Wafer processing method

Abstract

The present invention provides a wafer processing method that allows for the proper removal of the chamfered portion of the first wafer when joining a first wafer and processing the first wafer. [Solution] The method comprises a modified layer formation step of positioning the focusing point of a laser beam LB on the inside adjacent to the chamfered portion 17A formed on the outer circumference of the first wafer 10A and irradiating it to form a ring-shaped modified layer 100, and a chamfered portion removal step of removing the chamfered portion 17A of the first wafer 10A starting from the modified layer 100, and before the chamfered portion removal step, a chamfered portion removal acceleration step of supplying water L from the outer circumference to the bonding surface 20 of the bonded wafer W and allowing the water L to penetrate to an area where the chamfered portion 17A can be removed to weaken the bonding force, and a confirmation step of confirming whether the penetration of water L is appropriate or inappropriate.

Description

Technical Field

[0001] The present invention relates to a method for processing a wafer by performing processing on a bonded wafer in which a first wafer and a second wafer are bonded together.

Background Art

[0002] A wafer on which a plurality of devices such as ICs and LSIs are partitioned by a dicing line and formed on the surface is thinned by grinding the back surface with a grinding device, and then divided into individual device chips by a dicing device or a laser processing device, and used in electrical devices such as mobile phones and personal computers.

[0003] In addition, a chamfer is formed on the outer periphery of the wafer. When the back surface of the wafer is ground, the chamfer becomes a sharp knife edge, cracks occur from the knife edge and reach the inside, causing problems such as damaging the device or injuring the operator. Therefore, a technique for removing the chamfer of the wafer has been proposed (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in a technique of bonding the surface of the first wafer and the surface of the second wafer in order to improve the function of the device, and then grinding the back surface of the first wafer to form a desired thickness, there is a problem that it is relatively difficult to remove the chamfer from the first wafer.

[0006] That is, (1) Wafers bonded by siloxane bonds, etc., have strong bonding forces, and even if a laser beam with a wavelength that is transparent to the first wafer is focused on the inside adjacent to the chamfered portion and irradiated to form a modified layer inside the first wafer, it is difficult to remove the chamfered portion well by this alone. (2) When removing the chamfered portion from the first wafer using a cutting blade, it is difficult to remove it completely without damaging the second wafer. This presents a problem.

[0007] The present invention has been made in view of the above facts, and its main technical problem is to provide a wafer processing method that can appropriately remove the chamfered portion of the first wafer when joining a first wafer and processing the first wafer. [Means for solving the problem]

[0008] To solve the above-mentioned main technical problems, the present invention provides a wafer processing method for a bonded wafer formed by joining a first wafer and a second wafer, comprising: a modified layer formation step of positioning the focusing point of a laser beam on the inside adjacent to a chamfer formed on the outer circumference of the first wafer and irradiating it to form a ring-shaped modified layer; a chamfer removal step of removing the chamfer of the first wafer starting from the modified layer, and before the chamfer removal step, a chamfer removal acceleration step of supplying water from the outer circumference to the bonding surface of the bonded wafer and allowing the water to penetrate to an area where the chamfer can be removed to weaken the bonding force; and a confirmation step of confirming whether the water penetration is appropriate or inappropriate.

[0009] The verification step preferably involves irradiating the chamfered area from the side of the first or second wafer with infrared wavelength light that is transparent to the bonded wafer and absorbent to water, and detecting the amount of transmitted or reflected light to confirm whether the water infiltration is appropriate or inappropriate. Furthermore, the infrared wavelength is preferably 800 nm to 10000 nm. More preferably, the infrared wavelength is around 1450 nm, around 1940 nm, or around 2900 nm.

[0010] In the verification step, if the water infiltration is appropriate, the chamfer removal step is performed. If the water infiltration is inappropriate, it is preferable to repeat the modified layer formation step or the chamfer removal acceleration step, or to repeat both the modified layer formation step and the chamfer removal acceleration step. Furthermore, in the modified layer formation step, it is preferable to form a radial modified layer from the ring-shaped modified layer outward. Moreover, it is preferable that the chamfer removal acceleration step is performed before, after, or simultaneously with the modified layer formation step. It is preferable that the chamfer removal step is performed simultaneously with the grinding step, which thins the upper surface of the first wafer by grinding it.

[0011] The first wafer and the second wafer are joined by siloxane bonds, which are Si-O-Si bonds. It is preferable that the Si-O-Si bonds are changed to Si-OH-OH-Si bonds by water that penetrates during the chamfer removal acceleration process, thereby weakening the bonding strength. [Effects of the Invention]

[0012] The wafer processing method of the present invention is a wafer processing method for a bonded wafer formed by joining a first wafer and a second wafer, comprising: a modified layer formation step of positioning the focusing point of a laser beam on the inside adjacent to a chamfer formed on the outer circumference of the first wafer and irradiating it to form a ring-shaped modified layer; a chamfer removal step of removing the chamfer of the first wafer starting from the modified layer, and before the chamfer removal step, a chamfer removal acceleration step of supplying water from the outer circumference to the bonding surface of the bonded wafer to allow the water to penetrate to an area where the chamfer can be removed to weaken the bonding force; and a confirmation step of confirming whether the water penetration is appropriate or inappropriate. As a result, it is possible to confirm in advance that the bonding force in the area where the chamfer is to be removed has been weakened, and before performing the chamfer removal step, the bonding force in the area where the chamfer is to be removed can be reliably weakened, and the chamfer can be easily removed from the first wafer starting from the modified layer. Therefore, wafers bonded by siloxane bonds, etc., have strong bonding forces. Even if a laser beam with a wavelength that is transparent to the first wafer is focused on the inside adjacent to the chamfered area and irradiated, forming a modified layer inside the first wafer, the problem that it is difficult to remove the chamfered area well by this method alone is resolved. Furthermore, since there is no need to use a cutting blade to remove the chamfered area from the first wafer, the problem that it is difficult to completely remove the chamfered area from the second wafer without damaging it when using a cutting blade is also resolved. [Brief explanation of the drawing]

[0013] [Figure 1] This is a perspective view showing a bonded wafer processed by the wafer processing method of this embodiment. [Figure 2] This is a perspective view showing how a bonded wafer is held on a chuck table during the modified layer formation process. [Figure 3](a) A perspective view showing an embodiment in which the modified layer formation process and the chamfer removal acceleration process are carried out at the same time; (b) A side view showing a part of the embodiment shown in (a) enlarged and a cross-sectional view of that part; (c) A plan view of a bonded wafer in which a radial modified layer has been formed on the first wafer. [Figure 4] (a) A perspective view showing an embodiment of the verification process, and (b) A conceptual diagram showing the relationship between the rotation angle of the chuck table and the amount of infrared light received. [Figure 5] This is a perspective view showing an embodiment of a grinding process that also serves as a chamfer removal process. [Modes for carrying out the invention]

[0014] Hereinafter, embodiments relating to a wafer processing method constructed according to the present invention will be described in detail with reference to the attached drawings.

[0015] The wafer processing method constructed according to the present invention involves processing a bonded wafer W formed by joining a first wafer 10A and a second wafer 10B, as illustrated in Figure 1.

[0016] The first wafer 10A shown in Figure 1 is, for example, a silicon (Si) wafer with a diameter of 300 mm and a thickness of 700 μm, and multiple devices 12A are formed on the surface 10Aa, partitioned by division lines 14A. The first wafer 10A has a surface 10Aa and a back surface 10Ab, and comprises a device region 16A near the center on which the devices 12A to be used as a product are formed, and an outer peripheral excess region 18A surrounding the device region 16A, which includes a chamfered portion 17A with a width of 0.5 to 2 mm formed on the outer periphery. Furthermore, the second wafer 10B, which together with the first wafer 10A constitutes the bonded wafer W, has the same configuration as the first wafer 10A, with a chamfered portion 17B formed on its outer circumference. Although not shown in the figure, the silicon (Si) wafer has a surface 10Ba facing downwards in the figure, on which multiple devices corresponding to the device 12A formed on the first wafer 10A are demarcated by planned division lines.

[0017] As shown in Fig. 1, the above-mentioned bonded wafer W is formed by bonding the surface 10Aa of the first wafer 10A and the surface 10Ba of the second wafer 10B, and forming a bonding surface 20 by a siloxane bond to be integrated. The siloxane bond is a Si-O-Si bond in which silicon (Si) and oxygen (O) are alternately bonded. Since the first wafer 10A and the second wafer 10B are bonded by heat treatment, a strong bonding state is maintained even at high temperatures.

[0018] If the above-mentioned bonded wafer W is produced, the method for processing the wafer according to the present embodiment described below is carried out.

[0019] (Modified layer forming step) When carrying out the method for processing the wafer according to the present embodiment, first, a modified layer forming step of forming a ring-shaped modified layer 100 for removing the chamfer portion 17A of the first wafer 10A is carried out by positioning the condensing point of the laser beam LB having a wavelength permeable to the first wafer 10A inside adjacent to the chamfer portion 17A formed on the outer periphery of the first wafer 10A and irradiating it. The procedure will be described more specifically below.

[0020] Once the above-described bonded wafer W is prepared, the bonded wafer W is transported to a laser processing apparatus 40 (only a part thereof is shown) shown in FIGS. 2 and 3. The laser processing apparatus 40 includes at least a holding means 41 for holding the bonded wafer W and a laser beam irradiation means 45 (see FIG. 3) for irradiating a laser beam LB having a wavelength that is transmissive to the first wafer 10A. As shown in FIG. 2, the holding means 41 includes a chuck table 42, a rotating shaft 43 disposed on the lower surface side of the chuck table 42, and a drive motor 44 for rotationally driving the chuck table 42 via the rotating shaft 43. The chuck table 42 includes a holding surface 42a formed of a breathable member and a frame body 42b surrounding the holding surface 42a, and a suction means (not shown) is connected via the frame body 42b. By operating the suction means, a negative pressure is generated on the holding surface 42a. The illustrated chuck table 42 is formed with a dimension smaller than the diameter of the bonded wafer W, for example, with a dimension about 30 mm smaller in diameter than the bonded wafer W.

[0021] The bonded wafer W transported to the laser processing apparatus 40 is placed on the chuck table 42 of the holding means 41 with the chamfered portion 17A removed of the first wafer 10A facing upward so that the centers coincide, and the suction means described above is operated to generate a negative pressure on the holding surface 42a, and is sucked and held by the chuck table 42. As a result, the bonded wafer W is held in a state where it protrudes 15 mm from the outer peripheral end of the chuck table 42. Next, using an alignment means (not shown) disposed in the laser processing apparatus 40, the bonded wafer W sucked and held by the chuck table 42 is imaged, and alignment is performed to detect a processing position where the condensing point of the laser beam LB should be irradiated inside adjacent to the chamfered portion 17A of the first wafer 10A.

[0022] Based on the processing position information detected by the alignment described above, the moving means (not shown in the figure) is activated to move the chuck table 42, and as shown in Figure 3(a), the processing position set on the first wafer 10A of the bonded wafer W is positioned directly below the concentrator 46 of the laser beam irradiation means 45. Next, as can be understood from Figure 3(b) in addition to Figure 3(a), the focal point of the laser beam LB is positioned and irradiated from the back surface 10Ab side of the first wafer 10A to the processing position on the first wafer 10A, and the chuck table 42 is rotated in the direction indicated by arrow R1 in Figure 3(a). In this way, a ring-shaped modified layer 100 is formed along the inside of the chamfered portion 17A of the first wafer 10A.

[0023] The modified layer 100 described above is preferably formed in multiple layers in the vertical direction, as shown in Figure 3(b). For example, the modified layer 100 shown in Figure 3(b) is composed of four modified layers in the vertical direction. When forming such a modified layer 100 including multiple layers, first, the focal point of the laser beam LB is positioned and irradiated so that the first modified layer is formed inside the first wafer 10A, adjacent to the chamfered portion 17A of the first wafer 10A, more specifically at a position 2.0 mm inward from the outer edge of the first wafer 10A, and within the first wafer 10A, for example, at a depth of 700 μm from the back surface 10Ab of the first wafer 10A, i.e., near the surface 10Aa of the first wafer 10A, and the chuck table 42 is rotated in the direction indicated by the arrow R1 described above to form the first ring-shaped modified layer. Subsequently, while rotating the chuck table 42, the focusing point is raised three times toward the back surface 10Ab side (upwards), so that a total of four ring-shaped modified layers 100 are formed along the chamfered portion 17A, for example, so that the depth to which the modified layer 100 is formed from the back surface 10Ab is 500 μm → 300 μm → 150 μm. Note that the modified layer 100 shown in Figure 3(b) is shown conceptually for illustrative purposes, and the size and depth position of each layer do not correspond to the actual dimensions. With this, the modified layer formation process is completed. Note that the modified layer 100 described above is not limited to being formed of four layers, and the appropriate number of layers is set depending on the wavelength and output of the laser beam LB irradiated by the laser beam irradiation means 42, the thickness of the first wafer 10A, the material constituting the first wafer 10A, etc.

[0024] The laser processing conditions used when carrying out the modified layer formation process described above are set to, for example, the following laser processing conditions. Wavelength: 1342nm Repetition frequency: 80kHz Machining feed rate: 60 rpm (rotational speed of chuck table 42) Average output: 2.0W

[0025] In the modified layer formation process described above, for example, as shown in Figure 3(c), a radial modified layer 102 may be formed extending outwards from the region where the modified layer 100 is formed, toward the area where the chamfered portion 17A is formed. The radial modified layer 102 shown in the figure is formed, for example, by irradiating a laser beam LB with the same wavelength, repetition frequency, and average power as used to form the modified layer 100, and is formed at multiple locations (four locations in the illustrated embodiment) at equal intervals on the outer circumference of the first wafer 10A. By forming this radial modified layer 102, when the chamfered portion 17A is removed from the first wafer 10A in the chamfered portion removal process described later, the chamfered portion 17A is divided into multiple fragments 17A' (see Figure 5), and the removal of the chamfered portion 17A is carried out smoothly.

[0026] (Process to accelerate removal of chamfered areas) In the wafer processing method of this embodiment, a chamfer removal acceleration step described below is performed before the chamfer removal step described later. This chamfer removal acceleration step can be performed before or after the modified layer formation step described above, or at the same time as the modified layer formation step. In the following description, it will be described as being performed at the same time as the modified layer formation step described above.

[0027] When performing the chamfer removal acceleration process of this embodiment, a fluid supply means 47 is positioned to the side of the bonded wafer W, as shown in Figures 3(a) and 3(b). The fluid supply means 47 is installed in the laser processing apparatus 40 described above and includes a nozzle 47a that sprays a fluid L (water, preferably pure water) horizontally to weaken the bonding force between the first wafer 10A and the second wafer 10B. The fluid supply means 47 includes a moving means (not shown) that moves the nozzle 47a vertically and horizontally toward the center of the chuck table 42. The moving means is activated to position the tip 47b of the nozzle 47a near the side of the bonding surface 20 of the bonded wafer W, and in conjunction with performing the modified layer formation process described above, that is, while rotating the chuck table 42 in the direction indicated by arrow R1, the modified layer 100 is formed, and a fluid supply source (not shown) is activated to supply fluid L from the tip 47b of the nozzle 47. The fluid L can be any fluid that has the effect of weakening the bonding force of the bonding surface 20, and is not limited to liquid water as described above, but may be supplied in the form of water vapor. As described above, when the fluid L is supplied from the fluid supply means 47, the fluid L penetrates into the bonding surface 20 on the outer circumference where the chamfered portions 17A and 17B are joined, and the region joined by siloxane bonds changes to Si-OH-OH-Si bonds. As a result, as shown in Figure 3(b), the fluid L penetrates to an area where the chamfered portion 17A of the first wafer 10A can be removed, and a ring-shaped delamination layer 21 is formed in which the bonding force of the bonding surface 20 is weakened.

[0028] In the embodiment described above, the chamfer removal acceleration process was carried out simultaneously with the modified layer formation process. However, it is also possible to carry out the modified layer formation process before or after the process. In that case, the fluid supply means 47 can be prepared separately from the laser processing apparatus 40, and a fluid L (water) that weakens the bonding force between the first wafer 10A and the second wafer 10B can be supplied to the bonding surface 20 between the first wafer 10A and the second wafer 10B and allowed to penetrate from the outer periphery.

[0029] (Confirmation process) As described above, once the modified layer formation process and the chamfer removal acceleration process have been carried out, a confirmation process is performed to confirm whether the infiltration of the fluid L supplied by the chamfer removal acceleration process is appropriate or inappropriate, that is, whether the fluid L has properly infiltrated to the desired area from which the chamfer portion 17A of the first wafer 10A can be removed. In this embodiment, the chamfer portion 17A of the first wafer 10A is removed starting from the modified layer 100 formed by the modified layer formation process described above. The area from which the chamfer portion 17A can be removed is, as shown in Figure 3(b), the area from the outer edge of the bonded wafer W to the vicinity of the position where the modified layer 100 is formed, and is the area from which a delamination layer 21, whose bonding force is weakened by the action of the infiltrated water, should be formed.

[0030] In order to carry out the verification process, in this embodiment, as shown in Figure 4, the system includes a light emitter 32 that emits infrared wavelength light that is transparent to the first wafer 10A and the second wafer 10B constituting the bonded wafer W and absorbed by water, a light receiver 34 that receives the infrared light emitted by the light emitter 32, and a control means 30 that controls the operation of the light emitter 32 and receives and stores the value of the amount of infrared light received by the light receiver 34.

[0031] In this embodiment, when performing the verification step, the light emitter 32 and light receiver 34 described above are positioned above and below the outer peripheral region of the bonded wafer W held by the holding means 41, more specifically, the region from the outer peripheral region of the bonded wafer W where the modified layer 100 is formed to the outer edge of the bonded wafer W. The chuck table 42 is rotated in the direction indicated by arrow R1 in Figure 4(a) while measuring the rotation angle, and the light emitter 32 is activated to irradiate infrared light so as to include the region from the outer edge of the bonded wafer W to the position where the modified layer 100 is formed. Then, the infrared light transmitted through the bonded wafer W is received by the photodetector 34, and the intensity of the infrared light received by the photodetector 34 (amount of light received P), along with information on the rotation angle (0°-360°) of the chuck table 42, is transmitted to the control means 30, and the amount of light received P corresponding to the rotation angle of the chuck table 42, as shown in Figure 4(b), is stored in an appropriate memory.

[0032] Here, the light emitted by operating the light-emitting element 32 is infrared radiation that is penetrating to the bonded wafer W and absorbent to water. Therefore, if the fluid L (water) properly penetrates the bonding surface 20 in the region of the bonded wafer W irradiated with infrared radiation and a delamination layer 21 is formed, the infrared radiation irradiated from the light-emitting element 32 is absorbed by the water that has penetrated the bonding surface 20, and the amount of infrared radiation received by the photodetector 34 that has passed through the bonded wafer W decreases. On the other hand, if the fluid L (water) does not properly penetrate and the delamination layer 21 is not formed, the infrared radiation irradiated from the light-emitting element 32 penetrates the bonded wafer W, and the amount of infrared radiation received by the photodetector 34 does not decrease and remains relatively large. Therefore, as shown in Figure 4(b), a reference light reception amount P0 (shown by a dashed line) obtained from experiments conducted in advance is set, and in regions where the light reception amount P received by the photodetector 34 is higher than the reference light reception amount P0, it is determined that the intrusion of water into the region where the chamfered portion 17A of the bonded wafer W is removed is "inappropriate", and in regions where the light reception amount P received by the photodetector 34 is lower than the reference light reception amount P0, it is determined that the intrusion of water into the region where the chamfered portion 17A of the bonded wafer W is removed is "appropriate".

[0033] In this embodiment, the wavelength of the infrared radiation emitted from the light emitter 32 is set in the range of 800 nm to 10000 nm. Preferably, the infrared radiation is set to around 1450 nm, around 1940 nm, or around 2900 nm. This is because, referring to the absorption spectrum of water at the above-mentioned infrared wavelengths, absorption peaks appear when the wavelength is 1450 nm, 1940 nm, or 2900 nm. Therefore, when infrared radiation with a wavelength of around 1450 nm, around 1940 nm, or around 2900 nm is selected and emitted from the light emitter 32, even when using low-power infrared radiation, the amount of light received by the photodetector 34 becomes distinct, making it possible to clearly determine whether the penetration of water into the bonding surface 20 is appropriate or inappropriate.

[0034] In the embodiment described above, the diameter of the chuck table 42 is smaller than that of the bonded wafer W, and the light-emitting element 32 is positioned above and the light-receiving element 34 below on the outer circumference where the chamfered portions 17A and 17B of the bonded wafer W are formed, sandwiching the bonded wafer W. However, the top and bottom of the light-emitting element 32 and the light-receiving element 34 may be reversed. Furthermore, if the bonded wafer W is held by a chuck table with a diameter larger than the diameter of the bonded wafer W instead of the chuck table 42 in the above-described embodiment, the light emitter 32 and the light receiver 34 may be arranged above the back surface 10Ab of the first wafer 10A constituting the bonded wafer W. The infrared light from the light emitter 32 is irradiated from above to the area from the outer circumference of the bonded wafer W where the modified layer 100 is formed to the outer edge of the bonded wafer W. The amount of infrared light that passes through the bonded wafer W and is reflected by the chuck table 42 is received by the light receiver 34, which is arranged above the bonded wafer W in close proximity to the light emitter 32. Even with such a configuration, the same effects as those achieved by the light emitter 32 and light receiver 34 described above can be obtained.

[0035] If, in the verification process described above, it is determined that the water infiltration into the area where the chamfered portion 17A can be removed is appropriate, the chamfered portion removal process described later is performed. Furthermore, if, despite performing the chamfered portion removal acceleration process and supplying water to the bonding surface 20 of the bonded wafer W from the outer circumference, it is determined that the water infiltration into the area where the chamfered portion 17A can be removed is inappropriate, at least one of the above-described modified layer formation process and chamfered portion removal acceleration process is performed again. More specifically, by performing the above-described modified layer formation process again after it is determined to be inappropriate in the verification process, the infiltration of the already supplied fluid L (water) is accelerated, bringing the state closer to one in which the water infiltration into the area where the chamfered portion 17A can be removed is determined to be appropriate. Also, by performing the chamfered portion removal acceleration process again after it is determined to be inappropriate in the verification process, fluid L is supplied again to the area where the infiltration of fluid L was insufficient, bringing the state in which the water infiltration into the area where the chamfered portion 17A can be removed is determined to be appropriate. Furthermore, if the above-mentioned modification layer formation process and chamfer removal acceleration process are performed after the above-mentioned verification process has determined to be unsuitable, it becomes possible to more reliably ensure that water penetrates the area from the outer circumference of the bonded wafer W to where the chamfer 17A can be removed. If the rotation angle information of the chuck table 42 indicates the angular position at which the penetration of water into the area where the chamfer 17A can be removed is determined to be unsuitable, the fluid L (water) may be concentrated and supplied to that angular position. If the above-mentioned verification process has determined that the penetration of water into the area where the chamfer 17A can be removed is unsuitable, and the modification layer formation process and chamfer removal process are performed again, it is preferable to perform the verification process again to determine whether the penetration of water into the area where the chamfer 17A can be removed is appropriate or unsuitable. If the bonded wafer W has been determined to be "unsuitable" multiple times in the verification process, it may be a defective product, so the above-mentioned processing method may not be continued, and the control means 30 may report this to the operator.

[0036] (Chamfering removal process, grinding process) After determining that the intrusion of water into the area from which the chamfered portion 17A can be removed is appropriate based on the verification process described above, a chamfered portion removal process is performed to remove the chamfered portion 17A of the first wafer 10A that constitutes the bonded wafer W. This chamfered portion removal process can be performed by means of applying an external force to the chamfered portion 17A of the first wafer 10A. For example, by applying an external force by activating an air supply means (not shown) and spraying air from the side of the bonded wafer W toward the delamination layer 21, the chamfered portion 17A can be removed starting from the modified layer 100. Alternatively, the chamfered portion 17A can be removed by inserting a wedge-shaped member (not shown) from the side into the area where the delamination layer 21 is formed and applying an external force. Furthermore, when performing the grinding process described below, it is possible to apply external force to the chamfered portion 17A mentioned above when thinning the back surface 10Ab of the first wafer 10A by grinding, and the grinding process can also serve as a chamfered portion removal process. The grinding process that also serves as a chamfered portion removal process will be described below.

[0037] As described above, the bonded wafer W, which has undergone the modified layer formation process, the chamfer removal promotion process, and the confirmation process, is transported to the grinding apparatus 50 (only a portion is shown) shown in Figure 5. The grinding apparatus 50 comprises at least the chuck table 51 shown in Figure 5 and a grinding means 52. The grinding means 52 is a means for grinding the back surface 10Ab of the first wafer 10A of the bonded wafer W held by suction on the chuck table 51, and comprises a rotating spindle 52a that is rotated by a rotational drive mechanism (not shown), a wheel mount 52b attached to the lower end of the rotating spindle 52a, and a grinding wheel 52c attached to the lower surface of the wheel mount 52b, with a plurality of grinding wheels 52d arranged in an annular pattern on the lower surface of the grinding wheel 52c.

[0038] Once the bonded wafer W is transported to the grinding apparatus 50, the first wafer 10A of the bonded wafer W is placed on the chuck table 51 of the grinding apparatus 50 with the second wafer 10B facing upwards, and the bonded wafer W is held in place by suction by activating a suction means (not shown).

[0039] Next, the rotating spindle 52a of the grinding means 52 is rotated at, for example, 6000 rpm in the direction indicated by arrow R2 in Figure 5, while the chuck table 51 is rotated at, for example, 300 rpm in the direction indicated by arrow R3. Then, grinding water is supplied onto the back surface 10Ab of the first wafer 10A by a grinding water supply means (not shown), and a grinding feed means (not shown) is activated to bring the grinding wheel 52d into contact with the back surface 10Ab of the first wafer 10A, and the grinding wheel 52c is fed downwards at, for example, a grinding feed speed of 1.0 μm / second, indicated by arrow R4. At this time, the thickness of the bonded wafer W is measured using a contact-type or non-contact-type measuring gauge (not shown) while grinding is carried out, and the wafer can be thinned until the desired thickness is reached.

[0040] Although not shown in the diagram, the grinding process described above can be carried out in two steps. For example, the grinding apparatus 50 may be equipped with a grinding means at a predetermined grinding feed rate (e.g., 1.0 μm / sec) including a coarse grinding wheel equipped with a coarse grinding wheel for rough grinding, and a grinding means at a predetermined grinding feed rate (e.g., 0.1 μm / sec) including a finish grinding wheel equipped with a fine grinding wheel for finish grinding. The process can then be carried out in a series of steps: a rough grinding step in which the back surface 10Ab of the first wafer 10A is roughly ground with the coarse grinding wheel and the chamfered portion 17A is removed from the first wafer 10A, and a finish grinding step in which the back surface 10Ab is finish ground with the finish grinding wheel.

[0041] Then, by performing the grinding process described above, as shown in Figure 5, the first wafer 10A constituting the bonded wafer W is thinned, and an external force is applied to the first wafer 10A to remove the chamfered portion 17A as a result of the grinding process performed by the grinding means 52, so that the chamfered portion 17A is removed as fragments 17A' starting from the modified layer 100 described above. If the radial modified layer 102 described above is formed in the modified layer formation process, then in this grinding process, when the chamfered portion 17A is removed from the first wafer 10A, the division of the chamfered portion 17A into multiple fragments 17A' starting from the radial modified layer 102 is promoted, and the removal of the chamfered portion 17A can be achieved effectively.

[0042] According to the embodiment described above, the confirmation step can be used to confirm that the bonding force in the area from which the chamfered portion 17A is removed has been weakened. Therefore, before performing the chamfered portion removal step, the bonding force in the area from which the chamfered portion 17A is removed can be reliably weakened, and the chamfered portion 17A can be easily removed from the first wafer 10A starting from the modified layer 100. Thus, the problem described in (1) above can be resolved. Furthermore, since there is no need to use a cutting blade to remove the chamfered portion 17A from the first wafer 10A, the problem described in (2) above can also be resolved.

[0043] In the embodiments described above, an example was described in which a bonded wafer W is bonded by a siloxane bond between a first wafer 10A and a second wafer 10B. However, the bonded wafer W processed according to the present invention is not limited to being bonded by a siloxane bond. The bonded wafer W processed according to the present invention may be, for example, a bonded wafer in which the first wafer 10A and the second wafer 10B are bonded by a SiCN bond using nitride bonding, a TEOS bond which changes tetraethyl orthosilicate molecules into a solid having Si-O-Si bonds, or a ThOx bond which forms a thermal oxide film by heating the silicon surface in an oxidizing atmosphere. Regardless of the type of bond, the bonding force can be weakened by the fluid L (water) supplied in the chamfer removal acceleration step described above, and the chamfered portion 17A can be easily removed by the wafer processing method of the embodiments described above. Furthermore, the present invention can also be applied to bonded wafers W that are bonded after being subjected to O2 plasma treatment and N2 plasma treatment as pretreatment to the bonding surface 20 of the bonded wafer W. [Explanation of Symbols]

[0044] 10A: First wafer 10Aa: Surface 10Ab: Reverse side 12A: Device 14A: Planned division line 16A: Device area 17A: Chamfered section 17A': Fragment 18A: Outer perimeter surplus area 10B: Second wafer 10Ba: Surface 10Bb: Back side 17B: Chamfered section 20: Joint surface 21: Exfoliation layer 40: Laser processing equipment 41: Holding means 42: Chuck Table 43: Rotation axis 44: Drive motor 45: Laser beam irradiation means 46: Light concentrator 47:Fluid supply means 47a: Nozzle 47b: Tip 50: Grinding equipment 51: Chuck Table 52: Grinding methods 52a: Rotating spindle 52b: Wheel mount 52c: Grinding Wheel 52d: Grinding wheel 100, 102: Modified layer L: Fluid (water) W: Bonded wafer

Claims

1. A wafer processing method comprising processing a bonded wafer formed by joining a first wafer and a second wafer, A modification layer formation step involves positioning the focal point of a laser beam on the inner side adjacent to the chamfered portion formed on the outer circumference of the first wafer and irradiating it to form a ring-shaped modification layer, A chamfering removal step, in which the chamfered portion of the first wafer is removed starting from the modified layer, Equipped with, Prior to the chamfer removal step, a chamfer removal acceleration step is performed in which water is supplied from the outer periphery to the bonding surface of the bonded wafer to allow the water to penetrate to an area where the chamfer can be removed, thereby weakening the bonding force. A wafer processing method comprising a verification step of confirming whether the ingress of water is appropriate or improper.

2. The wafer processing method according to claim 1, wherein the verification step involves irradiating the chamfered area from the side of the first wafer or the second wafer with infrared wavelength light that is transparent to the bonded wafer and absorbent to water, and detecting the amount of transmitted or reflected light to confirm whether the water infiltration is appropriate or inappropriate.

3. The wafer processing method according to claim 2, wherein the infrared wavelength is 800 nm to 10000 nm.

4. The wafer processing method according to claim 3, wherein the infrared wavelength is one of approximately 1450 nm, approximately 1940 nm, or approximately 2900 nm.

5. The wafer processing method according to claim 1, wherein in the confirmation step, if the water infiltration is appropriate, the chamfer removal step is performed, and if the water infiltration is inappropriate, the modified layer formation step or the chamfer removal acceleration step is performed again, or the modified layer formation step and the chamfer removal acceleration step are performed again.

6. The wafer processing method according to claim 1, wherein in the modified layer formation step, a radial modified layer is formed from the ring-shaped modified layer toward the outside.

7. The wafer processing method according to claim 1, wherein the chamfer removal acceleration step is performed before, after, or simultaneously with the modified layer formation step.

8. The wafer processing method according to claim 1, wherein the chamfer removal step is performed simultaneously with a grinding step in which the upper surface of the first wafer is ground to thin it.

9. The first wafer and the second wafer are joined by a siloxane bond, which is a Si-O-Si bond. The wafer processing method according to claim 1, wherein the Si-O-Si bond is changed to a Si-OH-OH-Si bond by water that penetrates during the chamfer removal acceleration step, thereby weakening the bonding strength.

Images