How the chip is manufactured

Abstract

To reliably divide a wafer by suppressing abnormal extension of a crack, and reduce the number of times in which the wafer is turned over.SOLUTION: A chip manufacturing method includes a modified layer forming step of irradiating with a first laser beam along a planned division line to form a modified layer and a crack in a state in which the back side of a substrate in the wafer having the substrate and a laminate is exposed, and a focal point of the first laser beam having a wavelength that passes through the substrate is positioned inside the substrate from the back side of the substrate, a grinding step of grinding the back side of the substrate exposed in the modified layer forming step to thin the wafer to a predetermined thickness after the modified layer forming step, a laser-processed groove forming step of forming a laser-processed groove in the laminate by irradiating with a second laser beam having a wavelength that is absorbed by the substrate from the front surface side of the wafer along the planned division line after the grinding step.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a method for manufacturing chips by dividing a wafer having a substrate and a laminate thereon. [Background technology]

[0002] Semiconductor chips are usually manufactured by dividing a wafer having a substrate made of a semiconductor such as silicon and a laminate of a low-k film, a conductive layer, etc., formed on the surface of the substrate, along streets (planned division lines) set on the surface of the substrate. Note that TEGs (Test Element Groups) may be formed in some areas of the laminate on the streets.

[0003] To divide the wafer along the streets, for example, a pulsed laser beam having a wavelength that is absorbed by the substrate and the laminate is first irradiated along the streets to form laser grooves along the streets, in which the laminate is partially removed to expose the surface of the substrate along the streets.

[0004] After forming the laser processed grooves, the focal point of a pulsed laser beam having a wavelength that transmits through the substrate is positioned inside the substrate, and the focal point and the wafer are moved relatively along the street to form a modified layer with reduced mechanical strength inside the substrate.

[0005] At this time, cracks are also formed that start from the modified layer and extend toward the front surface of the substrate. After the modified layer is formed, an external force is applied to the wafer by grinding or the like to further extend the cracks, and the wafer is divided into multiple semiconductor chips (see, for example, Patent Document 1).

[0006] However, when forming a laser groove in a laminate, the surface side of the substrate is subjected to the thermal effect of the laser beam, which may change the crystal orientation of a portion of the substrate near the bottom of the groove.

[0007] The direction in which the cracks that extend from the modified layer depend on the crystal orientation of the substrate, so if the crystal orientation changes, the cracks will appear on the surface of the substrate, avoiding the laser-processed grooves, and the wafer will not be separated along the streets (i.e., processing defects will occur).

[0008] To address this issue, a processing method can be considered in which after forming a modified layer, laser grooves are formed along the streets, and then the back side of the substrate is ground to separate the wafer. If this processing method is adopted, it is believed that the abnormal propagation of cracks caused by changes in crystal orientation can be suppressed.

[0009] However, after exposing the back surface of the wafer upward to form the modified layer, the front surface of the wafer must be exposed upward to form the laser-processed grooves, and then the back surface of the wafer must be exposed upward again to grind the substrate. This means that the wafer must be turned over at least twice between the formation of the modified layer and grinding, reducing productivity. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] JP 2007-173475 A Summary of the Invention [Problem to be solved by the invention]

[0011] The present invention has been made in consideration of these problems, and aims to reliably divide a wafer having a laminate on the surface of a substrate by suppressing abnormal extension of cracks when dividing the wafer, while reducing the number of times the wafer needs to be flipped over. [Means for solving the problem]

[0012] According to one aspect of the present invention, there is provided a chip manufacturing method for manufacturing chips by dividing a wafer having devices formed in each area partitioned by a plurality of intersecting planned division lines set on a surface of a substrate and a stack formed at least on the plurality of planned division lines, the method comprising: a modified layer forming step of exposing a back side of the substrate and irradiating a first laser beam having a wavelength that transmits the substrate along the planned division lines with the focal point of the first laser beam being positioned inside the substrate from the back side of the substrate, to form a modified layer along the planned division lines and form cracks extending from the modified layer to the front side of the substrate; a grinding step after the modified layer forming step of grinding the back side of the substrate exposed in the modified layer forming step to thin the wafer to a predetermined thickness; and a laser processing groove forming step after the grinding step of irradiating a second laser beam having a wavelength that is absorbed by the substrate from the front side of the wafer along the planned division lines to form laser processing grooves in the stack.

[0013] In the laser processed groove forming step, the laser processed groove may be formed so as to cover the crack extending from the modified layer to the front surface side of the wafer.

[0014] In the laser groove forming step, the laser groove may be formed so as not to completely divide the laminate, and in the laser groove forming step, the laser groove may be formed so as to completely divide the laminate.

[0015] In the modified layer formation step, the crack extending toward the surface side of the wafer may be formed so as not to divide the laminate, and in the modified layer formation step, the crack extending toward the surface side of the wafer may be formed so as to divide the laminate.

[0016] The method for manufacturing the chip may further include a tape application step of applying an elastic tape to the back side of the substrate after the grinding step and before the laser groove formation step, and an expansion step of expanding the tape applied to the back side of the substrate after the laser groove formation step. Effect of the Invention

[0017] In a chip manufacturing method according to one embodiment of the present invention, a modified layer forming step and a grinding step are performed in that order with the back surface side exposed, and then after the grinding step, a laser groove forming step is performed with the front surface side exposed, thereby limiting the number of times the wafer is flipped from the modified layer forming step to the laser groove forming step to one time.

[0018] In addition, since the laser groove forming step is performed after the modified layer forming step, abnormal crack propagation due to changes in crystal orientation can be suppressed, and the number of times the wafer is turned over can be reduced, and the wafer having the laminate on the substrate surface can be reliably divided along the planned division line. [Brief description of the drawings]

[0019] [Figure 1] FIG. [Diagram 2] FIG. 2 is a flow diagram of a method for manufacturing a chip according to the first embodiment. [Diagram 3] FIG. 4 is a partial cross-sectional side view showing a modified layer forming step. [Figure 4] FIG. 4(A) is an enlarged cross-sectional view of a wafer showing an example of a crack, and FIG. 4(B) is an enlarged cross-sectional view of a wafer showing another example of a crack. [Diagram 5] FIG. [Figure 6] FIG. 6(A) is an enlarged cross-sectional view showing an example of a wafer after a grinding step, and FIG. 6(B) is an enlarged cross-sectional view showing another example of a wafer after a grinding step. [Figure 7] FIG. 1 is a diagram showing an overview of a peeling step. [Figure 8] FIG. 4 is a partial cross-sectional side view showing a laser processing groove forming step. [Figure 9] FIG. 9(A) is an enlarged cross-sectional view of a wafer showing a laser-processed groove, and FIG. 9(B) is an enlarged cross-sectional view of a wafer showing a laser-processed groove having a sufficiently wide width. [Figure 10] FIG. 10(A) is an enlarged cross-sectional view of a wafer showing two laser-processed grooves, and FIG. 10(B) is an enlarged cross-sectional view of a wafer showing a wide laser-processed groove. [Figure 11] FIG. 11(A) is an enlarged cross-sectional view of the wafer showing a laser-processed groove in the first stage, and FIG. 11(B) is an enlarged cross-sectional view of the wafer showing a laser-processed groove in the second stage. [Figure 12] FIG. 11 is a flow diagram of a method for manufacturing a chip according to a second embodiment. [Figure 13] FIG. 1 is an enlarged cross-sectional view of a wafer showing laser machined grooves that do not reach the surface of the substrate. [Figure 14] FIG. 14(A) is a partial cross-sectional side view of the expansion device, and FIG. 14(B) is a partial cross-sectional side view of the expansion device after the expansion step. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] An embodiment according to one aspect of the present invention will be described with reference to the accompanying drawings. Fig. 1 is a perspective view of a wafer 11 to be processed. The wafer 11 has a substantially disk-shaped substrate 13. The substrate 13 is made of a semiconductor material such as silicon (Si).

[0021] There are no limitations on the material, shape, structure, size, etc. of the substrate 13. For example, the substrate 13 may be formed of other semiconductor materials (GaAs, InP, GaN, etc.), sapphire, glass, ceramics, resin, composite oxides (LiNbO3, LiTaO3), etc.

[0022] In this example, the diameter of the substrate 13 is approximately 300 mm (12 inches), and the thickness of the substrate 13 from the front surface 13a to the back surface 13b is approximately 700 μm, but the diameter and thickness of the substrate 13 are not limited to this example.

[0023] A plurality of planned division lines (streets) 17 are set in a lattice pattern on the surface 13a of the substrate 13, and devices 19 such as ICs (Integrated Circuits) are formed in each of the areas partitioned by the intersecting planned division lines 17. The external shape of the device 19 in this example is a rectangle of 4 mm x 5 mm, but there are no limitations on the type, quantity, shape, structure, size, etc. of the device 19.

[0024] A laminate 15 including a low dielectric constant insulator layer (a so-called insulating layer made of a low-k material) and a conductive layer made of a metal is formed on a surface 13a of the substrate 13. The thickness of the laminate 15 is, for example, about several μm, which is sufficiently thinner than the thickness of the substrate 13 before processing.

[0025] The laminate 15 is formed over the entire surface 13a, including the planned division lines 17. In the laminate 15, the region in which the devices 19 are formed protrudes compared to the region in which the planned division lines 17 are set.

[0026] Next, a method for manufacturing a plurality of device chips (chips) 35 (see FIG. 9(A) and the like) by dividing the wafer 11 into units of devices 19 will be described. FIG. 2 is a flow diagram of the method for manufacturing the device chips 35 according to the first embodiment.

[0027] First, a first protective tape 21 is attached to the laminate 15 side of the wafer 11 (i.e., the surface 11a of the wafer 11) and to one side of a metal annular frame 23 (see Figure 3) having an opening with a larger diameter than the wafer 11 (tape attachment step S10).

[0028] In this example, a wafer unit 25 is formed in which the wafer 11 is supported by the annular frame 23 via the first protective tape 21, but the annular frame 23 is not required, and the first protective tape 21, which has approximately the same diameter as the wafer 11, may be attached to the surface 11a side.

[0029] The first protective tape 21 is, for example, a resin film having a glue layer (adhesive layer) provided on a base layer made of resin. In this example, the base layer is made of polyolefin resin, and the adhesive layer is made of ultraviolet-curable resin.

[0030] After the tape attachment step S10, the first laser processing device 2 is used to form a modified layer 13c on the substrate 13 (modified layer forming step S20). In the modified layer forming step S20, the first laser processing device 2 shown in FIG. 3 is used.

[0031] The first laser processing apparatus 2 includes a disk-shaped chuck table 4 having a holding surface 4a that suction-holds the front surface 11a side of the wafer 11. The chuck table 4 includes a disk-shaped frame and a disk-shaped porous plate (not shown) having a smaller diameter than the frame.

[0032] A negative pressure from a suction source such as an ejector acts on the porous plate through a flow path (not shown) formed in the frame. The chuck table 4 is configured to be rotatable around a predetermined rotation axis. The rotation of the chuck table 4 is performed by a rotation drive source (not shown) that rotates a base member on which the chuck table 4 is placed.

[0033] The chuck table 4 is configured to be movable along a processing feed direction 6 and an indexing feed direction that is perpendicular to the processing feed direction 6 in a horizontal plane. Movement in the processing feed direction 6 is achieved by a ball screw type X-axis movement mechanism, and movement in the indexing feed direction is achieved by a ball screw type Y-axis movement mechanism.

[0034] A plurality of clamp mechanisms 8 are provided on the sides of the chuck table 4 along the circumferential direction of the chuck table 4. A head portion 12 of a laser beam irradiation unit 10 is provided above the holding surface 4a.

[0035] The laser beam irradiation unit 10 includes a laser oscillator (not shown) and a head unit 12 equipped with a condenser lens (not shown). A pulsed laser beam (first laser beam) 14 having a wavelength that transmits the substrate 13 is irradiated from the head unit 12 toward the holding surface 4a.

[0036] In the modified layer forming step S20, first, the front surface 11a side of the wafer 11 is suction-held by the holding surface 4a, and the annular frame 23 is clamped by each clamp mechanism 8. At this time, the back surface 11b of the wafer 11 (i.e., the back surface 13b of the substrate 13) is exposed upward.

[0037] Then, the chuck table 4 is rotated appropriately to correct the deviation (so-called θ deviation) of the intended division line 17 with respect to the processing feed direction 6. Next, with the focusing point 14a of the laser beam 14 positioned at a predetermined depth inside the substrate 13 from the rear surface 13b side of the substrate 13, the focusing point 14a and the chuck table 4 are moved relatively along the processing feed direction 6.

[0038] In the modified layer forming step S20, the wafer 11 is processed under the following processing conditions, for example. Laser beam wavelength: 1064nm Average power output: 1W Pulse repetition frequency: 100kHz Machining feed speed: 800mm / s Distance from focal point to substrate surface: 70μm Number of passes: 1

[0039] The number of passes means the number of times the laser beam 14 is irradiated along one planned division line 17. However, the number of passes may be two or more, and two or more modified layers 13c may be formed at different depth positions on the front surface 11a side. For example, it is preferable to set the number of passes to three, and form three modified layers 13c at different depth positions on the front surface 11a side.

[0040] Furthermore, the distance from the focusing point 14a to the surface 13a of the substrate 13 (i.e., the distance from the modified layer 13c to the surface 13a) is appropriately adjusted so as to be greater than the finished thickness of the substrate 13. For example, when the finished thickness of the substrate 13 is 50 μm, the distance from the modified layer 13c to the surface 13a of the substrate 13 is adjusted to 70 μm, etc.

[0041] After forming a modified layer 13c along one of the planned dividing lines 17, the chuck table 4 is indexed and fed a predetermined length in the indexing feed direction to similarly form modified layers 13c along other planned dividing lines 17 adjacent to the one of the planned dividing lines 17.

[0042] After forming modified layers 13c along all of the planned division lines 17 along one direction, the chuck table 4 is rotated 90 degrees by the rotary drive source, and modified layers 13c are similarly formed along all of the planned division lines 17 along the other direction perpendicular to the one direction.

[0043] The modified layer 13c means a region having a lower mechanical strength than a region where the modified layer 13c is not formed in the substrate 13. In the modified layer forming step S20, when the modified layer 13c is formed, cracks 13d extend from the modified layer 13c toward the front surface 13a side and the back surface 13b side.

[0044] 4(A) is an enlarged cross-sectional view of the wafer 11 showing an example of a crack 13d formed by performing the modified layer forming step S20 under the above-mentioned processing conditions. In the example shown in FIG. 4(A), the crack 13d reaches the surface 13a of the substrate 13 but does not divide the laminate 15.

[0045] In this way, when the crack 13d does not divide the laminate 15, the wafer 11 is not divided into individual device chips 35, and therefore the device chips 35 do not move during the tape attachment step S50 described below. The same applies when the crack 13d does not divide the laminate 15 and the extension of the crack 13d stops within the laminate 15.

[0046] If the device chips 35 do not move, the spacing between the device chips 35 will be constant, which has the advantage that the laser processing groove formation step S70 described below will be easier to perform than when the spacing between the device chips 35 becomes variable due to movement of the device chips 35.

[0047] However, by applying one or more of the following, a crack 13d may be formed in the laminate 15 and the laminate 15 may be divided by the crack 13d: increasing at least one of the pulse energy and the repetition frequency, moving the focal point 14a closer to the surface 13a, and decreasing the processing feed rate.

[0048] For example, by forming the modified layer 13c further toward the front surface 11a than in the above example, the laminate 15 can be divided by the crack 13d as shown in Fig. 4(B). This has the advantage that the wafer 11 can be divided more reliably. Fig. 4(B) is an enlarged cross-sectional view of the wafer 11 showing another example of the crack 13d.

[0049] 4(A) and 4(B), the crack 13d does not reach the back surface 11b (i.e., the back surface 13b). After the modified layer forming step S20, the back surface 11b side exposed in the modified layer forming step S20 is ground to apply an external force and thin the wafer 11 to a predetermined finishing thickness (grinding step S30).

[0050] In the grinding step S30, a grinding device 20 shown in Fig. 5 is used. The grinding device 20 has a disk-shaped chuck table 22. The chuck table 22 has a structure similar to that of the above-mentioned chuck table 4, so a duplicated description will be omitted.

[0051] A rotary drive source (not shown) is connected to the lower part of the chuck table 22, and the chuck table 22 rotates at high speed around a predetermined rotation axis 24. Above the holding surface of the chuck table 22, a grinding unit 26 is provided.

[0052] The grinding unit 26 has a cylindrical spindle 28. A motor (not shown) is provided at the upper end of the spindle 28, and a disk-shaped mounter 30 is provided at the lower end of the spindle 28.

[0053] An annular grinding wheel 32 is connected to the underside of the mounter 30. The grinding wheel 32 has an annular metallic wheel base 34. A plurality of grinding stones 36 are fixed to the bottom side of the wheel base 34 along the circumferential direction of the wheel base 34.

[0054] The grinding wheel 36 is formed by mixing abrasive grains such as diamond, cBN (cubic boron nitride), etc. with a binder such as metal, ceramics, resin, etc. Fig. 5 is a perspective view showing the grinding step S30.

[0055] However, before performing the grinding step S30, first, the first protective tape 21 is cut along the contour of the wafer 11 without turning over the wafer unit 25, to form a wafer unit 27 of the wafer 11 and the first protective tape 21. If the annular frame 23 is not used in the tape attachment step S10, it is not necessary to cut off the first protective tape 21.

[0056] Next, in the grinding step S30, the surface 11a side is suction-held by the chuck table 22, the chuck table 22 is rotated at high speed (e.g., 100 rpm), and further, while grinding water such as pure water is supplied to the processing point, the grinding wheel 32 is rotated at a predetermined speed (e.g., 3000 rpm), and the grinding unit 26 is ground and fed downward at a predetermined grinding feed speed (e.g., 0.5 μm / sec).

[0057] As a result, the back surface 13b side is ground, and the substrate 13 is thinned to a predetermined finished thickness (for example, about 50 μm). Fig. 6(A) is an enlarged cross-sectional view showing an example of the wafer 11 after the grinding step S30.

[0058] Due to the external force applied to the wafer 11 in the grinding step S30, the cracks 13d extend to the front surface 13a and the back surface 13b of the substrate 13, and for example, each crack 13d continues from the front surface 13a to the back surface 13b as shown in FIG. 6(A).

[0059] In addition, if cracks 13d have already been formed in stack 15 (see FIG. 4(B)) or if the external force applied to wafer 11 in grinding step S30 is relatively large, cracks 13d may continue from front surface 11a to back surface 11b of wafer 11 (see FIG. 6(B)). FIG. 6(B) is an enlarged cross-sectional view showing another example of wafer 11 after grinding step S30.

[0060] After the grinding step S30, the wafer 11 is transported to a tape replacement device (not shown), where a UV (ultraviolet ray) irradiation step S40 and a tape attachment step S50, which will be described later, are performed.

[0061] The tape replacement device has an ultraviolet ray irradiation unit. The ultraviolet ray irradiation unit has a table made of a transparent material that transmits UV rays. A UV lamp is disposed below the table.

[0062] In the UV irradiation step S40 after the grinding step S30, first, the wafer 11 is transferred onto a table with the front surface 11a of the wafer 11 facing downward. Then, the first protective tape 21 is irradiated with UV light to reduce the adhesive force of the first protective tape 21.

[0063] The tape replacement device includes an application unit that applies second protective tape 29 to wafer 11 or the like, and a cutting unit that cuts out, into a circular shape, second protective tape 29 applied to wafer 11 or the like. The cutting unit includes an arm and a cutting blade provided at the tip of the arm.

[0064] After the UV irradiation step S40, in a state where a metal annular frame 31 having a diameter larger than that of the wafer 11 is arranged on the outer periphery of the wafer 11, a bonding unit bonds a second protective tape 29 to one surface of the annular frame 31 and to the back surface 11b side of the wafer 11 (tape bonding step S50). Next, a cutting unit cuts the second protective tape 29 to a predetermined diameter.

[0065] The tape replacement device further includes an inverting and moving unit that inverts the wafer 11 upside down, and a peeling unit that peels off the first protective tape 21. The wafer 11 with the second protective tape 29 attached to the back surface 11b side is inverted upside down by the inverting and moving unit, and is transported to a holding table of the peeling unit.

[0066] As a result, the first protective tape 21 is exposed upward, and the second protective tape 29 is sucked and held by the holding table. Then, the peeling unit peels off the first protective tape 21, whose adhesive strength has been reduced in the UV irradiation step S40 (peeling step S60).

[0067] In this manner, the wafer 11 is transferred from the first protective tape 21 to the second protective tape 29, and a wafer unit 33 (see FIG. 7) is formed in which the wafer 11 is supported by the annular frame 31 via the second protective tape 29. FIG. 7 is a diagram showing an overview of the peeling step S60.

[0068] Like the first protective tape 21, the second protective tape 29 is a resin film having a glue layer (adhesive layer) provided on a base layer formed of resin, but is a film that is particularly elastic.

[0069] After the peeling step S60, a laser processed groove forming step S70 is performed to form laser processed grooves 15a (see FIG. 8) along each of the intended division lines 17. In the laser processed groove forming step S70, a second laser processing device 40 shown in FIG. 8 is used.

[0070] The second laser processing apparatus 40 has a chuck table 42 similar to the chuck table 4, and a clamping mechanism 44 similar to the clamping mechanism 8. Similarly to the first laser processing apparatus 2, the second laser processing apparatus 40 has a rotation drive source, an X-axis movement mechanism, a Y-axis movement mechanism, etc. A head portion 48 of a laser beam irradiation unit 46 is provided above the holding surface 42a.

[0071] Laser beam projection unit 46 The laser includes a head 48 equipped with a laser oscillator (not shown) and a condenser lens (not shown). A pulsed laser beam (second laser beam) 50 having a wavelength that is absorbed by the substrate 13 and the laminate 15 is irradiated from the head 48 toward the holding surface 42a.

[0072] The laser beam 50 is a laser beam irradiation unit 46 In the plane perpendicular to the optical axis of the focusing lens at , the intensity distribution is close to a Gaussian distribution.

[0073] When performing the laser processing groove forming step S70, first, a coating / cleaning device (not shown) is used to coat the entire surface 11a with a water-soluble resin such as polyvinyl alcohol to form a protective film (not shown).

[0074] Next, the back surface 11b side of the wafer 11 is suction-held by the holding surface 42a of the second laser processing device 40, and the annular frame 31 is attached to each clamping mechanism. 44 At this time, the surface 11a of the wafer 11 is exposed upward.

[0075] Then, the chuck table 4 is rotated to correct the deviation of the intended division line 17 with respect to the processing feed direction 6, and then, with the focusing point 50a of the laser beam 50 positioned on the surface 11a, the focusing point 50a and the chuck table 4 are moved relatively along the processing feed direction 6. Fig. 8 is a partial cross-sectional side view showing the laser groove forming step S70.

[0076] In the laser processing groove forming step S70, for example, the processing conditions are set as follows to process the wafer 11. Laser beam wavelength: 355nm Average output: 2W Pulse repetition frequency: 200kHz Machining feed speed: 400mm / s Number of passes: 1

[0077] Next, a focal point 50a is positioned on the surface 11a so as to overlap with the irradiation area of ​​the laser beam 14 in the modified layer formation step S20 in the thickness direction A1 of the wafer 11 (see Figure 9 (A)), and the laser beam 50 is irradiated along each of the planned division lines 17 to form laser processed grooves 15a.

[0078] 9(A) is an enlarged cross-sectional view of the wafer 11 showing an example of a laser processed groove 15a. The laminate 15 is subjected to ablation processing, and the laminate 15 is formed with laser processed grooves 15a in which the laminate 15 is removed from the surface 11a of the wafer 11 to the surface 13a of the substrate 13 along each of the planned division lines 17 (i.e., the laminate 15 is completely divided).

[0079] After the laser grooves 15a are formed along all of the dividing lines 17, the coating and cleaning device is used again to wash the protective film together with the debris generated by the laser processing with pure water or the like, and then the protective film is dried. In this manner, a plurality of device chips 35 are formed by dividing the wafer 11 along the dividing lines 17.

[0080] In this embodiment, the modified layer forming step S20 and the grinding step S30 are performed with the back surface 11b side exposed, and after the grinding step S30, the laser processing groove forming step S70 is performed with the front surface 11a side exposed, so that the number of times the wafer 11 is flipped from the modified layer forming step S20 to the laser processing groove forming step S70 can be reduced to one.

[0081] In addition, since the laser processing groove forming step S70 is performed after the modified layer forming step S20, it is possible to suppress abnormal extension of the crack 13d caused by a change in the crystal orientation of the substrate 13. Therefore, it is possible to reduce the number of times the wafer 11 is turned over, and the wafer 11 having the laminate 15 on the front surface 13a of the substrate 13 can be reliably divided along the intended division lines 17.

[0082] Incidentally, the crack 13d may meander in the thickness direction A1 of the substrate 13 or in the extension direction A2 of the division line 17. If the crack 13d exposed on the surface 13a exists in a region different from the bottom of the laser processed groove 15a, there is a risk that the laser processed groove 15a and the crack 13d will not be connected.

[0083] Therefore, the width of the laser processed groove 15a (the width in a direction perpendicular to the extension direction A2 of the division line 17 in a plane approximately parallel to the back surface 11b) may be made wider than the meandering width of the crack 13d exposed on the front surface 13a.

[0084] 9B is an enlarged cross-sectional view of the wafer 11 showing a laser-processed groove 15a1 that overlaps the crack 13d in the thickness direction A1 and has a width 15b1 that is sufficiently wider than the meandering width of the crack 13d. The laser-processed groove 15a1 completely divides the laminate 15.

[0085] The laser processed groove 15a shown in FIG. 9(A) can also cover the crack 13d that is exposed on the surface 13a, but the laser processed groove 15a1 shown in FIG. 9(B) can sufficiently cover the crack 13d that is exposed on the surface 13a, so that the laser processed groove 15a1 and the crack 13d can be more reliably connected.

[0086] (Second Modification) In the above example, the laser processed groove 15a is formed by irradiating the laser beam 50 once along one of the intended division lines 17, but the laser beam 50 may be irradiated a plurality of times.

[0087] For example, a laser beam 50 having an intensity distribution close to a Gaussian distribution, similar to that used when forming the laser-processed grooves 15a and 15a1, is split into two using a diffractive optical element or the like, and two focal points 50a are irradiated along one planned division line 17.

[0088] In this manner, two laser processed grooves 15a are formed to completely divide the laminate 15 along each of the intended dividing lines 17. Fig. 10(A) is an enlarged cross-sectional view of the wafer 11 showing the two laser processed grooves 15a formed by the first irradiation of the laser beam 50.

[0089] Next, the laminate 15 remaining between the two laser-processed grooves 15a is irradiated with a laser beam 50 (a so-called top hat beam) that has a substantially uniform intensity distribution in a plane perpendicular to the optical axis of the focusing lens and is wider than the first laser beam 50.

[0090] 10(B) is an enlarged cross-sectional view of the wafer 11 showing a wide laser processed groove 15a2 formed by the second irradiation of the laser beam 50. The second modification has the advantage that the formation of two relatively narrow laser processed grooves 15a can eliminate the problem of the laminate 15 tearing and peeling off in the vicinity of the intended division line 17 (so-called delamination).

[0091] In addition, since the width 15b2 of the laser processed groove 15a2 is wider than the crack 13d, there is an advantage that the laser processed groove 15a2 and the crack 13d can be reliably connected.

[0092] (Third Modification) Incidentally, the laminate 15 does not necessarily have to be completely divided by the first irradiation of the laser beam 50. For example, the first irradiation of the laser beam 50 forms two shallow laser processed grooves 15a3 that do not divide the laminate 15.

[0093] 11A is an enlarged cross-sectional view of the wafer 11 showing two shallow laser grooves 15a3. In the second irradiation of the laser beam 50 following the first irradiation, the above-mentioned top hat beam is irradiated to form the laser grooves 15a2.

[0094] 11(B) is an enlarged cross-sectional view of the wafer 11 showing a wide laser processed groove 15a2 formed by the second irradiation of the laser beam 50. The third modified example also has the advantage that the laser processed groove 15a2 and the crack 13d can be reliably connected.

[0095] In addition, in the third modified example, since thermal damage to the substrate 13 can be reduced when the two shallow laser processed grooves 15a3 are formed, the strength of the device chip 35 can be increased compared to the second modified example.

[0096] In addition to the above-mentioned modified examples, a wide laser groove 15a2 may be formed by arranging multiple focal points 50a of the relatively thin laser beam 50 used to form the laser groove 15a along a direction intersecting the intended division line 17.

[0097] According to the above-mentioned manufacturing method of the device chip 35, the number of times the wafer 11 is turned over from the modified layer formation step S20 to the laser groove formation step S70 can be reduced, and the wafer 11 having the laminate 15 on the surface 13a of the substrate 13 can be reliably divided along the intended division lines 17.

[0098] Second Embodiment Next, a second embodiment will be described. Fig. 12 is a flow diagram of a method for manufacturing a device chip 35 according to a second embodiment. In the laser groove forming step S70 of the second embodiment, as shown in Fig. 13, a laser groove 15a4 having a predetermined depth that does not completely divide the laminate 15 is formed.

[0099] 13 is an enlarged cross-sectional view of the wafer 11 showing a laser-processed groove 15a4 that does not reach the surface 13a of the substrate 13. For example, the laser-processed groove 15a4 can be formed by decreasing the average output and the repetition frequency, increasing the processing feed rate, or changing the height position of the focal point 50a.

[0100] By forming the laser processed groove 15a4 that does not reach the surface 13a, damage to the surface 13a side of the substrate 13 in the laser processed groove forming step S70 can be reduced.

[0101] In the second embodiment, after the laser processed groove forming step S70, an expanding step S80 is performed using an expanding device 52 that expands the second protective tape 29. FIG.

[0102] The expansion device 52 has a cylindrical drum 54. A plurality of rollers 56 are provided on the upper end of the drum 54 along the circumferential direction of the opening of the drum 54. A plurality of (three or more, for example, four) legs 58 are arranged on the outside of the drum 54.

[0103] 14(A) shows two legs 58. An air cylinder (not shown) is provided at the lower end of each of the legs 58 for vertically moving the legs 58. An annular table 60 is disposed at the upper ends of the legs 58.

[0104] The annular table 60 has a rectangular outer shape and has a circular opening through which the drum 54 can pass. A plurality of clamp mechanisms 62 are arranged on the outer periphery of the annular table 60, each clamping the annular frame 31 of the wafer unit 33.

[0105] In the expansion step S80, the upper end of the drum 54 and the upper surface of the annular table 60 are positioned at approximately the same height, and the wafer unit 33 is placed on the drum 54 and the annular table 60. Then, the annular frame 31 is clamped by a plurality of clamp mechanisms 62.

[0106] Next, the air cylinder is operated to pull down the annular table 60 relative to the drum 54, thereby radially expanding the second protective tape 29. Figure 14(B) is a partial cross-sectional side view of the expanding device 52 after the expanding step S80.

[0107] As the second protective tape 29 expands, an external force is also applied to the wafer 11 attached to the second protective tape 29, and the wafer 11 is divided into a plurality of device chips 35 along the cracks 13d and the laser processed grooves 15a4.

[0108] In the second embodiment, by using the expansion device 52, even if the laminate 15 is not completely divided in the laser processing groove forming step S70, the wafer 11 can be divided into a plurality of device chips 35. Of course, in the second embodiment, the number of times the wafer 11 is turned over can be reduced, and the wafer 11 having the laminate 15 can be reliably divided along the intended division lines 17.

[0109] Incidentally, in the second embodiment, the laser processed groove 15a4 does not completely divide the laminate 15, but the laminate 15 may be completely divided. In addition, while a case where the crack 13d does not divide the laminate 15 is shown in FIG. 13, the crack 13d may divide the laminate 15.

[0110] If the wafer 11 has already been divided into device chips 35 before the expanding step S80, the expanding step S80 may be performed to increase the spacing between the device chips 35 to facilitate picking up the device chips 35.

[0111] In addition, the structures, methods, etc. according to the above-mentioned embodiments can be appropriately modified and implemented without departing from the scope of the present invention. For example, even in the first embodiment, the expansion step S80 can be performed after the laser processing groove formation step S70. [Explanation of symbols]

[0112] 2: first laser processing device, 4: chuck table, 4a: holding surface 6: Processing feed direction, 8: Clamping mechanism 11: wafer, 11a: front surface, 11b: back surface 10: laser beam irradiation unit, 12: head part 13: substrate, 13a: front surface, 13b: back surface, 13c: modified layer, 13d: crack 14: laser beam, 14a: focusing point 15: laminate, 15a, 15a1, 15a2, 15a3, 15a4: laser processed grooves 15b1,15b2:Width 17: Division line, 19: Device 20: grinding device, 22: chuck table, 24: rotating shaft 21: first protective tape, 23: annular frame, 25, 27: wafer unit 26: Grinding unit, 28: Spindle, 30: Mounter 29: second protective tape, 31: annular frame, 33: wafer unit 32: Grinding wheel, 34: Wheel base, 36: Grinding stone 35: Device chip 40: Second laser processing device 42: chuck table, 42a: holding surface, 44: clamping mechanism 46: laser beam irradiation unit, 48: head part 50: laser beam, 50a: focusing point 52: Expansion device, 54: Drum, 56: Roller, 58: Leg 60: Annular table, 62: Clamping mechanism, A1: Thickness direction, A2: Extension direction

Claims

1. A method for manufacturing chips by dividing a wafer having devices formed in areas partitioned by a plurality of intersecting division lines set on a surface of a substrate and a stack formed at least on the plurality of division lines, comprising the steps of: a modified layer forming step of exposing the back side of the substrate and positioning a focal point of a first laser beam having a wavelength that transmits the substrate from the back side to inside the substrate, irradiating the first laser beam along the intended division line to form a modified layer along the intended division line and form cracks extending from the modified layer to the front side of the substrate; a grinding step of grinding the back side of the substrate exposed in the modified layer forming step to thin the wafer to a predetermined thickness after the modified layer forming step; a laser groove forming step of irradiating the wafer from the front side along the dividing lines with a second laser beam having a wavelength absorbed by the substrate after the grinding step, thereby forming a laser groove in the laminate; A method for manufacturing a chip, comprising:

2. 2. The method for manufacturing a chip according to claim 1, wherein in the laser groove forming step, the laser groove is formed so as to cover the crack extending from the modified layer to the front surface side of the wafer.

3. 3. The method for manufacturing a chip according to claim 2, wherein in the laser groove forming step, the laser groove is formed so as not to completely divide the laminate.

4. 3. The method for manufacturing a chip according to claim 2, wherein in the laser groove forming step, the laser groove is formed so as to completely divide the laminate.

5. 5. The method for manufacturing a chip according to claim 1, wherein in the modified layer forming step, the crack extending toward the front surface side of the wafer is formed so as not to divide the laminate.

6. 5. The method for manufacturing a chip according to claim 1, wherein in the modified layer forming step, the crack extending toward the front surface side of the wafer is formed so as to divide the laminate.

7. The method for producing the chip comprises the steps of: a tape attachment step of attaching an elastic tape to the back side of the substrate after the grinding step and before the laser processing groove forming step; an expanding step of expanding the tape attached to the back side of the substrate after the laser processing groove forming step; The method for manufacturing a chip according to any one of claims 1 to 6, further comprising:

Images