outer surface of the tip of the spindle 50b. A rotational drive source (not shown), such as a motor, is connected to the base end (other end) of the spindle 50b. This rotational drive source causes the spindle 50b to rotate around a rotational axis that is roughly parallel to the Y-axis direction.

A flange mechanism 52b, to which the second cutting blade 56b is attached, is fixed to the tip of the spindle 50b. The flange mechanism 52b has a rear flange 64 that supports the second cutting blade 56b from the rear side, and a front flange 66 that supports the second cutting blade 56b from the front side.

The rear flange 64 comprises a disk-shaped flange portion 68 and a cylindrical support shaft (boss portion) 70 that protrudes from the center of the surface 68a of the flange portion 68. The flange portion 68 has an opening 64a that penetrates the center of the flange portion 68 and the center of the support shaft 70. In addition, a threaded portion (male thread) 70a is formed on the outer peripheral surface of the tip of the support shaft 70.

The rear flange 64 of the flange mechanism 52b is attached to the spindle 50b so that the tip of the spindle 50b is inserted into the opening 64a. In this state, the annular fixing nut 72 is screwed onto the threaded portion 53 formed on the tip of the spindle 50b and tightened. This secures the rear flange 64 of the flange mechanism 52b to the tip of the spindle 50b.

An annular front flange (pressing flange) 66 made of metal or the like is attached to the rear flange 64. The front flange 66 has a first surface (front surface) 66a and a second surface (back surface) 66b (see Figure 6) that are generally parallel to each other. A circular opening 66c is provided in the center of the front flange 66, penetrating the front flange 66 in the thickness direction.

When the support shaft 70 of the rear flange 64 is inserted sequentially into the opening 57c of the second cutting blade 56b and the opening 66c of the front flange 66, the second cutting blade 56b and the front flange 66 are attached to the rear flange 64. In this state, when the fixing nut 74 is tightened onto the threaded portion 70a formed on the support shaft 70, the second cutting blade 56b and the front flange 66 are fixed to the flange mechanism 52b. As a result, the second cutting blade 56b is clamped between the rear flange 64 and the front flange 66 and fixed to the tip of the spindle 50b.

When the second cutting blade 56b is attached to the tip of the spindle 50b and the spindle 50b is rotated by a rotary drive source connected to the spindle 50b, the second cutting blade 56b rotates at a predetermined rotation speed around an axis of rotation roughly parallel to the Y-axis direction.

Figure 6 is a cross-sectional view showing the second cutting unit 24b equipped with the second cutting blade 56b. An annular protrusion 68b protruding from the surface 68a is provided along the outer periphery of the flange portion 68 of the rear flange 64. Furthermore, a plurality of through-holes (slits) 68c penetrating the flange portion 68 in the thickness direction are provided in the area inside the protrusion 68b of the flange portion 68. For example, four arc-shaped through-holes 68c are formed in the flange portion 68 at approximately equal intervals along the circumferential direction of the flange portion 68 (see Figure 5).

On the other hand, an annular protrusion 66d protruding from the second surface 66b is provided along the outer periphery of the front flange 66. Furthermore, a plurality of through holes (slits) 66e are provided in the area inside the protrusion 66d of the front flange 66, penetrating the front flange 66 in the thickness direction. For example, four arc-shaped through holes 66e are formed in the front flange 66 at approximately equal intervals along the circumferential direction of the front flange 66 (see Figure 5).

An annular support member 76a that supports the first surface 57a of the second cutting blade 56b is provided on the tip surface of the convex portion 68b of the flange portion 68. An annular support member 76b that supports the second surface 57b of the second cutting blade 56b is provided on the tip surface of the convex portion 66d of the front flange 66. The support members 76a and 76b are made of a synthetic resin or the like, and come into contact with the second cutting blade 56b to clamp it. It is preferable that the support members 76a and 76b be made of a material with a hardness of 40 or higher as measured by a durometer type D.

A vibrator 78a is provided inside the convex portion 68b of the flange portion 68. Furthermore, a vibrator 78b is provided inside the convex portion 66d of the front flange 66. For example, the vibrators 78a and 78b are fixed to the rear flange 64 and the front flange 66, respectively, with adhesive. The vibrators 78a and 78b constitute a vibration imparting unit 78 that vibrates the second cutting blade 56b at a frequency within the ultrasonic band.

When the second cutting blade 56b and front flange 66 are attached to the rear flange 64, the second cutting blade 56b is clamped and fixed between the support members 76a and 76b. The second cutting blade 56b is then positioned so that its first surface 57a faces the vibrator 78a and its second surface 57b faces the vibrator 78b.

The vibrator 78a includes an annular piezoelectric element 80a arranged circumferentially around the flange portion 68 of the rear flange 64, a pair of electrodes 82a arranged to sandwich the piezoelectric element 80a, and an insulator 84a covering the piezoelectric element 80a and the pair of electrodes 82a. Similarly, the vibrator 78b includes an annular piezoelectric element 80b arranged circumferentially around the front flange 66, a pair of electrodes 82b arranged to sandwich the piezoelectric element 80b, and an insulator 84b covering the piezoelectric element 80b and the pair of electrodes 82b. For example, the piezoelectric elements 80a and 80b are made of piezoelectric ceramics such as barium titanate, lead zirconate titanate, or lithium tantalate.

Furthermore, wiring 86a and 86b are provided inside the flange mechanism 52b. One end of each of the wiring 86a and 86b branches into two and is exposed on the side of the support shaft 70. One electrode 82a of the vibrator 78a is connected to wiring 86a via lead wire 88a. The other electrode 82a of the vibrator 78a is connected to wiring 86b via lead wire 88b.

The front flange 66 has connection electrodes 90a and 90b exposed through the opening 66c. One electrode 82b of the vibrator 78b is connected to the wiring 86a via a lead wire 88c and the connection electrode 90a. The other electrode 82b of the vibrator 78b is connected to the wiring 86b via a lead wire 88d and the connection electrode 90b.

The second cutting unit 24b also includes a voltage supply unit 92 that supplies AC voltage to the vibrators 78a and 78b that make up the vibration imparting unit 78. The voltage supply unit 92 includes a power receiving unit (power receiving unit) 94 provided on the rear side (housing 51b side) of the rear flange 64, and a power supply unit (power supply unit) 100 located on the front side (flange mechanism 52b side) of the housing 51b and facing the power receiving unit 94.

The power receiving unit 94 includes an annular core 96 provided on the rear surface of the rear flange 64. An annular recess 96a is formed on the front surface (the power supply unit 100 side) of the core 96, and an annular coil (power receiving coil) 98 is provided inside this recess 96a. One end of the coil 98 is connected to the wiring 86a, and the other end of the coil 98 is connected to the wiring 86b.

The power supply unit 100 includes an annular core 102. For example, the core 102 is fixed to the surface side of the housing 51b with bolts (not shown). An annular recess 102a is formed on the surface side (the power receiving unit 94 side) of the core 102, and an annular coil (power supply coil) 104 is provided inside this recess 102a. The coil 104 is connected to an AC power source 108 via wiring 106a and 106b. In addition, a frequency converter 110 that controls the frequency of the AC voltage supplied from the AC power source 108 is connected to the AC power source 108.

The power receiving unit 94 and power supply unit 100 form a rotary transformer that transmits AC voltage supplied from the AC power supply 108 to the vibration imparting unit 78. When the wafer 1 is processed by the second cutting blade 56b, AC voltage is applied to the coil 104 of the power supply unit 100 by the AC power supply 108. At this time, the frequency of the AC voltage is controlled by the frequency converter 110.

The AC voltage applied to the coil 104 is supplied to the pair of electrodes 82a of the vibrator 78a and the pair of electrodes 82b of the vibrator 78b via the coil 98 of the power receiving unit 94, the wiring 86a, 86b, and the lead wires 88a, 88b, 88c, and 88d. This causes the piezoelectric element 80a of the vibrator 78a to vibrate radially together with the rear flange 64. Furthermore, the piezoelectric element 80b of the vibrator 78b to vibrate radially together with the front flange 66.

As a result, the second cutting blade 56b, which is sandwiched between the rear flange 64 and the front flange 66, vibrates in the radial direction of the second cutting blade 56b. The frequency of the AC voltage supplied from the AC power supply 108 is controlled by the frequency converter 110 so that the second cutting blade 56b vibrates at a frequency in the ultrasonic band.

Here, the flange portion 68 of the rear flange 64 has multiple through holes 68c, and the front flange 66 has multiple through holes 66e (see Figure 5). This reduces the radial rigidity of the flange portion 68 and the front flange 66, making the flange portion 68 and the front flange 66 more likely to vibrate in the radial direction. As a result, the second cutting blade 56b also more likely to vibrate in the radial direction.

Next, we will explain the grinding device used to thin the wafer 1 in the wafer processing method according to this embodiment. Figure 8 is a perspective view showing a schematic diagram of the grinding device 112. An opening 114a is provided on the upper surface of the base 114 of the grinding device 112. Inside the opening 114a, there is an X-axis moving table 118 on whose upper surface rests a holding table 116 that holds the wafer 1 by suction.

The X-axis moving table 118 can be moved in the X-axis direction by an X-axis moving mechanism (not shown). The X-axis moving table 118 is positioned between a loading/unloading area 120 where the wafer 1 is loaded and unloaded onto the holding table 116 by the X-axis moving mechanism, and a processing area 122 where the wafer 1 held by suction on the holding table 116 is ground.

A porous member having an upper surface with the same diameter as the wafer 1 is disposed on the upper surface of the holding table 116, and this upper surface of the porous member serves as the holding surface 116a that holds the wafer 1. The holding table 116 has an internal suction path (not shown) with one end connected to the porous member and the other end connected to a suction source (not shown). When the suction source is activated, negative pressure acts on the wafer 1 placed on the holding surface 116a, and the wafer 1 is held by suction to the holding table 116. The holding table 116 can also rotate around an axis perpendicular to the holding surface 116a.

A grinding unit 124 that grinds the wafer 1 is disposed above the processing area 122. A support part 126 is erected on the rear side of the base 114, and this support part 126 supports the grinding unit 124. A pair of Z-axis guide rails 128 extending in the Z-axis direction are provided on the front surface of the support part 126, and a Z-axis moving plate 130 is slidably attached to each Z-axis guide rail 128.

A nut portion (not shown) is provided on the back side (rear side) of the Z-axis moving plate 130, and a Z-axis ball screw 132 parallel to the Z-axis guide rail 128 is threadedly engaged with this nut portion. A Z-axis pulse motor 134 is connected to one end of the Z-axis ball screw 132. When the Z-axis pulse motor 134 rotates the Z-axis ball screw 132, the Z-axis moving plate 130 moves in the Z-axis direction along the Z-axis guide rail 128.

A grinding unit 124, which grinds the wafer 1, is fixed to the lower front side of the Z-axis moving plate 130. When the Z-axis moving plate 130 is moved in the Z-axis direction, the grinding unit 124 also moves in the Z-axis direction.

The grinding unit 124 has a spindle 138 aligned along the Z-axis, a housing 136 that rotatably accommodates the upper end of the spindle 138, and a disk-shaped wheel mount 140 fixed to the lower end of the spindle 138. The housing 136 accommodates a rotary drive source such as a motor that is connected to the upper end of the spindle 138 and rotates the spindle 138 around the Z-axis.

A circular grinding wheel 142 is fixed to the underside of the wheel mount 140. A grinding stone 144, consisting of abrasive grains such as diamond dispersed and fixed in a binder, is arranged in an annular shape on the underside of the grinding wheel 142.

When the spindle 138 is rotated to rotate the grinding wheel 142, the grinding stone 144 rotates on a circular orbit. The grinding unit 124 is then lowered to bring the rotating grinding stone 144 into contact with the surface of the wafer 1 to be ground, thereby grinding the wafer 1. The grinding device 112 has a thickness measuring device (not shown) that monitors the thickness of the wafer 1 while grinding, and when the thickness of the wafer 1 reaches the specified finished thickness, the lowering of the grinding unit 124 is stopped and grinding is completed.

When the wafer 1 is ground with the grinding wheel 144, processing debris and frictional heat are generated from the wafer 1 and the grinding wheel 144. The grinding device 112 is equipped with a grinding water supply nozzle 146 (see Figure 10, etc.), and while the wafer 1 is being ground with the grinding wheel 144, grinding water 148, such as pure water, is supplied to the wafer 1, etc. from the grinding water supply nozzle 146. The processing debris and frictional heat are removed by the grinding water 148.

The grinding device 112 has a drain groove 114b in the opening 114a through which used grinding water flows. The grinding water 148 absorbs processing debris and frictional heat before flowing into the drain groove 114b. A drain outlet 114c is formed at the bottom of the drain groove 114b, and the used grinding water that flows into the drain groove 114b is discharged from the drain outlet 114c.

Because a chamfer is formed on the outer periphery 1c of the wafer 1, thinning the wafer 1 to the finishing thickness in this state leaves the chamfer partially intact, creating a knife-edge shape and making the wafer 1 susceptible to damage. Therefore, before grinding the wafer 1 from the back surface 1b side, an edge trimming process is performed in which the outer periphery of the wafer 1 is cut from the front surface 1a side to a depth greater than the finishing thickness, thereby partially removing the chamfer.

However, when the wafer 1 is subsequently ground and thinned from the back surface 1b side using the grinding device 112, scrap material remaining on the outer periphery of the wafer 1 breaks off just before the grinding wheel 144 reaches the terrace portion 13 (see Figure 7, etc.) formed by the edge trimming process, causing relatively large scrap material to fall. Because the large scrap material that falls accumulates in the drain groove 114b, which is used to drain the grinding water, or clogs the drain outlet 114c, it has traditionally been necessary to frequently clean the drain groove 114b of the grinding device 112.

Therefore, in the wafer processing method according to this embodiment, before grinding the wafer 1 from the back surface 1b side, an annular cut groove with a fractured layer is formed near the terrace portion 13 formed by the edge trimming process. When the wafer 1 is ground from the back surface 1b side in this state, the annular cut groove and fractured layer act as starting points for dividing the terrace portion, resulting in the generation of fine scraps. When the fine scraps fall into the drain groove 114b, they are easily washed away and do not accumulate, making them less likely to clog the drain outlet 114c. Therefore, there is no need to frequently clean the drain groove 114b.

The wafer processing method according to this embodiment will be described below. Figure 12 is a flowchart showing the flow of each step of the wafer processing method according to this embodiment, in which a disk-shaped wafer 1 having a chamfered portion on its outer periphery 1c is ground from the back surface 1b side to thin it to its finish thickness.

In the wafer processing method according to this embodiment, first, a trimming step S10 is performed in which the wafer 1 is cut along its outer periphery to form an annular terrace on the chamfered portion. The trimming step S10 is performed using the cutting device 2 shown in Figure 2.

First, the cassette 10 containing the frame unit, which is an integrated combination of the wafer 1, tape, and annular frame, is carried to the cassette support table 8. The wafer 1 (frame unit) is then pulled out of the cassette 10 and placed on the holding surface 18a of the holding table 18, and the wafer 1 is held by suction on the holding table 18 via the tape. At this time, the front surface 1a of the wafer 1 is exposed upward.

Edge trimming of the wafer 1 is performed by the first cutting unit 24a. Figure 3 is a cross-sectional view showing a schematic diagram of the wafer 1 undergoing edge trimming. First, the camera unit 46a photographs the wafer 1 to detect the position of the outer periphery 1c of the wafer 1. Then, the positions of the first cutting unit 24a and the holding table 18 are adjusted so that the grinding wheel portion 60a of the first cutting blade 56a is positioned above the outer periphery 1c of the wafer 1.

Next, the spindle 50a is rotated to rotate the first cutting blade 56a, and the first cutting unit 24a is lowered until the bottom end of the grinding wheel portion 60a reaches a first depth position 13a (see Figure 7) that is greater than or equal to the finishing thickness from the surface 1a of the wafer 1. For example, if the thickness of the wafer 1 is 700 μm, this first depth position 13a is preferably approximately 300 μm below the surface 1a of the wafer 1.

With the first cutting blade 56a cutting into the chamfered portion from the front surface 1a of the wafer 1, the holding table 18 is rotated 360° or more around a table rotation axis perpendicular to the holding surface 18a. The wafer 1 is then cut along its outer periphery, partially removing the chamfered portion and forming an annular terrace portion 13 (see Figure 4, etc.) on the wafer 1.

Next, in the cutting device 2, a first ultrasonic vibration cutting step S21 is performed in which the second cutting unit 24b cuts the wafer 1 to form a first annular cutting groove. Figure 4 is a cross-sectional view showing a schematic diagram of the first ultrasonic vibration cutting step S21. In the first ultrasonic vibration cutting step S21, a second cutting blade 56b, which is thinner than the first cutting blade 56a, is inserted from the front surface 1a side into the position where the terrace portion 13 of the wafer 1 will be formed.

In the first ultrasonic vibration cutting step S21, first, the positions of the second cutting unit 24b and the holding table 18 are adjusted to position the second cutting blade 56b above a predetermined position on the terrace portion 13. Then, while rotating the second cutting blade 56b, the vibration imparting unit 78 (see Figure 6) vibrates the second cutting blade 56b in the radial direction of the second cutting blade 56b.

At this time, the vibration imparting unit 78 vibrates the second cutting blade 56b at a frequency in the ultrasonic band. For example, the vibration frequency of the second cutting blade 56b is set to 20 kHz or higher. When vibration is imparted to the second cutting blade 56b by the vibration imparting unit 78, the second cutting blade 56b vibrates so that its diameter increases or decreases. The amount of variation in the diameter of the second cutting blade 56b (the difference between the maximum and minimum diameter values) is, for example, approximately 5 μm.

In this state, the second cutting unit 24b is lowered until the lower end of the second cutting blade 56b reaches a second depth position 17a (see Figure 7), where the depth from the surface 1a of the wafer 1 exceeds the first depth position 13a. For example, this second depth position 17a is preferably approximately 500 μm below the surface 1a of the wafer 1.

While the second cutting blade 56b is cutting into the terrace portion 13 from the front surface 1a of the wafer 1, the holding table 18 is rotated 360° or more around the table rotation axis perpendicular to the holding surface 18a. This cuts the wafer 1 along its outer periphery, forming a first annular cutting groove 15a in the annular terrace portion 13.

At this time, the vibrations of the second cutting blade 56b at a frequency within the ultrasonic band (ultrasonic vibrations) cause the abrasive grains exposed from the lower end of the second cutting blade 56b to collide with the wafer 1 at a frequency corresponding to the ultrasonic vibrations. This causes the wafer 1 to fracture, forming a fractured layer (not shown) around the first annular cutting groove 15a.

In the wafer processing method according to this embodiment, in addition to the first annular cutting groove 15a, it is preferable to form an additional annular cutting groove with the second cutting blade 56b in the terrace portion 13 formed on the outer periphery 1c of the wafer 1. Next, the second ultrasonic vibration cutting step S22 for forming the second annular cutting groove 15b (see Figure 7) will be described.

In the second ultrasonic vibration cutting step S22, first, the position of the second cutting unit 24b in the Y direction is adjusted to position the second cutting blade 56b above a predetermined position radially outward of the wafer 1 and beyond the first annular cutting groove 15a formed in the terrace portion 13. Then, as in the first ultrasonic vibration cutting step S21, the second cutting blade 56b is rotated while the vibration imparting unit 78 (see Figure 6) vibrates the second cutting blade 56b in the radial direction of the second cutting blade 56b.

In this state, the second cutting unit 24b is lowered until the bottom end of the second cutting blade 56b reaches a third depth position 17b (see Figure 7), which is a depth from the surface 1a of the wafer 1 that exceeds the second depth position 17a. In other words, the second depth position 17a is closer to the surface 1a than the third depth position 17b. For example, it is preferable that the third depth position 17b be approximately 550 μm below the surface 1a of the wafer 1.

With the second cutting blade 56b cutting into the terrace portion 13 from the front surface 1a of the wafer 1, the holding table 18 is rotated 360° or more around a table rotation axis perpendicular to the holding surface 18a. The wafer 1 is then cut along its outer periphery, forming a second annular cutting groove 15b in the annular terrace portion 13. At this time, the wafer 1 also fractures around the second annular cutting groove 15b, forming a fracture layer (not shown) around the first annular cutting groove 15a.

Further, annular cutting grooves may be formed in the terrace portion 13 of the wafer 1. Figure 7 is an enlarged cross-sectional view schematically showing the vicinity of the outer periphery 1c of the wafer 1 in which a first annular cutting groove 15a, a second annular cutting groove 15b, and a third annular cutting groove 15c are formed in the terrace portion 13.

Here, as an example, the third annular cut groove 15c is formed so that its lower end reaches a fourth depth position 17c that is deeper from the surface 1a of the wafer 1 than the third depth position 17b at which the lower end of the second annular cut groove 15b is located. For example, the fourth depth position 17c is preferably located approximately 600 μm below the surface 1a of the wafer 1. However, the depth position of the lower end of the third annular cut groove 15c is not limited to this.

Since each annular cutting groove is formed by rotating the holding table 18 at the same position, the first annular cutting groove 15a, the second annular cutting groove 15b, and the third annular cutting groove 15c are formed concentrically around the rotation axis of the holding table 18. When multiple annular cutting grooves 15a, 15b, and 15c are formed in this manner on the terrace portion 13 formed on the outer periphery 1c of the wafer 1, it is preferable that the outermost annular cutting grooves have their lower ends reach deeper.

Up to this point, we have explained an example in which the first cutting blade 56a used in edge trimming is a hub-type cutting blade, and the second cutting blade 56b used in cutting with ultrasonic vibration is a washer-type cutting blade. However, the first cutting blade 56a and the second cutting blade 56b are not limited to this.

That is, a washer-type first cutting blade 56a may be used in the trimming step S10. Furthermore, a hub-type second cutting blade 56b may be used in the first ultrasonic vibration cutting step S21 and the second ultrasonic vibration cutting step S22. If the second cutting blade 56b is a hub-type, a vibration imparting unit is incorporated into the base supporting the grinding wheel, for example.

After the trimming step S10, the first ultrasonic vibration cutting step S21, and the second ultrasonic vibration cutting step S22, a protective member attachment step S30 is performed in which a protective member is attached to the front surface 1a of the wafer 1 in preparation for grinding the wafer 1. Figure 9 is a cross-sectional view schematically showing the wafer 1 placed on the holding table 116 of the grinding device 112 with the protective member 17 attached to the front surface 1a and the back surface 1b facing upward.

The protective member 17 is a disk-shaped member with the same diameter as the wafer 1, and is made of a material such as resin. When the protective member 17 is attached to the front surface 1a of the wafer 1, the front surface 1a does not come into direct contact with the holding table 116 when the wafer 1 is ground from the back surface 1b side. As a result, the front surface 1a of the wafer 1 is not damaged. The wafer 1 with the protective member 17 attached to its front surface 1a side is placed on the holding table 116 and held by suction to the holding table 116.

After the protective member attachment step S30, the grinding step S40 is performed, in which the wafer 1 is ground from the back surface 1b to thin it down to its finish thickness. Figure 10 is a cross-sectional view showing a schematic representation of the initial stage of grinding step S40, and Figure 11 is a cross-sectional view showing a schematic representation of the final stage of grinding step S40.

When grinding the wafer 1, the holding table 116 that holds the wafer 1 by suction is moved to the processing area 122 (see Figure 8), and rotation of the spindle 138 and holding table 116 begins. Then, while supplying grinding water 148 to the wafer 1 from the grinding water supply nozzle 146, the grinding unit 124 is lowered, and the grinding wheel 144 moving on the circular orbit is brought into contact with the back surface 1b of the wafer 1, thereby commencing grinding of the wafer 1.

When the wafer 1 is ground from the back surface 1b side, the thickness of the wafer 1 gradually decreases, and the deepest third annular cutting groove 15c is first exposed on the back surface 1b side. At this time, the wafer 1 is divided by the third annular cutting groove 15c, generating scrap material, which is then crushed starting from the fracture layer formed around the third annular cutting groove 15c.

As grinding continues, the next deepest second annular cut groove 15b is exposed on the back surface 1b. At this point, the wafer 1 is divided at the second annular cut groove 15b, producing scrap material, which is then crushed starting from the fracture layer formed around the second annular cut groove 15b. As grinding continues, the next deepest first annular cut groove 15a is exposed on the back surface 1b. At this point, the wafer 1 is divided at the first annular cut groove 15a, producing scrap material, which is then crushed starting from the fracture layer formed around the first annular cut groove 15a.

Then, as shown in FIG. 11, the grinding unit 124 is lowered until the wafer 1 reaches the finishing thickness, completing the grinding of the wafer 1. The grinding unit 124 is then raised, the holding table 116 is returned to the loading/unloading area 120, the suction holding of the wafer 1 by the holding table 116 is released, and the wafer 1 is unloaded from the grinding device 112.

In this way, in the wafer processing method according to this embodiment, when the wafer 1 is thinned and the grinding wheel 144 reaches the terrace portion 13, the annular cutting groove and crushed layer act as the starting point for division, and the vicinity of the outer periphery 1c of the wafer 1 is subdivided. Since the subdivided scraps are easily washed away when they fall into the drainage groove 114b and are not likely to accumulate in the drainage groove 114b, there is no need to clean the drainage groove 114b frequently.

If the depth relationship between the annular cutting grooves 15a, 15b, and 15c were reversed, the annular cutting groove exposed on the back surface 1b would first cut away the portion outside the annular cutting groove. However, because the portion cut away is before it has been subdivided, the resulting scrap material would be relatively large. For this reason, it is preferable that the depth relationship is not reversed. However, even if the depth relationship is reversed, the resulting scrap material will be smaller than when no annular cutting grooves are formed, making it less likely to accumulate in the drain groove 114b.

Next, we will explain an example in which edge trimming was performed on a disk-shaped wafer 1 having a chamfered portion on the outer periphery 1c, two annular cutting grooves 15a, 15b were formed in the formed terrace portion 13, and the wafer 1 was ground from the back surface 1b side, resulting in the recovery and observation of the scrap material. In the wafer processing method of this example, the two annular cutting grooves 15a, 15b were formed using a second cutting blade 56b to which ultrasonic vibrations were applied.

First, multiple silicon wafers with a thickness of 700 μm and a diameter of 8 inches were prepared as wafers 1. No devices 5 were formed on the surface 1a of these wafers 1. These wafers 1 were loaded into a cutting device 2 and processed one by one using the procedure below. The cutting device 2 used for edge trimming of wafers 1 was a "DFD6361" manufactured by Disco Corporation. First, wafer 1 was loaded onto holding table 18 and held by suction on holding table 18.

Then, a washer-type first cutting blade 56a manufactured by Disco Corporation with a blade thickness of 2 mm or more was used to cut along the outer periphery 1c of the wafer 1. At this time, an area 2 mm wide from the outer periphery 1c of the wafer 1 and 300 μm deep from the surface 1a was removed, forming a terrace portion 13 along the outer periphery 1c on the wafer 1. The rotation speed of the first cutting blade 56a was set to 30,000 rpm, the rotation speed of the holding table 18 was set to 1 degree per second, and the holding table 18 was rotated 370 degrees.

Next, the wafer 1 was loaded into a Disco Corporation ultrasonic vibration cutting machine, DAD3350. A hub-type second cutting blade 56b, Disco Corporation's U09RA-SDC600-BB200-50, with a blade thickness of 300 μm, was used to cut the terrace portion 13 along the outer periphery 1c, forming a first annular cutting groove 15a in the wafer 1. The first annular cutting groove 15a was formed 500 μm from the outer periphery 1c of the wafer 1, and the second cutting blade 56b was inserted to a height position 550 μm lower than the surface 1a of the wafer 1.

The rotation speed of the second cutting blade 56b was set to 30,000 revolutions per minute, the rotation speed of the holding table 18 was set to 1 degree per second, and the holding table 18 was rotated 370 degrees. During this time, ultrasonic vibrations with a radial amplitude of 5 μm and a frequency of 41 kHz were applied to the second cutting blade 56b.

Furthermore, the second cutting blade 56b was used to cut the terrace portion 13 along the outer periphery 1c, forming a second annular cutting groove 15b in the wafer 1. At this time, the second annular cutting groove 15b was formed radially inward of the first annular cutting groove 15a so that the distance from the first annular cutting groove 15a was 500 μm. At this time, the second cutting blade 56b was cut to a height position 500 μm lower than the surface 1a of the wafer 1.

The rotation speed of the second cutting blade 56b was set to 30,000 revolutions per minute, the rotation speed of the holding table 18 was set to 1 degree per second, and the holding table 18 was rotated 370 degrees. Again, ultrasonic vibrations with a radial amplitude of 5 μm and a frequency of 41 kHz were applied to the second cutting blade 56b.

Next, the wafer 1, on which the terrace portion 13, first annular cutting groove 15a, and second annular cutting groove 15b were formed, was loaded into a grinding device 112 such as that shown in Figure 8, and ground using the following procedure. In the wafer processing method according to this example, the attachment of a protective member to the front surface 1a of the wafer 1 was omitted. The grinding device 112 used to grind the wafer 1 was a "DAG810" manufactured by Disco Corporation. First, the wafer 1 was loaded onto the holding table 116 with the back surface 1b facing upward, and was held by suction on the holding table 116.

Then, the wafer 1 was ground from the backside 1b side using a grinding wheel 142 manufactured by Disco Corporation equipped with a grinding stone 144. The rotation speed of the holding table 116 was set to 300 rpm, the rotation speed of the spindle 138 was set to 3,500 rpm, and the lowering speed of the grinding unit 124 was set to 0.7 μm per second. The finished thickness of the wafer 1 was set to 300 μm, and the grinding unit 124 was lowered until the thickness of the wafer 1 reached the finished thickness.

When wafer 1 is ground by grinding device 112, scrap material is generated from wafer 1 and falls into drainage groove 114b. In this example, after grinding of wafer 1 was completed, the scrap material that fell into drainage groove 114b was collected, observed, and its size was measured.

For comparison, two annular cutting grooves 15a, 15b were formed with the second cutting blade 56b on a terrace portion 13 formed in a similar manner on another wafer 1 by edge trimming without applying ultrasonic vibrations, and the scrap material generated when the wafer 1 was ground from the back surface 1b side was observed. The only difference between the wafer processing method of the example and the wafer processing method of the comparative example is whether or not ultrasonic vibrations were applied to the second cutting blade 56b when forming the two annular cutting grooves 15a, 15b.

Figure 13 is a photograph of the collected scraps placed on a mesh fabric. The left side of the photograph shows three of the scraps 19a collected when the wafer processing method of the example was carried out. The right side of the photograph shows one of the scraps 19b collected when the wafer processing method of the comparative example was carried out. As is clear from the photograph shown in Figure 13, scrap 19a is significantly smaller than scrap 19b.

More specifically, when the sizes of the recovered scraps 19a and 19b were measured, it was confirmed that the average width of scrap 19a was approximately 75% smaller than scrap 19b. This is thought to be because when ultrasonic vibrations were applied to the second cutting blade 56b to form the annular cutting grooves 15a and 15b, a fractured layer was formed at the bottom of the annular cutting grooves 15a and 15b. It is thought that when the wafer 1 was ground from the back surface 1b side, this fractured layer became the starting point for splitting, causing the wafer 1 to be split into smaller pieces near the terrace portion 13.

Furthermore, as is clear from the photograph shown in Figure 13, scrap 19a is smaller than scrap 19b not only in width but also in length. This suggests that the fractured layer formed on wafer 1 promotes fragmentation of wafer 1 in the length direction as well. The examples and comparative examples described above confirmed that by forming annular cutting grooves 15a, 15b with second cutting blade 56b to which ultrasonic vibrations are applied, small fragmented scraps are ultimately produced when wafer 1 is ground.

The present invention is not limited to the above-described embodiment and can be implemented with various modifications. For example, the above embodiment describes the case where three annular cutting grooves 15a, 15b, and 15c are formed in the terrace portion 13 of the wafer 1, but the wafer processing method according to one aspect of the present invention is not limited to this. Only the first annular cutting groove 15a may be formed in the terrace portion 13, and the second ultrasonic vibration cutting step S22 may not be performed. Even in this case, the amount of scrap generated when grinding the wafer 1 is smaller than in conventional methods.

In addition, in the above embodiment, the first ultrasonic vibration cutting step S21 is performed after the trimming step S10, but this aspect of the present invention is not limited to this. That is, the first ultrasonic vibration cutting step S21, etc. may be performed first to form one or more annular cutting grooves 15a, 15b, 15c near the outer periphery 1c of the wafer 1, and then the trimming step S10 may be performed to form the terrace portion 13.

Even in this case, annular cutting grooves 15a, 15b, and 15c are formed in the terrace portion 13 before grinding step S40 is performed, thereby reducing the amount of scrap material generated in grinding step S40. Before the terrace portion 13 is formed, the outer periphery 1c of the wafer 1 can be easily detected using the camera unit 46b located near the second cutting unit 24b.

However, if the annular cutting grooves 15a, 15b, and 15c are to be formed when the terrace portion 13 has not yet been formed, the second cutting blade 56b must be cut deep into the wafer 1 from the surface 1a, which requires a second cutting blade 56b with a long cutting edge.

The structures, methods, etc., described in the above embodiments can be modified as appropriate without departing from the scope of the present invention.

1 Wafer 1a Front surface 1b Back surface 1c Outer periphery 3 Planned division line 5 Device 7 Device area 9 Outer surplus area 13 Terrace portion 13a, 17a, 17b, 17c Depth position 15a, 15b, 15c Annular cutting groove 17 Protective member 19a, 19b Scrap material 2 Cutting device 4 Base 4a, 4b, 4c Opening 8 Cassette support base 10 Cassette 12 Guide rail 14 Table cover 16 Dustproof/waterproof cover 18 Holding table 18a Holding surface 18b Clamp 18c Porous member 20 Support structure 22a, 22b Moving unit 24a, 24b Cutting unit 26, 34a, 34b Guide rail 28a, 28b, 36a, 36b Moving plate 30a, 30b, 38a, 38b Ball screw 32a, 40a, 40b Pulse motor 46a, 46b Camera unit 48 Cleaning unit 50a, 50b Spindle 51b Housing 52a, 52b Flange mechanism 53 Threaded portion 54a Fixing nut 56a, 56b Cutting blade 57a First surface 57b Second surface 57c Opening 58a Base 60a Grinding stone portion 62a, 62b Cutting fluid supply nozzle 64 Rear flange 64a Opening 66 Front flange 66a First surface 66b Second surface 66c Opening 66d Convex portion 66e Through hole 68 Flange portion 68a Surface 68b Convex portion 68c Through hole 70 Support shaft 70a Threaded portion 72, 74 Fixing nuts 76a, 76b Support member 78 Vibration imparting unit 78a, 78b Vibrator 80a, 80b Piezoelectric body 82a, 82b Electrode 84a, 84b Insulator 86a, 86b Wiring 88a, 88b, 88c, 88d Lead wire 90a, 90b Connection electrode 92 Voltage supply unit 94 Power receiving portion 96, 102 Core 96a, 102a Recess 98, 104 Coil 100 Power supply portion 106a, 106b Wiring 108 AC power supply 110 Frequency converter 112 Grinding device 114 Base 114a Opening 114b Drain groove 114c Drain port 116 Holding table 116a Holding surface 118 X-axis moving table 120 Loading/unloading area 122 Processing area 124 Grinding unit 126 Support 128 Z-axis guide rail 130 Z-axis moving plate 132 Z-axis ball screw 134 Z-axis pulse motor 136 Housing 138 Spindle 140 Wheel mount 142 Grinding wheel 144 Grinding stone 146 Grinding water supply nozzle 148 Grinding water

Claims (4)

A wafer processing method for thinning a disk-shaped wafer having a chamfered outer periphery to a finish thickness by grinding the wafer from its back surface, comprising the steps of:
a trimming step in which a first cutting blade is inserted into the chamfered portion from the front surface side of the wafer to a first depth position that is equal to or greater than the finishing thickness, thereby cutting the wafer along the outer periphery, thereby forming an annular terrace portion;
a protective member attaching step of attaching a protective member to the front surface side of the wafer after the trimming step;
a grinding step of grinding the wafer from the back surface to thin it to the finished thickness after the protective member attaching step,
The wafer processing method further comprises a first ultrasonic vibration cutting step, before carrying out the grinding step, in which ultrasonic vibrations are applied to a second cutting blade having a blade thickness thinner than that of the first cutting blade, and the second cutting blade is caused to cut into the terrace portion from the surface side of the wafer along the outer peripheral edge to a second depth position whose depth from the surface exceeds the first depth position , thereby forming a first annular cutting groove in the terrace portion (except when the first annular cutting groove is formed on the inner peripheral edge of the terrace portion) . A wafer processing method for thinning a disk-shaped wafer having a chamfered outer periphery to a finish thickness by grinding the wafer from its back surface, comprising the steps of:
a trimming step in which a first cutting blade is inserted into the chamfered portion from the front surface side of the wafer to a first depth position that is equal to or greater than the finishing thickness, thereby cutting the wafer along the outer periphery, thereby forming an annular terrace portion;
a protective member attaching step of attaching a protective member to the front surface side of the wafer after the trimming step;
a grinding step of grinding the wafer from the back surface to thin it to the finished thickness after the protective member attaching step,
a first ultrasonic vibration cutting step of, before the grinding step, applying ultrasonic vibration to a second cutting blade having a blade thickness thinner than that of the first cutting blade, cutting the second cutting blade into the terrace portion from the front surface side of the wafer along the outer peripheral edge to a second depth position whose depth from the surface exceeds the first depth position, thereby forming a first annular cutting groove in the terrace portion;
a second ultrasonic vibration cutting step of, before the grinding step, cutting the second cutting blade into the terrace portion from the front surface side of the wafer along the outer circumferential edge to a third depth position whose depth from the front surface exceeds the first depth position while applying ultrasonic vibration to the second cutting blade, thereby forming a second annular cutting groove in the terrace portion ;
A wafer processing method, wherein the first annular cutting groove and the second annular cutting groove are formed concentrically. The first annular cutting groove is formed more inward than the second annular cutting groove,
3. The wafer processing method according to claim 2, wherein the second depth position is closer to the surface than the third depth position. A wafer processing method according to any one of claims 1 to 3, characterized in that the first ultrasonic vibration cutting step is performed after the trimming step.