Stacked wafer processing method

Abstract

To completely remove a first wafer over the entire outer periphery of a stacked wafers when edge trimming is performed on the first wafer of the stacked wafers, but a bonding layer is left such that the outer periphery of the second wafer is not exposed.SOLUTION: A stacked wafer processing method includes a first cutting step for forming a first step portion at the outer periphery of a first wafer, the first step portion having a depth that does not reach a second wafer but leaves the first wafer, and a second cutting step for removing the outer periphery of the stacked wafers by rotating the holding table around the rotation axis while positioning the lower end of a second cutting blade at a position where a bonding layer is present, and in the second cutting step, the second cutting blade and the holding table are rotated such that the direction of the velocity vector of the lower end of the second cutting blade is opposite to the direction of the velocity vector of the stacked wafers at the position corresponding to the lower end of the second cutting blade.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method for processing a laminated wafer in which a first wafer and a second wafer are laminated with a bonding layer sandwiched therebetween. [Background technology]

[0002] In recent years, there has been an increasing number of cases where laminated wafers are handled as workpieces to be processed, in which a first wafer and a second wafer, each having a chamfered portion (also called a bevel portion) on the outer periphery of both sides, are bonded together via a bonding layer.

[0003] It is known that before thinning the first wafer of the stacked wafers by grinding, edge trimming is performed on the outer periphery of the first wafer using a cutting blade to remove the chamfered portions on both sides of the first wafer over the entire circumferential direction (see, for example, Patent Document 1).

[0004] When edge trimming is performed on the outer periphery of a first wafer, if the cutting blade is positioned at a height corresponding to the boundary between the first wafer and the bonding layer to cut only the first wafer without cutting the second wafer, the area of ​​the annular region on the first wafer that has been edge trimmed may not be completely removed and may remain partially.

[0005] If voids due to poor bonding occur between the remaining region of the first wafer and the bonding layer, defects will occur in the stacked wafers when the edge-trimmed stacked wafers are subjected to heat treatment in the subsequent process. Therefore, it is desirable to completely remove the first wafer in the annular region where edge trimming has been performed.

[0006] Therefore, in order to eliminate the remaining area of ​​the first wafer in the annular region where edge trimming has been performed, it is possible to lower the lower end of the cutting blade to the position of the second wafer, thereby cutting not only the first wafer and the bonding layer, but also the area of ​​the outer periphery of the second wafer near the bonding layer.

[0007] However, if the outer periphery of the second wafer is exposed by cutting in this manner, the outer periphery of the second wafer will also be etched in the subsequent etching process (specifically, the etching process for removing grinding damage that occurs on the grinding surface of the first wafer when grinding the first wafer).

[0008] If a device is formed on one side of the second wafer near the bonding layer, etching the outer periphery of the second wafer can cause damage to the device located near the outer periphery of one side of the second wafer.

[0009] Therefore, in edge trimming, it is preferable to leave the bonding layer over the entire outer periphery of one side of the second wafer without cutting the second wafer so that the outer periphery of one side of the second wafer is not exposed.

[0010] However, in edge trimming, the cutting depth can vary to some extent in the circumferential direction of the stacked wafers due to the nature of cutting using a cutting blade, which makes it difficult to control the process so that the first wafer is completely removed from the entire outer periphery of the stacked wafers while leaving the bonding layer so that the outer periphery of the second wafer is not exposed. [Prior art documents] [Patent documents]

[0011] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-92413 Summary of the Invention [Problem to be solved by the invention]

[0012] The present invention has been made in consideration of the above problems, and aims to completely remove the first wafer over the entire outer periphery of the stacked wafers when edge trimming is performed on the first wafer of the stacked wafers, while leaving the bonding layer so that the outer periphery of the second wafer is not exposed. [Means for solving the problem]

[0013] According to one aspect of the present invention, a method for processing laminated wafers in which a first wafer and a second wafer are stacked with a bonding layer sandwiched therebetween includes: a holding step of holding the laminated wafers on a holding table rotatable around a predetermined rotation axis so that the first wafer is exposed; a first cutting step of, after the holding step, rotating a first cutting blade attached to a tip of a first spindle around the first spindle and rotating the holding table around the rotation axis with the first cutting blade cutting into the outer periphery of the first wafer, thereby forming a first step portion in the outer periphery of the first wafer, the first step portion having a depth that does not reach the second wafer but leaves the first wafer; and a second cutting step of removing the outer periphery of the laminated wafers by rotating a second cutting blade attached to the tip of a second spindle around the second spindle after the first cutting step and positioning the lower end of the second cutting blade closer to the second wafer in the thickness direction of the laminated wafers than the bottom surface of the first step portion and at a position where the bonding layer is present, and rotating the holding table around the rotation axis, wherein the second cutting step rotates the second cutting blade and the holding table so that the direction of the velocity vector of the lower end of the second cutting blade is opposite to the direction of the velocity vector of the laminated wafers at a position corresponding to the lower end.

[0014] Preferably, the method for processing stacked wafers further includes a remaining thickness measurement step, after the first cutting step and before the second cutting step, of measuring the remaining thickness of the first wafer remaining between the bottom surface of the first step portion and the second wafer, and in the second cutting step, the position of the lower end of the second cutting blade is set based on the remaining thickness obtained in the remaining thickness measurement step.

[0015] Preferably, in the remaining thickness measuring step, the remaining thickness of the first step portion is measured with a non-contact film thickness meter that can measure the thickness without contacting the stacked wafers. [Effects of the Invention]

[0016] In one embodiment of the method for processing stacked wafers according to the present invention, a first step portion is formed in a first cutting step, the first step portion having a depth that does not reach the second wafer and in which the bonding layer remains, and then, in a second cutting step, the second cutting blade and the holding table are rotated (i.e., a so-called up-cut) so that the direction of the velocity vector of the lower end of the second cutting blade is opposite to the direction of the velocity vector of the stacked wafers at a position corresponding to the lower end of the second cutting blade, thereby cutting the bonding layer that is closer to the second wafer than the bottom surface of the first step portion.

[0017] In up-cutting, the cutting blade is less likely to bite into the cutting target than in down-cutting, so the cutting load is greater and the cutting blade is more likely to escape in a direction away from the bonding layer (e.g., upward). In addition, the material that makes up the bonding layer generally has a higher cutting load than the first and second wafers, which also causes the cutting blade to more easily escape in a direction away from the bonding layer.

[0018] Therefore, in the second cutting step, if the lower end of the second cutting blade is positioned closer to the second wafer than the bottom surface of the first step portion and at a position where the bonding layer is present, and the stacked wafers are cut by up-cutting, the first wafer can be completely removed without completely removing the bonding layer over the entire circumferential direction of the stacked wafers due to the high cutting load on the bonding layer (i.e., while leaving the bonding layer). [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a flow chart showing a method for processing laminated wafers. [Figure 2] FIG. [Figure 3] FIG. 3(A) is a plan view showing the first cutting step, and FIG. 3(B) is a partially cross-sectional side view showing the first cutting step. [Figure 4] 3 is an enlarged cross-sectional view of the outer periphery of the laminated wafer showing a first step portion. FIG. [Figure 5] FIG. 5(A) is a plan view showing the remaining thickness measuring step, and FIG. 5(B) is a partial cross-sectional side view showing the remaining thickness measuring step. [Figure 6] FIG. 6(A) is a plan view showing the second cutting step, and FIG. 6(B) is a partially cross-sectional side view showing the second cutting step. [Figure 7] FIG. 10 is an enlarged cross-sectional view showing a second cutting step. [Figure 8] 4 is an enlarged cross-sectional view of the outer periphery of the laminated wafer showing a second step portion. FIG. [Figure 9] FIG. 9(A) is a schematic diagram showing a first comparative example in which the second cutting step is performed by down cutting, and FIG. 9(B) is a schematic diagram showing a second comparative example in which the second cutting step is performed by down cutting. [Figure 10] FIG. 10 is a schematic diagram showing a state in which a second cutting step is performed by up-cutting. [Figure 11] FIG. 10 is a diagram showing the results of an experiment in which the difference between down-cutting and up-cutting was confirmed. DETAILED DESCRIPTION OF THE INVENTION

[0020] An embodiment according to one aspect of the present invention will be described with reference to the accompanying drawings. Fig. 1 is a flow diagram showing a method for processing a laminated wafer 11 (see Fig. 2). In this embodiment, the steps of a holding step S10, a first cutting step S20, a residual thickness measuring step S30, and a second cutting step S40 are performed in this order. First, the laminated wafer 11 to be processed by this processing method will be described.

[0021] The stacked wafers 11 each include a disk-shaped first wafer 13 and a disk-shaped second wafer 15. The first wafer 13 and the second wafer 15 each include a silicon single crystal substrate having the same diameter (e.g., 300±0.2 mm) defined by a predetermined standard such as the SEMI (Semiconductor Equipment and Materials International) standard.

[0022] The first wafer 13 has a circular front surface 13a and a circular back surface 13b located on opposite sides in the thickness direction. Similarly, the second wafer 15 has a circular front surface 15a and a circular back surface 15b located on opposite sides in the thickness direction.

[0023] Chamfered portions are formed on the outer peripheries of the front surface 13a and the back surface 13b of the first wafer 13. The chamfered portions are exaggerated in Fig. 2. It is planned that the back surface 13b of the first wafer 13 will be subjected to a grinding process after the second cutting step S40.

[0024] In this embodiment, similar chamfered portions are formed on the outer peripheries of the front surface 15a and the back surface 15b of the second wafer 15. However, the second wafer 15 does not necessarily have to have chamfered portions on the outer peripheries of the front surface 15a and the back surface 15b.

[0025] The ratio of diameter to thickness is sufficiently large in the first wafer 13 and the second wafer 15, and the chamfered portion is formed in a very narrow region on the outer periphery. For example, in the first wafer 13 and the second wafer 15, each of which has a diameter of 300 mm, the chamfered portion is formed in a range of less than 1.0 mm in width in the radial direction from the outer periphery 13c, 15c.

[0026] On the front surface 13a of the first wafer 13, a plurality of devices (not shown) such as ICs (Integrated Circuits) are formed on the center side of the chamfered portion in the radial direction of the first wafer 13.

[0027] Similarly, a plurality of devices such as ICs (not shown) are formed on the front surface 15a of the second wafer 15, closer to the center than the chamfered portion in the radial direction of the second wafer 15. However, the second wafer 15 does not necessarily have to be provided with any devices.

[0028] In this case, the second wafer 15 functions as a support substrate that supports the first wafer 13 during the thinning process of the first wafer 13. The support substrate may be any substrate having approximately the same diameter as the first wafer 13, and may be made of a semiconductor, resin, metal, ceramic, glass, or the like.

[0029] If the second wafer 15 does not have any devices, the etching step performed after edge trimming will not damage devices located near the periphery of the surface 15a of the second wafer 15. Therefore, it may seem acceptable to remove the periphery of the first wafer 13 and the bonding layer 17 so that the second wafer 15 is exposed.

[0030] However, when edge trimming is performed on the stacked wafers 11 so that the second wafer 15 is exposed, the surface 15a of the second wafer 15 is almost inevitably cut, so that the first wafer 13 cannot be peeled off from the second wafer 15 and the second wafer 15 cannot be reused, and the second wafer 15 must be discarded.

[0031] Therefore, even if no device is provided on the second wafer 15, when edge trimming is performed on the first wafer 13 of the stacked wafers 11, the first wafer 13 is completely removed over the entire outer periphery of the stacked wafers 11, but the bonding layer 17 is left in place so that the outer periphery of the second wafer 15 is not exposed, which has the advantage that the support substrate can be reused.

[0032] The surface 13a of the first wafer 13 and the surface 15a of the second wafer 15 are aligned so that the centers of the surfaces 13a and 15a are substantially aligned. The surfaces 13a and 15a are bonded together via a bonding layer 17. That is, in the laminated wafer 11, the first wafer 13 and the second wafer 15 are laminated with the bonding layer 17 sandwiched therebetween.

[0033] The bonding layer 17 is, for example, an oxide film such as silicon oxide, a nitride film such as silicon nitride, an oxynitride film such as silicon oxynitride, or a resin film made of an adhesive such as epoxy resin. The bonding layer 17 in this embodiment is an oxide film provided on the surface 13a of the first wafer 13 and functions as a passivation film or the like.

[0034] However, the bonding layer 17 may be provided on the surface 15a of the second wafer 15, or may be provided on each of the surfaces 13a and 15a. Furthermore, the bonding layer 17 may be a film provided separately from the first wafer 13 and the second wafer 15 in order to bond the first wafer 13 and the second wafer 15 together.

[0035] For example, the bonding layer 17 is a film formed by plasma CVD (Chemical Vapor Deposition) or the like separately from the first wafer 13 and the second wafer 15. This separate bonding layer 17 is an oxide film, a nitride film, an oxynitride film, or the like, and is formed on the surface 13 a of the first wafer 13 and / or the surface 15 a of the second wafer 15.

[0036] When the bonding layer 17 is an inorganic film such as an oxide film, this inorganic film is harder than a silicon single crystal substrate. Therefore, during cutting, a cutting blade (e.g., the second cutting blade 24 shown in FIGS. 6(A) and 6(B)) tends to move away from the bonding layer 17 (e.g., upward) due to a high cutting load.

[0037] Furthermore, when bonding layer 17 is a resin film such as an epoxy resin, bonding layer 17 is not as hard as an inorganic film, but the second cutting blade 24 is more likely to become clogged when cutting bonding layer 17, resulting in a high cutting load. Due to this high cutting load, second cutting blade 24 is more likely to move away from bonding layer 17, as in the case of an inorganic film.

[0038] The bonding layer 17 in this embodiment is provided in a circular region from each center of the surface 13a to a predetermined position of the chamfered portion in the radial direction of the surface 13a, and does not reach the outer peripheral edge 13c of the first wafer 13. Therefore, the bonding layer 17 is in contact with the second wafer 15 in a circular region that does not reach the outer peripheral edge 15c in the radial direction of the second wafer 15.

[0039] Next, we will explain the cutting device 2 used in the method for processing the laminated wafer 11. As shown in Fig. 2, the cutting device 2 has a disk-shaped chuck table (holding table) 4. The chuck table 4 has a disk-shaped frame body 4a.

[0040] A disk-shaped recess 4b is provided in the center of the top surface of the frame 4a. An annular suction groove 4c is provided on the top surface of an annular protrusion located outside the recess 4b. Negative pressure is transmitted to the suction groove 4c from a vacuum generating device (not shown) such as a vacuum pump or ejector.

[0041] The upper surface of the annular convex portion functions as a holding surface 4d that suction-holds the laminated wafers 11. The annular holding surface 4d is disposed approximately parallel to the XY plane defined by the X-axis and Y-axis directions of the cutting device 2.

[0042] A rotation shaft 4e is connected to the bottom of the chuck table 4. The rotation shaft 4e is arranged along the Z-axis direction, which is perpendicular to the X-axis and Y-axis directions of the cutting device 2, by a motor (not shown). When the rotation shaft 4e rotates, the chuck table 4 can rotate around the rotation shaft 4e (i.e., the rotation center 4e1).

[0043] The rotation direction of the rotation shaft 4e can be set to either clockwise or counterclockwise in plan view, and the rotation speed of the rotation shaft 4e (i.e., rotation angle per unit time) can be set to any value.

[0044] 3(A) and 3(B), a first cutting unit 6 is disposed above the chuck table 4. In this embodiment, the first cutting unit 6 cooperates with the chuck table 4 to perform a so-called down cut on the stacked wafers 11.

[0045] The first cutting unit 6 has a spindle housing 8 whose longitudinal portion is arranged along the Y-axis direction. A part of a cylindrical first spindle 10 is housed in the spindle housing 8 so that it can rotate by an air bearing (i.e., a hydrostatic air bearing). The tip of the first spindle 10 protrudes outside the spindle housing 8.

[0046] A motor (not shown) is provided near the base end of the first spindle 10. In Fig. 3(B), when the first spindle 10 is viewed in the Y-axis direction with the chuck table 4 sandwiched therebetween, the first spindle 10 is shown rotating counterclockwise, but the first spindle 10 can also rotate clockwise.

[0047] The longitudinal direction of the first spindle 10 is arranged approximately parallel to the Y-axis direction. A first cutting blade 14 having an annular cutting edge is attached to the tip of the first spindle 10 using a blade attachment mechanism 12.

[0048] The blade mounting mechanism 12 has a disk-shaped receiving flange portion 12a. The radial center of the receiving flange portion 12a is fixed to the tip of the first spindle 10 by a bolt or the like (not shown). The receiving flange portion 12a has a cylindrical boss portion (not shown) in its radial center.

[0049] The boss portion has a circular opening inserted therein for the first cutting blade 14. The first cutting blade 14 of this embodiment is a so-called hubless type (i.e., washer type) cutting blade, and is composed only of a cutting edge containing abrasive grains, a bond material, etc.

[0050] The first cutting blade 14 is used for so-called rough cutting, and therefore has a relatively large average grain size of abrasive grains. For example, the abrasive grains of the first cutting blade 14 have a predetermined grain size (e.g., #400) of #240 or more and less than #1500. Generally, the smaller the number indicating the grain size, the larger the average grain size of the abrasive grains.

[0051] The notation of particle size in this embodiment conforms to JIS R 6001-2:2017 (Particle size of abrasives for grinding wheels - Part 2: Fine powder) described in the JIS (Japanese Industrial Standards).

[0052] The grain size is determined, for example, using a sedimentation tube test method, an electrical resistance test method, etc. The grain size may be determined according to or in accordance with the notation generally used in the industry of manufacturing and selling grinding wheels.

[0053] The circular opening of the pressing flange 12b is inserted into the boss of the receiving flange 12a so that the receiving flange 12a and the annular pressing flange 12b sandwich the first cutting blade 14. A male thread is formed at the tip of the boss.

[0054] The male thread of the boss portion is fastened to the female thread of the annular press nut 12c, which has a female thread formed on its annular inner surface. In this way, the first cutting blade 14 is fixed to the first spindle 10 while being sandwiched between the receiving flange portion 12a and the press flange portion 12b.

[0055] The cutting device 2 is provided with a non-contact film thickness meter 8a (see FIGS. 5(A) and 5(B)). The non-contact film thickness meter 8a is a spectroscopic film thickness meter that uses interference of light over a wide wavelength band, a film thickness meter that uses infrared rays, a film thickness meter that uses ultrasonic waves, or the like. The non-contact film thickness meter 8a can measure the thickness of a thin film without coming into contact with the laminated wafer 11.

[0056] For example, if the non-contact film thickness meter 8a is a spectroscopic film thickness meter, the non-contact film thickness meter 8a includes an SLD (Super Luminescent Diode) light source (not shown), a sensor head 8a1, a spectrometer (not shown), a waveform analysis unit (not shown) realized by executing a program on a processor, etc. The sensor head 8a1 is fixed to the side of the spindle housing 8.

[0057] The cutting device 2 is a dicing saw with a parallel dual structure in which two spindles are arranged in parallel, and as shown in Figures 6(A) and 6(B), a second cutting unit 16 is arranged above the chuck table 4. The second cutting unit 16 has a spindle housing 18 whose longitudinal portion is arranged along the Y-axis direction.

[0058] A portion of a cylindrical second spindle 20 is rotatably accommodated in the spindle housing 18 by an air bearing. The longitudinal direction of the second spindle 20 is disposed substantially parallel to the Y-axis direction. A motor (not shown) is provided near the base end of the second spindle 20.

[0059] In Figure 6(A), when the second spindle 20 is viewed in the Y-axis direction with the chuck table 4 sandwiched between them, the second spindle 20 is shown rotating clockwise, but the second spindle 20 can also rotate counterclockwise.

[0060] The tip of the second spindle 20 protrudes outside the spindle housing 18. A second cutting blade 24 having an annular cutting edge is attached to the tip of the second spindle 20 using a blade attachment mechanism 22.

[0061] The blade mounting mechanism 22 has a disk-shaped receiving flange portion 22a, a holding flange portion 22b, and a holding nut 22c, similar to the blade mounting mechanism 12. The second cutting blade 24 of this embodiment is also a hubless type cutting blade.

[0062] The second cutting blade 24 is used for so-called finish cutting, and therefore has a relatively small average grain size of abrasive grains. For example, the abrasive grains of the second cutting blade 24 have a predetermined grain size (e.g., #1500) of #1500 or more and #3000 or less.

[0063] The second cutting blade 24 is inserted into the boss portion of the receiving flange portion 22a, and then inserted into the circular opening of the pressing flange portion 22b, and then the pressing nut 22c is tightened to the tip of the boss portion, whereby the second cutting blade 24 is fixed to the second spindle 20 while being clamped between the receiving flange portion 22a and the pressing flange portion 22b.

[0064] Next, a method for processing the laminated wafers 11 will be described with reference to Fig. 2 to Fig. 8. Fig. 2 is a partial cross-sectional side view showing the holding step S10. In the holding step S10, first, the laminated wafers 11 are placed on the chuck table 4 so that the back surface 13b of the first wafer 13 is exposed upward. At this time, the radial center of the laminated wafers 11 and the rotation center 4e1 are approximately aligned.

[0065] Next, negative pressure is applied to the suction groove 4c, whereby the stacked wafers 11 are suction-held on the holding surface 4d. After the holding step S10, in the first cutting step S20, the first cutting blade 14 cuts the outer periphery of the first wafer 13, thereby forming a first step portion 13d (see FIG. 4) that does not reach the bonding layer 17.

[0066] Fig. 3(A) is a plan view showing the first cutting step S20, and Fig. 3(B) is a partially cross-sectional side view showing the first cutting step S20. In Fig. 3(A) and Fig. 3(B), an extension line of the rotation center 10a of the first spindle 10 is indicated by a dashed line. In Fig. 3(A), the rotation center 4e1 of the rotation shaft 4e is indicated by an X, and in Fig. 3(B), the rotation center 4e1 is indicated by a dashed line.

[0067] In the first cutting process S20, the first cutting blade 14 is rotated around the first spindle 10 while the chuck table 4 is rotated around the rotation axis 4e while the first cutting blade 14 is cutting into the outer periphery of the first wafer 13, thereby forming a first step portion 13d on the outer periphery of the first wafer 13.

[0068] Specifically, first, the lower end 14a of the first cutting blade 14, which is rotating at high speed (e.g., 30,000 rpm), is positioned a predetermined distance (e.g., 10 μm) above the surface 13a of the first wafer 13, and the chuck table 4 is translated in the X-axis direction to cause the first cutting blade 14 to cut into the outer periphery of the first wafer 13 (i.e., perform a slide-in cut).

[0069] After the chuck table 4 is moved along the X-axis direction to a position where an extension of the rotation center 10a of the first spindle 10 intersects with the rotation center 4e1 of the rotation axis 4e (see FIG. 3(A)), rotation of the chuck table 4 is started. The rotation angle of the rotation axis 4e per unit time is set to, for example, 5° / s.

[0070] By rotating the chuck table 4 at least once, the outer periphery of the first wafer 13 is cut over the entire circumferential direction of the first wafer 13, and 50% to 99% of the thickness of the first wafer 13 is removed from the outer periphery of the first wafer 13. In this manner, an annular first step portion 13d is formed.

[0071] In the first cutting step S20, the rotational direction of the first cutting blade 14 (i.e., the rotational direction of the first spindle 10) is adjusted so that the stacked wafers 11 are down-cut.

[0072] To cut the stacked wafers 11 using down-cutting, the first cutting blade 14 and the chuck table 4 are rotated so that the direction of the velocity vector of the lower end 14a of the first cutting blade 14 is the same as the direction of the velocity vector of the stacked wafers 11 at the position corresponding to the lower end 14a in the XY plane.

[0073] At this time, the first cutting blade 14 rotates from the back surface 13b of the first wafer 13 that is in contact with the first cutting blade 14 toward the lower end 14a of the first cutting blade 14, and the outer periphery of the stacked wafers 11 rotates so as to approach the first cutting blade 14.

[0074] In this embodiment, when the first cutting blade 14 is viewed in the Y-axis direction with the chuck table 4 sandwiched therebetween, the first cutting blade 14 is rotated counterclockwise, and the chuck table 4 is rotated clockwise in a plan view.

[0075] By employing down-cutting in the first cutting step S20, chipping on the rear surface 13b and peeling of the coating film provided on the rear surface 13b can be prevented compared to when up-cutting is employed, and the depth of the actually formed first step portion 13d can be matched with high precision to the set depth of the first step portion 13d.

[0076] However, down cutting is not essential in the first cutting step S20. Although down cutting has many advantages over up cutting in the first cutting step S20, up cutting may also be employed.

[0077] In addition, when performing up-cutting in the first cutting step S20, this can be achieved by reversing the rotation direction of the first cutting blade 14 or the rotation direction of the chuck table 4 in Figures 3(A) and 3(B).

[0078] 4 is an enlarged cross-sectional view of the outer periphery of the laminated wafer 11 showing the first step portion 13d. By forming the first step portion 13d, the outer periphery 13c moves to a position closer to the second wafer 15 than before the formation of the first step portion 13d.

[0079] 4, when viewed from the rear surface 13b as the starting point, the first step portion 13d does not reach the second wafer 15 and has a depth that allows the first wafer 13 to remain. A remaining thickness 13e of the first wafer 13 that remains below the first step portion 13d is, for example, approximately 10 μm.

[0080] In the first cutting step S20, the first step portion 13d is formed using the first cutting blade 14 having abrasive grains with a relatively large grain size, thereby shortening the working time and reducing wear on the cutting blade compared to forming the first step portion 13d using the second cutting blade 24 having abrasive grains with a relatively small grain size.

[0081] After the first cutting step S20 and before the second cutting step S40, the remaining thickness 13e of the first wafer 13 remaining between the bottom surface 13d1 of the first step portion 13d and the second wafer 15 is measured by a non-contact film thickness meter 8a (remaining thickness measuring step S30). Fig. 5(A) is a plan view showing the remaining thickness measuring step S30, and Fig. 5(B) is a partial cross-sectional side view showing the remaining thickness measuring step S30.

[0082] In the remaining thickness measuring step S30, first, the sensor head 8a1 of the non-contact film thickness meter 8a is positioned directly above the first step portion 13d by adjusting the position of the first cutting unit 6 in the Y-axis direction and the position of the chuck table 4 in the X-axis direction. At the same time, the position of the first cutting unit 6 in the Z-axis direction is adjusted to adjust the position of the sensor head 8a1 in the Z-axis direction.

[0083] Next, the sensor head 8a1 irradiates light, ultrasonic waves, or the like, and rotates the chuck table 4 at a predetermined rotation angle per unit time. Based on the data obtained via the sensor head 8a1, the non-contact film thickness meter 8a measures the remaining thickness 13e of the first wafer 13 remaining below the first step portion 13d.

[0084] After the remaining thickness measuring step S30, the second cutting step S40 is performed. Figure 6(A) is a plan view showing the second cutting step S40, and Figure 6(B) is a partially cross-sectional side view showing the second cutting step S40. Figure 7 is an enlarged cross-sectional view showing the second cutting step S40.

[0085] In the second cutting process S40, first, the second cutting blade 24 is rotated around the second spindle 20, and the lower end 24a of the second cutting blade 24 is positioned at a predetermined position 17a (see Figure 7) that is closer to the second wafer 15 in the thickness direction 11c of the stacked wafers 11 than the bottom surface 13d1 of the first step portion 13d and where the bonding layer 17 is present.

[0086] In the second cutting step S40, the bottom surface 13d1 of the first step portion 13d is used as a reference, and it is determined how far below the bottom surface 13d1 the lower end 24a of the second cutting blade 24 should be positioned. In other words, by setting the position of the lower end 24a of the second cutting blade 24 based on the remaining thickness 13e obtained in the remaining thickness measurement step S30, the position of the lower end 24a can be set to the predetermined position 17a with high accuracy.

[0087] In the second cutting step S40, the second cutting blade 24 and the chuck table 4 are rotated so that the direction 24b of the velocity vector at the lower end 24a of the second cutting blade 24 is opposite to the direction 11b of the velocity vector of the stacked wafers 11 at a position corresponding to the lower end 24a on the XY plane (see FIG. 7). In FIG. 7, the outline arrow indicates the rotation direction at the outer periphery of the stacked wafers 11.

[0088] In the second cutting process S40, the second cutting blade 24 rotates from the lower end 24a of the second cutting blade 24 toward the contact point 13d2 of the bottom surface 13d1 of the first step portion 13d that contacts the second cutting blade 24 (i.e., an up-cut), and the outer periphery of the stacked wafer 11 rotates to approach the second cutting blade 24.

[0089] The rotation speed is, for example, 30,000 rpm. Next, with the lower end 24a of the second cutting blade 24 rotating at high speed positioned at the predetermined position 17a, the chuck table 4 is translated along the X-axis direction until the extension of the rotation center 20a intersects with the rotation center 4e1 (i.e., a slide-in cut is performed) (see FIG. 6(A)).

[0090] Then, the chuck table 4 is rotated around the rotation axis 4e at a rate of, for example, 5° / s to remove the outer periphery of the laminated wafers 11. By rotating the chuck table 4 at least once, the outer periphery of the laminated wafers 11 is cut over the entire circumferential direction of the laminated wafers 11.

[0091] In this way, the first wafer 13 directly below the first step portion 13d is completely removed, but the bonding layer 17 located near the surface 13a of the first wafer 13 is not completely removed in the circumferential direction, forming a second step portion 23d (see FIG. 8). That is, in the second step portion 23d, the bonding layer 17 remains in the circumferential direction of the stacked wafers 11.

[0092] 8 is an enlarged cross-sectional view of the second step portion 23d formed in the second cutting step S40 at the outer periphery of the laminated wafer 11. As shown in Fig. 8, at the bottom surface 23d1 of the second step portion 23d, the bonding layer 17 is not completely removed in the thickness direction 11c, and the bonding layer 17 remains over the entire circumferential direction of the laminated wafer 11.

[0093] In the up-cutting method used in the second cutting step S40, the second cutting blade 24 is less likely to bite into the cutting target than in the down-cutting method, resulting in a larger cutting load, and the second cutting blade 24 is more likely to move away from the bonding layer 17 (e.g., upward).

[0094] In addition, the material used as the bonding layer 17 generally has a higher cutting load than the first wafer 13 and the second wafer 15, which also causes the second cutting blade 24 to more easily escape in a direction away from the bonding layer 17.

[0095] Therefore, in the second cutting step S40, if the lower end 24a of the second cutting blade 24 is positioned closer to the second wafer 15 than the bottom surface 13d1 of the first step portion 13d and at a predetermined position 17a where the bonding layer 17 is present, and the stacked wafers 11 are cut by up-cutting, the first wafer 13 can be completely removed without completely removing the bonding layer 17 over the entire circumferential direction of the stacked wafers 11 due to the high cutting load on the bonding layer 17 (i.e., while leaving the bonding layer 17).

[0096] In this way, in this embodiment, by actively utilizing the difficulty of cutting the bonding layer 17 in up-cutting, it is possible to achieve both the retention of the bonding layer 17 over the entire circumferential direction of the stacked wafers 11 and the complete removal of the first wafer 13.

[0097] Next, a comparison between a case where down-cutting is performed (comparative example) and a case where up-cutting is performed (the above-described embodiment) in the second cutting step S40 will be described with reference to Figures 9(A), 9(B), and 10. In Figures 9(A), 9(B), and 10, the outline arrows indicate the rotation direction at the outer periphery of the laminated wafer 11, and for convenience of explanation, indicate the bottom surface 13d1 of the first step portion 13d.

[0098] When edge trimming is performed on the stacked wafer 11 using the second cutting blade 24, the lower end 24a of the second cutting blade 24 cannot always maintain a constant depth, but moves up and down within a certain range (for example, a range of ±0.5 μm) along the Z-axis direction from the set depth position.

[0099] Therefore, for example, if the thickness of the bonding layer 17 is 1.0 μm and the set depth position 24a1 is within the bonding layer 17 located above the depth position 17b, which is half the thickness of the bonding layer 17 based on the bottom surface 13d1 of the first step portion 13d (i.e., a position closer to the first wafer 13), as shown in Figure 9(A), a region of the first wafer 13 that remains uncut will be generated.

[0100] 9A is a schematic diagram showing a first comparative example in which the second cutting step S40 is performed by down-cutting. In FIG. 9A, the trajectory 24a2 of the lower end 24a in the second cutting step S40 is shown by a solid line, and the uncut region of the first wafer 13 is hatched.

[0101] In contrast, as shown in Figure 9(B), if the set depth position 24a1 is within the bonding layer 17 located below the depth position 17b that is half the thickness of the bonding layer 17 based on the bottom surface 13d1 of the first step portion 13d (i.e., a position closer to the second wafer 15), a region in which the second wafer 15 is exposed will be created when the bonding layer 17 is removed.

[0102] 9(B) is a schematic diagram showing a second comparative example in which the second cutting step S40 is performed by down-cutting. In FIG. 9(B), the trajectory 24a2 of the lower end 24a in the second cutting step S40 is shown by a solid line, and the exposed region of the second wafer 15 is hatched.

[0103] 10 is a schematic diagram showing the second cutting step S40 being performed by up-cutting, as in the present embodiment. As described above, up-cutting imposes a higher cutting load than down-cutting.

[0104] Therefore, when the set depth position 24a1 is within the thickness range of the bonding layer 17 and the lower end 24a of the second cutting blade 24 reaches the bonding layer 17 during actual cutting, the actual cutting depth will be, for example, approximately 35% of the set depth position 24a1.

[0105] In this case, the 1.0 μm thick bonding layer 17 appears to be 2.9 μm (=1 μm / 0.35). Therefore, if the set depth position 24a1 is set to the depth position 17b that is half the thickness of the bonding layer 17, the lower end 24a of the second cutting blade 24 will be within the thickness range of the bonding layer 17 even if it is shifted 0.5 μm above or below the half depth position 17b.

[0106] Furthermore, even if the depth position 24a1 is set to a position in the Z-axis direction where the second wafer 15 is located near the surface 15a, if an upcut is used in the second cutting process S40, the first wafer 13 can be completely removed in the circumferential direction of the stacked wafers 11 without cutting the second wafer 15 due to the relatively high cutting load, and the bonding layer 17 can be exposed.

[0107] Next, experimental results regarding the difference between down-cutting and up-cutting will be described with reference to Fig. 11. Fig. 11 is a diagram showing experimental results confirming the difference between down-cutting and up-cutting. #1 to #4 in Fig. 11 are experimental results when the above-mentioned second cutting blade 24 was cut into a silicon single crystal substrate (i.e., a film-less wafer) that does not have a silicon oxide film corresponding to the bonding layer 17.

[0108] #5 to #7 in Fig. 11 are experimental results when the second cutting blade 24 was caused to cut only into the silicon oxide film (i.e., the silicon oxide film was cut with the lower end 24a of the second cutting blade 24 positioned within the silicon oxide film) in a silicon single crystal substrate (i.e., a film-coated wafer) having a silicon oxide film corresponding to the bonding layer 17 on its entire surface. #8 in Fig. 11 is an experimental result when the second cutting blade 24 was caused to cut into the film-coated wafer beyond the silicon oxide film and further into the silicon single crystal substrate.

[0109] The shaded bars indicate the target depth [μm] (i.e., the set depth position 24a1), and the open bars indicate the actual machining depth [μm] (i.e., the depth actually cut when the depth position 24a1 is set). DOWN is the above-mentioned down-cutting, and UP is the above-mentioned up-cutting.

[0110] The actual processing depth is the average value of the depths at eight points spaced at approximately equal intervals in the circumferential direction of the film-free wafer and the film-coated wafer. The actual processing ratio [%] was calculated by {(actual processing depth / target depth) × 100}.

[0111] For wafers without a film, the actual processing rate for #1 was approximately 93%. In contrast, the actual processing rate for #2 was approximately 44%, for #3 approximately 60%, and for #4 approximately 53%. As is clear from #2, #3, and #4, the cutting load is high with up-cutting, and it is not possible to cut to the desired depth.

[0112] On the other hand, for the film-coated wafers, the actual processing ratio for #5 was approximately 73%, while the actual processing ratio for #6 was approximately 36%, the actual processing ratio for #7 was approximately 34%, and the actual processing ratio for #8 was approximately 56%.

[0113] As is clear from the comparison of #5, #6, and #7, when cutting a silicon oxide film by up-cutting, the cutting load increases significantly. In the above-described embodiment, by actively utilizing the difficulty of cutting a silicon oxide film by up-cutting, it is possible to both leave the bonding layer 17 over the entire circumferential direction of the stacked wafers 11 and completely remove the first wafer 13.

[0114] In addition, the structures, methods, etc. according to the above-described embodiments can be modified as appropriate without departing from the scope of the present invention. For example, a dicing saw having one spindle (i.e., the first spindle 10 and the second spindle 20 are the same) can be used.

[0115] In this case, in the first cutting step S20 and the second cutting step S40, the same cutting blade is used as the first cutting blade 14 and the second cutting blade 24 after being attached to the single spindle.

[0116] If the rough cutting blade corresponding to the first cutting blade 14 is used consistently, the processing work can be carried out relatively efficiently in the first cutting step S20, and if the finish cutting blade corresponding to the second cutting blade 24 is used consistently, the processing quality in the second cutting step S40 can be maintained relatively high. [Explanation of symbols]

[0117] 2:Cutting device 4: chuck table (holding table), 4a: frame body, 4b: recess, 4c: suction groove 4d: Holding surface, 4e: Rotation axis, 4e1: Rotation center 6: First cutting unit, 8: Spindle housing 8a: Non-contact film thickness meter, 8a1: Sensor head 10: first spindle, 10a: rotation center 11: stacked wafer, 11b: orientation, 11c: thickness direction 12: Blade attachment mechanism 12a: Receiving flange portion, 12b: Pressing flange portion, 12c: Pressing nut 13: first wafer, 13a: front surface, 13b: back surface, 13c: outer periphery 13d: first step, 13d1: bottom surface, 13d2: contact point, 13e: remaining thickness 14: first cutting blade, 14a: lower end 15: second wafer, 15a: front surface, 15b: back surface, 15c: outer periphery 16: Second cutting unit, 18: Spindle housing 17: bonding layer, 17a: predetermined position, 17b: half depth position 20: second spindle, 20a: rotation center 22: Blade attachment mechanism 22a: receiving flange portion, 22b: holding flange portion, 22c: holding nut 23d: second step, 23d1: bottom surface 24: second cutting blade, 24a: lower end, 24a1: depth position, 24a2: trajectory 24b: Orientation S10: Holding process, S20: First cutting process S30: Residual thickness measurement process, S40: Second cutting process

Claims

1. A method for processing a laminated wafer in which a first wafer and a second wafer are laminated with a bonding layer sandwiched therebetween, comprising: a holding step of holding the stacked wafers on a holding table rotatable around a predetermined rotation axis so that the first wafer is exposed; a first cutting step in which, after the holding step, a first cutting blade attached to a tip of a first spindle is rotated around the first spindle and the holding table is rotated around the rotation axis while the first cutting blade is cutting into the outer periphery of the first wafer, thereby forming a first step portion in the outer periphery of the first wafer, the first step portion having a depth that does not reach the second wafer and leaves the first wafer; a second cutting step of removing the outer periphery of the laminated wafers by rotating a second cutting blade attached to a tip of a second spindle around the second spindle after the first cutting step and rotating the holding table around the rotation axis while positioning a lower end of the second cutting blade closer to the second wafer in the thickness direction of the laminated wafers than the bottom surface of the first step portion and at a position where the bonding layer is present; Equipped with A method for processing laminated wafers, characterized in that in the second cutting process, the second cutting blade and the holding table are rotated so that the direction of the velocity vector of the lower end of the second cutting blade is opposite to the direction of the velocity vector of the laminated wafer at the position corresponding to the lower end.

2. a remaining thickness measuring step of measuring a remaining thickness of the first wafer remaining between a bottom surface of the first step portion and the second wafer after the first cutting step and before the second cutting step, 2. The method for processing laminated wafers according to claim 1, wherein in the second cutting step, the position of the lower end of the second cutting blade is set based on the remaining thickness obtained in the remaining thickness measurement step.

3. 3. The method for processing laminated wafers according to claim 2, wherein the remaining thickness measuring step measures the remaining thickness of the first step portion using a non-contact film thickness meter that can measure thickness without contacting the laminated wafers.

Images