Semiconductor substrate cutting method

DESCRIPTION

Semiconductor Substrate Cutting Method

Technical Field [0001] The present invention relates to a semiconductor substrate 5 cutting method used for cutting a semiconductor substrate having a front face formed with a functional device in a process of making a semiconductor device and the like.

Background Art [0002] As a conventional technique of this kind, Patent Document 1 10 and Patent Document 2 disclose the following technique. First, an adhesive sheet is attached to the rear face of a. semiconductor wafer by way of a die bonding resin, and a blade cuts the semiconductor wafer while the semiconductor wafer is held on the adhesive sheet, so as to yield semiconductor chips. When picking up the semiconductor chips 15 on the adhesive sheet, the die bonding resin is peeled off together with the individual semiconductor chips. This can bond each semiconductor chip onto a lead frame while saving steps such as the step of applying an adhesive to the rear face of the semiconductor chip.

Patent Document 1: Published Japanese Patent Application No. 2002158276

Patent Document 2: Published Japanese Patent Application No.2000-104040

Disclosure of the Invention

Problem to be Solved by the Invention 25 {0003] However, when cutting the semiconductor wafer held on the adhesive sheet with a blade, it is necessary for techniques such as the FP04 00 one mentioned above to reliably cut the die bonding resin layer existing between the semiconductor wafer and the adhesive sheet without cutting the adhesive sheet. Therefore, care must be taken in particular when cutting the semiconductor wafer with a blade in such a case.

[0004] In view of such circumstances, it is an object of the present invention to provide a semiconductor substrate cutting method which can efficiently cut a semiconductor substrate having a front face formed with a functional device together with a die bonding resin layer.

Means for Solving Problem [0005] For achieving the above-mentioned object, the present invention provides a semiconductor substrate cutting method for cutting a semiconductor substrate having a front face formed with a functional device along a line along which the substrate should be cut, the method comprising the steps of irradiating the semiconductor substrate with laser light while positioning a light-converging point within the semiconductor substrate with a rear face of the semiconductor substrate acting as a laser light incident face, so as to form a modified region, and causing the modified region to form a starting point region for cutting along the line along which the substrate should be cut inside by a predetermined distance from the laser light incident face; and attaching an expandable holding member to a rear face of the semiconductor substrate by way of a die bonding resin layer after forming the starting point region for cutting; and cutting the semiconductor substrate and die bonding resin layer along the line along which the substrate should be cut by expanding the holding member after attaching the holding member. 2 FP04 00 [0006] A semiconductor substrate having a front face formed with a functional device is an object to be processed in this semiconductor substrate cutting method. Such a semiconductor substrate is irradiated with laser light while positioning a light-converging point within the semiconductor substrate with the rear face of the semiconductor substrate acting as a laser light incident face, whereby multiphoton absorption or optical absorption equivalent thereto, for example, is generated, and a starting point region for cutting due to the modified region is formed within the semiconductor substrate along the line along which the substrate should be cut. Here, the rear face of the semiconductor substrate is employed as the laser light incident face, since there is a fear of the functional device restraining laser light from entering when the front face is used as the laser light incident face. When the starting point region for cutting is formed within the semiconductor substrate as such, a fracture can be generated from the starting point region for cutting acting as a start point naturally or with a relatively small force applied thereto, so as to reach the front face and rear face of the semiconductor substrate. Therefore, after the starting point region for cutting is formed, an expandable holding member is attached to the rear face of the semiconductor substrate by way of a die bonding resin layer and then is expanded, whereby cut surfaces of the semiconductor substrate cut along the line along which the substrate should be cut are released from their close contact state as the holding member expands. This also cuts the die bonding resin layer existing between the semiconductor substrate and holding member along the line along which the substrate should be cut. Hence, the semiconductor 3 FP04 00 substrate and die bonding resin layer can be cut along the line along which the substrate should be cut much more efficiently than in the case cut with a blade. Also, since the cut surfaces of the semiconductor substrate cut along the line along which the substrate should be cut are initially in close contact with each other, the cut individual pieces of the semiconductor substrate and the cut individual pieces of the die bonding resin layer have substantially the same outer shape, whereby the die bonding resin can be prevented from protruding from the cut surface of each piece of the semiconductor substrate.

[0007] Here, the starting point region for cutting refers to a region to become a cut start point when the semiconductor substrate is cut. The starting point region for cutting may be formed when a modified region is formed continuously or intermittently, The functional device refers to semiconductor active layers formed by crystal growth, light-receiving devices such as photodiodes, light-emitting devices such as laser diodes, and circuit devices formed as circuits, for example.

[0008] Preferably, the method further comprises the step of grinding file rear face of the semiconductor substrate such that the semiconductor substrate attains a predetermined thickness before forming the starting point region for cutting. When the rear face of the semiconductor substrate is thus ground beforehand such that the semiconductor substrate attains a predetermined thickness, the semiconductor substrate and die bonding resin layer can be cut more accurately along the line along which the substrate should be cut. Here, the grinding encompasses cutting, polishing, chemical etching, etc.

[0009] The modified region may include a molten processed region. 4 FP04 00

When the object to be processed is a semiconductor substrate, a molten processed region may be formed upon irradiation with laser light. Since this molten processed region is an example of the above-mentioned modified region, the semiconductor substrate can be cut easily in this case as well, whereby the semiconductor substrate and die bonding resin layer can be cut efficiently along the line along which the substrate should be cut.

[0010] The modified region may include a molten processed region and a minute void positioned on the opposite side of the molten processed region from the laser light incident face. When the object to be processed is a semiconductor substrate, the molten processed region and minute void may be formed upon irradiation with laser light. Since the molten processed region and minute void constitute an example of the modified region, the semiconductor substrate can easily be cut in this case as well, whereby the semiconductor substrate and die bonding resin layer can efficiently be cut along the line along which the substrate should be cut.

[0011] When forming the starting point region for cutting in the semiconductor substrate cutting method in accordance with the present invention explained in the foregoing, a fracture may be allowed to reach the front face of the semiconductor substrate from the starting point region for cutting acting as a start point, the rear face of the semiconductor from the starting point region for cutting acting as a start point, or the front face and rear face of the semiconductor substrate from the starting point region for cutting acting as a start point.

[0012] Preferably, the method further comprises the step of heating the 5 FP04 00 die bonding resin layer before the step of cutting the semiconductor substrate and die bonding resin layer along the line along which the substrate should be cut by expanding the holding member. When the die bonding resin layer is heated before expanding the holding member, 5 the die bonding resin layer can be cut more accurately and easily along the line along which the substrate should be cut by expanding the holding member.

Effect of the Invention [0013] In the present invention, a semiconductor substrate having a 10 front face formed with a functional device can efficiently be cut together with a die bonding resin layer.

Brief Description of the Drawings [0014] [0001] Fig. 1 is apian view of a semiconductor substrate during laser processing by the laser processing method in accordance with an 15 embodiment of the present invention; [0002] Fig. 2 is a sectional view of the semiconductor substrate taken along the line II-II of Fig. 1; [0003] Fig. 3 is a plan view of the semiconductor substrate after laser processing by the laser processing method in accordance with the 20 embodiment; [0004] Fig. 4 is a sectional view of the semiconductor substrate taken along the line IV-IV of Fig. 3; [0005] Fig, 5 is a sectional view of the semiconductor substrate taken along the line V-V of Fig. 3; 25 [0006] Fig. 6 is a plan view of the semiconductor substrate cut by the laser processing method in accordance with the embodiment; 6 FP04 00 [0007] Fig. 7 is a view showing a photograph of a cut section of a silicon wafer formed with a molten processed region by the laser processing method in accordance with the embodiment; [0008] Fig. 8 is a graph showing relationships between the laser light 5 wavelength and the internal transmittance of a silicon substrate in the laser processing method in accordance with the embodiment; [0009] Fig. 9 is a sectional view of the semiconductor substrate formed with a molten processed region and a minute void by the laser processing method in accordance with the embodiment; 10 [0010] Fig. 10 is a sectional view for explaining a principle by which the molten processed region and minute void are formed by the laser processing method in accordance with the embodiment; [0011] Fig. 11 is a view showing photographs of a cut section of a silicon wafer formed with molten processed regions and minute voids 15 by the laser processing method in accordance with this embodiment; [0012] Fig. 12 is a plan view of a silicon wafer to become an object to be processed in the semiconductor substrate cutting method in accordance with the embodiment; [0013] Fig. 13 is a schematic view for explaining the semiconductor 20 substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where a protective film is attached to the silicon wafer, the silicon wafer is thinned, and the protective film is irradiated with UV rays; [0014] Fig. 14 is a schematic view for explaining the semiconductor 25 substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where the silicon wafer and FP04 00 protective film are secured onto a mounting table, the silicon wafer is irradiated with laser light, and a starting point region for cutting is formed within the silicon wafer; [0015] Fig. 15 is a schematic view for explaining the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where a die bonding resin bearing film is attached to the silicon wafer, the protective film is peeled off from the silicon wafer, and the expansion film is irradiated with UV rays; [0016] Fig. 16 is a schematic view for explaining the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where the expansion film is expanded, semiconductor chips are picked up together with cut pieces of a die bonding resin layer, and the semiconductor chip is joined to a lead frame by way of the die bonding resin layer; [0017] Fig. 17 is a schematic view showing the relationship between the silicon wafer and the starting point region for cutting in the semiconductor substrate cutting method in accordance with the embodiment, in which (a) and (b) illustrate respective states where no fracture is generated from the starting point region for cutting acting as a start point, and a fracture from the starting point region for cutting acting as a start point reaches the front face and rear face of the silicon wafer; [0018] Fig. 18 is a schematic view showing the relationship between the silicon wafer and the starting point region for cutting in the semiconductor substrate cutting method in accordance with die FP04 00 embodiment, in which (a) and (b) illustrate respective states where a fracture from the starting point region for cutting acting as a start point reaches the front face of the silicon wafer, and a fracture from the starting point region for cutting acting as a start point reaches the rear face of the silicon wafer; [0019] Fig. 19 is a schematic view for explaining a specific example of the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where the silicon wafer and protective film are secured onto a mounting table, the silicon wafer is irradiated with laser light, and the starting point region for cutting is formed within the silicon wafer; [0020] Fig. 20 is a schematic view for explaining the specific example of the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where a die bonding resin layer is secured to the silicon wafer, the die bonding resin layer is irradiated with laser light, and the die bonding resin layer is formed with a modified region; [0021] Fig. 21 is a schematic view for explaining the specific example of the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where an expansion film is attached to the die bonding layer by way of an adhesive layer, the protective film is peeled off from the silicon wafer, and the expansion film is expanded; [0022] Fig. 22 is a schematic view for explaining another specific example of the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective FP04 00 states where a die bonding resin bearing film is attached to the silicon wafer, the die bonding resin layer is irradiated with laser light, and the die bonding resin is formed with a modified region; and [0023] Fig. 23 is a view for explaining another specific example of the semiconductor substrate cutting method in accordance with the embodiment, in which (a), (b), and (c) illustrate respective states where the silicon wafer is removed from a mounting table of a laser processing apparatus, the protective film is peeled off from the silicon wafer, and the expansion film is expanded.

Explanations of Numerals [0015] 1.semiconductor substrate; 3.front face ; 5.line along which the semiconductor substrate should be cut; 7.modified region; 8.starting point region for cutting; 11.silicon wafer (semiconductor substrate); 13.molten processed region; 14. minute void; 15.functional device; 17.rear face (laser light incident face); 21. expansion film (holding member); 23.die bonding resin layer; 28. fracture; L. laser light; P. light-converging point.

Best Modes for Carrying Out the Invention [0016] In the following, a preferred embodiment of the semiconductor substrate cutting method in accordance with the present invention will be explained in detail with reference to the drawings. This embodiment utilizes a phenomenon of multiphoton absorption for forming a modified region within a semiconductor substrate. Therefore, a laser processing method for forming a modified region due to multiphoton absorption will be explained at first.

[0017] A material becomes optically transparent if its absorption 10 FP04 00 bandgap Eo is greater than a photon energy hv. Hence, the condition under which absorption occurs in foe material is hv > Eg· However, even when optically transparent, the material yields absorption under the condition of nhv > EG (n = 2, 3,4,.) if the intensity of laser light is very high. This phenomenon is known as multiphoton absorption. In the case of pulse waves, the intensity of laser light is determined by the peak power density (W/cm2) of laser light at a light-converging point thereof. The multiphoton absorption occurs, for example, at a peak power density (W/cm2) of 1 x 10s (W/cm2) or higher. The peak power density is determined by (energy per pulse of laser light at the light-converging point)/(laser light beam spot cross-sectional area x pulse width). In the case of a continuous wave, the intensity of laser light is determined by the electric field strength (W/cm2) of laser light at the light-converging point.

[0018] The laser processing method in accordance with an embodiment of the present invention utilizing such multiphotori absoiption will be explained with reference to Figs. 1 to 6. As shown in Fig. 1, a semiconductor substrate 1 has a front face 3 provided with a line along which the substrate should be cut 5 for cutting the semiconductor substrate 1. The line along which the substrate should be cut 5 is a virtual line extending straight. In the laser processing method in accordance with this embodiment, the semiconductor substrate 1 is irradiated with laser light L while a light-converging point P is positioned within the semiconductor substrate 1 under a condition where multiphoton absorption occurs. The light-converging point P is a location where the laser light L converges. The line along which the 11 FP04 00 substrate should be cut 5 may be either straight or curved, and is not limited to the virtual line, but may be a line actually drawn on the semiconductor substrate 1.

[0019] Then, the laser light L is relatively moved along the line along which the substrate should be cut 5 (i.e., in the direction of arrow A in Fig. 1), so as to move the light-converging point P along the line along which the substrate should be cut 5. This forms a modified region 7 within the semiconductor substrate 1 along the line along which the substrate should be cut 5 as shown in Figs. 3 to 5, and this modified region 7 becomes a starting point region for cutting 8. The laser processing method of this embodiment forms no modified region 7 by causing the semiconductor substrate 1 to absorb the laser light L upon heating the semiconductor substrate 1. Instead, the laser light L is transmitted through the semiconductor substrate 1, so as to generate multiphoton absorption within the semiconductor substrate 1, thereby forming the modified region 7. Hence, the front face 3 of the semiconductor substrate 1 hardly absorbs the laser light L, and thus does not melt.

[0020] When the starting point region for cutting 8 is formed within the semiconductor substrate 1, a fracture is likely to occur from the starting point region for cutting 8 acting as a start point, whereby the semiconductor substrate 1 can be cut as shown in Fig. 6 with a relatively small force. Therefore, the semiconductor substrate 1 can be cut with a high accuracy without generating unnecessary fractures in the front face 3 of the semiconductor substrate 1.

[0021] There seem to be the following two cases in the cutting of the 12 FP04 00 semiconductor substrate 1 from the starting point region for cutting 8 acting as a start point. The first case is where, after forming the starting point region for cutting 8, an artificial force is applied to the semiconductor substrate 1, so that the semiconductor substrate 1 fractures from the starting point region for cutting 8 acting as a start point, whereby the semiconductor substrate 1 is cut. This is the cutting in tiie case where the semiconductor substrate 1 has a large thickness, for example. The application of an artificial force encompasses application of bending stress and shearing stress along the starting point region for cutting 8 of the semiconductor substrate 1, and exertion of a temperature difference upon the semiconductor substrate 1 to generate thermal stress, for example. The other case is where the starting point region for cutting 8 is formed, so that the semiconductor substrate 1 is naturally fractured in a cross-sectional direction (thickness direction) of the semiconductor substrate 1 from the starting point region for cutting 8 acting as a start point, whereby the semiconductor substrate 1 is cut. This is enabled, for example, by forming the starting point region for cutting 8 by a single row of modified regions 7 when the semiconductor substrate 1 has a small thickness, and by a plurality of rows of modified regions 7 aligned in the thickness direction when the semiconductor substrate 1 has a large thickness. Even in the case of natural fracturing, fractures do not extend to the front face 3 at a location not formed with the starting point region for cutting 8 in the part to cut, whereby only the part corresponding to the location formed with the starting point region for cutting 8 can be fractured. Thus, fracturing can be regulated well. Such a fracturing method with favorable controllability is quite effective, 13 FP04 00 since the semiconductor substrate 1 such as a silicon wafer has recently been apt to become thinner.

[0022] The modified region formed by multiphoton absorption in this embodiment includes the following cases (1) and (2): [0023] (1) Case where the modified region is a molten processed region A semiconductor material is irradiated with laser light while a light-converging point is positioned therewithin under a condition with an electric field intensity of at least 1 x 108 (W/cm2) at the light-converging point and a pulse width of 1 ps or less. As a consequence, the inside of the semiconductor substrate is locally heated by multiphoton absorption. This heating forms a molten processed region within the semiconductor substrate. The molten processed region refers to a region once melted and then re-solidified, a region just in a melted state, or a region in the process of re-solidifying from its melted state, and may also be defined as a phase-changed region or a region having changed its crystal structure. The molten processed region may also be regarded as a region in which a certain structure has changed into another structure in monocrystal, amorphous, and polycrystal structures. Namely, it refers to a region in which a monocrystal structure has changed into an amorphous structure, a region in which a monocrystal structure has changed into a polycrystal structure, and a region in which a monocrystal structure has changed into a structure including an amorphous structure and a polycrystal structure, for example. When the semiconductor substrate has a silicon monocrystal structure, the molten processed region is an amorphous silicon structure, for example. The upper limit of electric field 14 FP04 00 10 15 20 25 17 7 intensity is 1 x 10 (W/cm ), for example. The pulse width is preferably 1 ns to 200 ns, for example.

[0024] By an experiment, the inventors have verified that a molten processed region is formed within a silicon wafer which is an example of semiconductor substrate. Conditions for the experiment are as follows: [0025] (A) Semiconductor Substrate: silicon wafer (having a thickness of 350 pm and an outer diameter of 4 inches) (B) Laser

Light source: semiconductor laser pumping Nd:YAG laser

Wavelength: 1064 nm

Laser light spot cross-sectional area: 3.14 x IO'8 cm2 Oscillation mode: Q-switch pulse Repetition frequency: 100 kHz Pulse width: 30 ns

Output: 20 pJ/pulse

Laser light quality: TEM00

Polarization characteristic: linear polarization (C) Light-converging lens Magnification: x50 N. A.: 0.55

Transmittance with respect to laser light wavelength: 60% (D) Moving speed of a mounting table mounting the semiconductor substrate: lOOmm/sec 15 FP04 00 [0026] Fig. 7 is a view showing a photograph of a cut section in a part of a silicon wafer cut by laser processing under the above-mentioned conditions. A molten processed region 13 is formed within a silicon wafer 11. The size of the molten processed region 13 formed under the above-mentioned conditions is about 100 pm in the thickness direction.

[0027] The fact that the molten processed region 13 is formed by multiphoton absorption will now be explained. Fig. 8 is a graph showing relationships between the wavelength of laser light and the transmittance within the silicon substrate. Here, respective reflecting components on the front face side and rear face side of the silicon substrate are eliminated, whereby only the transmittance therewithin is represented. The above-mentioned relationships are shown in the cases where the thickness t of the silicon substrate is 50 pm, 100 pm, 200 pm, 500 pm, and 1000 pm, respectively.

[0028] For example, it is seen that laser light is transmitted through the silicon substrate by at least 80% at 1064 nm, where the wavelength of Nd:YAG laser is located, when the silicon substrate has a thickness of 500 pm or less. Since the silicon wafer 11 shown in Fig. 7 has a thickness of 350 pm, the molten processed region 13 due to multiphoton absorption is formed near the center of the silicon wafer, i.e., at a part separated from the front face by 175 pm. The transmittance in this case is 90% or greater with reference to a silicon wafer having a thickness of 200 pm, whereby the laser light is absorbed within the silicon wafer 11 only slightly and is substantially transmitted therethrough. This means that the molten processed region 13 is not 16 FP04 00 formed by laser light absorption within the silicon wafer 11 (i.e., not formed upon usual heating with laser light), but by multiphoton absorption. The forming of a molten processed region by multiphoton absorption is described, for example, in "Processing Characteristic Evaluation of Silicon by Picosecond Pulse Laser”, Preprints of the National Meeting of Japan Welding Society. No. 66 (April 2000), pp. 72-73.

[0029] Here, a fracture is generated in the cross-sectional direction while using a molten processed region as a start point, whereby the silicon wafer is cut when the fracture reaches the front face and rear face of the silicon wafer. The fracture reaching the front face and rear face of the silicon wafer may grow naturally or grow as a force is applied to the silicon wafer. The fracture may naturally grow from the starting point region for cutting to the front face and rear face of the silicon wafer in any of the cases where the fracture grows from the molten processed region in a melted state and where the fracture grows from the molten processed region in the process of re-solidifying from the melted state. In any of. these cases, the molten processed region is formed only within the silicon wafer. In the cut section after cutting, the molten processed region is formed only therewithin as shown in Fig. 7. When a molten processed region is formed within the semiconductor substrate as such, unnecessary fractures deviating from a line along which the substrate should be cut are hard to occur at the time of fracturing, which makes it easier to control the fracturing.

[0030] (2) Case where the modified region is constituted by a molten processed region and a minute void 17 FP04 00 A semiconductor substrate is irradiated with laser light while a light-converging point is positioned therewithin under a condition with an electric field intensity of at least 1 x 108 (W/cm2) at the light-converging point and a pulse width of 1 ps or less. This may form a molten processed region and a minute void within the semiconductor substrate. When the laser light L is incident on the semiconductor substrate 1 from the front face 3 side as shown in Fig. 9, a minute void 14 is formed on the rear face 17 side of the molten processed region 13. Though the molten processed region 13 and the minute void 14 are formed so as to be separated from each other in Fig. 9, there is a case where the molten processed region 13 and the minute void 14 are formed continuously with each other. Namely, when the molten processed region and minute void are formed as a pair by multiphoton absorption, the minute void is formed on the opposite side of the molten processed region from the laser light incident face in the semiconductor substrate. The upper limit for the electric field intensity is 1 x 1012 (W/cm2), for example. The pulse width is preferably 1 ns to 200 ns, for example.

[0031] A principle by which minute voids 14 are formed so as to correspond to respective molten processed regions 13 when the laser light L is transmitted through the semiconductor substrate 1, so as to generate multiphoton absorption, thereby forming the molten processed regions 13 is not totally clear. Here, two hypotheses presumed by the inventors concerning the principle by which the molten processed regions 13 and minute voids 14 are formed in pairs will be explained. [0032] The following is the first hypothesis presumed by the inventors. 18 FP04 00

When the semiconductor substrate 1 is irradiated with laser light L while its focal point is positioned at a light-converging point P within the semiconductor substrate 1 as shown in Fig. 10, a molten processed region 13 is formed near the light-converging point P. Conventionally, light components (L4 and L5 in Fig. 10) in the center part of the laser light L emitted from a laser light source have been used as the laser light L. This aims at employing the center part of the Gaussian distribution of laser light L. In order to restrain the laser light L from affecting the front face 3 of the semiconductor substrate 1, the inventors have decided to widen the laser light L. As a technique therefor, the laser light L emitted from the laser light source is expanded by a predetermined optical system, so as to widen skirts of the Gaussian distribution, thereby relatively increasing the laser intensity of light components (LI to L3 and L6 to L8 in Fig. 10) in marginal parts of the laser light. When thus expanded laser light L is transmitted through the semiconductor substrate 1, a molten processed region 13 is formed near the light-converging point P as explained above, and a minute void 14 is formed at a part corresponding to the molten processed region 13. Namely, the molten processed region 13 and minute void 14 are formed at respective positions along the optical axis (the dash-single-dot line in Fig. 10) of the laser light. The position where the minute void 14 is formed corresponds to a part where the light components (LI to L3 and L6 to L8 in Fig. 10) in marginal parts of the laser light L are theoretically converged. It seems to be because of the spherical aberration of a lens converging the laser light L that light components (L4 and L5 in Fig. 10) in the center part of the laser light are converged 19 FP04 00 at a position different from positions where the light components (LI to L3 and L6 to L8 in Fig. 10) in marginal parts of the laser light L are converged as such in terms of the thickness direction of die semiconductor substrate 1. The first hypothesis presumed by the inventors lies in that this difference in light-converging positions exerts some influences.

[0033] The following is the second hypothesis presumed by the inventors. The part where the light components (LI to L3 and L6 to L8 in Fig. 10) in marginal parts of the laser light L converge is a theoretical laser light-converging point, and thus has such a high optical intensity that a minute structural change occurs, thereby forming the minute void 14 whose surroundings do not change in terms of die crystal structure, whereas the part formed with the molten processed region 13 is thermally affected so much as to be simply melted and then re-solidified.

[0034] Here, the molten processed region is as stated in (1) mentioned above, whereas the minute void is one whose surroundings do not change in terms of the crystal structure. When the semiconductor substrate has a silicon monocrystal structure, the surroundings of the minute void are likely to keep the silicon monocrystal structure.

[0035] By an experiment, the inventors have verified that molten processed regions and minute voids are formed within a silicon wafer which is an example of the semiconductor substrate. Conditions for the experiment are as follows: [0036] (A) Object to be processed: silicon wafer (having a thickness of 100 pm) 20 FP04 00 (B) Laser

Light source: semiconductor laser pumping Nd:YAG laser

Wavelength: 1064 nm Repetition frequency: 40 kHz Pulse width: 30 nsec Pulse pitch: 7 pm Processing depth: 8 pm Pulse energy: 50 pJ/pulse (C) Light-converging lens N. A.: 0.55 (D) Moving speed of a mounting table mounting the object: 280 mm/sec [0037] Fig. 11 is a view showing photographs of a cut section of a silicon wafer cut by laser processing under the above-mentioned conditions. In Figs. 11, (a) and (b) are photographs showing the same cut section in respective scales different from each other. As depicted, pairs of molten processed regions 13 and minute voids 14, each pair being formed upon irradiation with one pulse of laser light L, are made with a predetermined pitch along the cut section (i.e., along a line along which the substrate should be cut). Each of the molten processed regions 13 of the cut section shown in Fig. 11 has a width of about 13 pm in the thickness direction of the silicon wafer 11 (the vertical direction in the drawing) and a width of about 3 pm in the direction of moving the laser light L (the horizontal direction in the drawing). Each of the minute voids 14 has a width of about 7 pm in the thickness 21 FP04 00 direction of the silicon wafer 11 and a width of about 1.3 pm in the direction of moving the laser light L. The gap between each molten processed region 13 and its corresponding minute void 14 is about 1.2 pm.

[0038] The cases of (1) and (2) are explained in the foregoing as a modified region formed by multiphoton absorption. When a starting point region for cutting is formed as follows in view of the crystal structure of the semiconductor substrate, its cleavage property, and the like, the semiconductor substrate can be cut accurately with a smaller force from the starting point region for cutting acting as a start point. [0039] Namely, in the case of a substrate made of a monocrystal semiconductor having a diamond structure such as silicon, the starting point region for cutting is preferably formed in a direction along the (111) plane (first cleavage plane) or (110) plane (second cleavage plane). In the case of a substrate made of a ΙΠ-V family compound semiconductor having a zinc ore type structure such as GaAs, the starting point region for cutting is preferably formed in a direction along the (110) plane.

[0040] When the substrate is formed with an orientation flat along a direction to be formed with the starting point region for cutting (e.g., in a direction along the (111) plane in the monocrystal silicon substrate) or a direction orthogonal to the direction to be formed with the starting point region for cutting, the starting point region for cutting extending along the direction to be formed with the starting point region for cutting can be formed in the substrate in an easy and accurate manner with reference to the orientation flat. 22 FP04 00 [0041] A preferred embodiment of the semiconductor substrate cutting method in accordance with the present invention will now be explained more specifically. Figs. 13 to 16 are partial sectional views of the silicon wafer taken along the line X1II-XIII of Fig. 12.

[0042] As shown in Fig. 12, on the front face 3 of the silicon wafer (semiconductor substrate) 11 to become an object to be processed, a plurality of functional devices 15 are formed into a matrix pattern in directions parallel and perpendicular to the orientation flat 16. In the following manner, such a silicon wafer 11 is cut into the individual functional devices 15.

[0043] First, as shown in Fig. 13(a), a protective film 18 is attached to the front face 3 side of the silicon wafer 11, so as to cover the functional devices 15. The protective film 18 protects the functional devices 15 and hold the silicon wafer 11. After attaching the protective film 18, as shown in Fig. 13(b), the rear face 17 of the silicon wafer 11 is subjected to surface grinding such that the silicon wafer 11 attains a predetermined thickness, and then is subjected to chemical etching so as to be smoothed. Thus, for example, the silicon wafer 11 having a thickness of 350 pm is thinned to a thickness of 100 pm. After the silicon wafer 11 is thinned, the protective film 18 is irradiated with UV rays. This hardens a UV-curable resin layer which is an adhesive layer of the protective film 18, thereby making the protective film 18 easier to peel off from the silicon wafer 11.

[0044] Subsequently, using a laser processing apparatus, a starting point region for cutting is formed within the silicon wafer 11. Namely, as shown in Fig. 14(a), the protective film 18 is secured by vacuum suction 23 FP04 00 onto a mounting table 19 of the laser processing apparatus such that the rear face 17 of the silicon wafer 11 faces up, and a line along which the substrate should be cut 5 is set like a grid (see dash-double-dot lines in Fig. 12) running between neighboring functional devices 15,15. Then, as shown in Fig. 14(b), the silicon wafer 11 is irradiated with laser light L under the above-mentioned condition generating multiphoton absorption while positioning a light-converging point P within the silicon wafer 11 with the rear face 17 acting as a laser light incident face, and the mounting table 19 is moved such that the light-converging point P is relatively moved along the line along which the substrate should be cut 5. Consequently, as shown in Fig. 14(c), molten processed regions 13 form starting point regions for cutting 8 within the silicon wafer 11 along the line along which the substrate should be cut 5.

[0045] Subsequently, the silicon wafer 11 having the protective film 18 attached thereto is removed from the mounting table 19, and a die bonding resin bearing film 20 (e.g., LE-5000 (product name) by Lintec Corporation) is attached to the rear face 17 of the silicon wafer 11 as shown in Fig. 15(a). The die bonding resin bearing film 20 comprises an expandable expansion film (holding member) 21 having a thickness of about 100 pm. On the expansion film 21, a die bonding resin layer (an adhesive resin layer) 23 functioning as a die bonding adhesive is disposed by way of a UV-curable resin layer having a thickness of several pm. Namely, the expansion film 21 is attached to the rear face 17 of the silicon wafer 11 by way of the die bonding resin layer 23. Film expanding means 30 are attached to marginal parts of the expansion film 21. After attaching the die bonding resin bearing film 24 FP04 00 20, the protective film 18 is peeled off from the front face 3 side of the silicon wafer 11 as shown in Fig. 15(b), and the expansion film 21 is irradiated with UV rays as shown in Fig. 15(c). This hardens a UV-curable resin layer which is an adhesive layer of the expansion film 21, thereby making the die bonding resin layer 23 easier to peel off from the expansion film 21.

[0046] Subsequently, as shown in Fig. 16(a), the film expanding means 21 pull the marginal parts of the expansion film 21 outward, thereby expanding the expansion film 21. Expanding the expansion film 21 generates fractures from the starting point regions for cutting 8 acting as start points, and these fractures reach the front face 3 and rear face 17 of the silicon wafer 11. As a consequence, the silicon wafer 11 is cut accurately along the line along which the substrate should be cut 5, whereby a plurality of semiconductor chips 25 each having one functional device 15 are obtained. Here, as the expansion film 21 expands, opposing cut surfaces 25a, 25a of neighboring semiconductor chips 25, 25 are released from their close contact state. Therefore, simultaneously with the cutting of the silicon wafer 11, the die bonding resin layer 23 closely in contact with the rear face 17 of the silicon wafer 11 is cut along the line along which the substrate should be cut 5. [0047] Then, as shown in Fig. 16(b), the semiconductor chips 25 are successively picked up by a suction collect or the like. Here, the die bonding resin layer 23 is cut into an outer shape equivalent to that of the semiconductor chip 25, whereas the adhesion force between the die bonding resin layer 23 and the expansion film 21 is lowered, whereby the semiconductor chip 25 is picked up while in a state where the cut 25 FP04 00 piece of the die bonding resin layer 23 is in close contact with the rear face thereof. Then, as shown in Fig. 16(c), the semiconductor chip 25 is mounted by way of the die bonding resin layer 23 in close contact with the rear face thereof onto a die pad of a lead frame 27, and is bonded to the latter with the filler upon heating.

[0048] In the method of cutting the silicon wafer 11 in the foregoing, the silicon wafer 11 having the front face 3 formed with the functional devices 15 is employed as an object to be processed, and the silicon wafer 11 is irradiated with the laser light L while positioning the light-converging point P within the silicon wafer 11 with the rear face 17 acting as a laser light incident face. This generates multiphoton absorption within the silicon wafer 11, thereby causing the molten processed region 13 to form the starting point region for cutting 8 within the silicon wafer 11 along the line along which the substrate should be cut 5. Here, the rear face of the semiconductor substrate is employed as the laser light incident face, since there is a fear of the functional device restraining laser light from entering when the front face is used as the laser light incident face. When the starting point region for cutting 8 is formed within the silicon wafer 11 as such, a fracture can be generated from the starting point region for cutting 8 acting as a start point naturally or with a relatively small force applied thereto, so as to reach the front face 3 and rear face 17 of the silicon wafer 11. Therefore, after the starting point region for cutting 8 is formed, the expandable holding member 21 is attached to the rear face 17 of the silicon wafer 11 by way of the die bonding resin layer 23, whereby the cut surfaces 25a, 25a of the semiconductor substrate cut along the line 26 FP04 00 along which the substrate should be cut 5 are released from their close contact state as the expansion film 21 expands. This also cuts the die bonding resin layer 23 existing between the silicon wafer 11 and expansion film 21 along the line along which the substrate should be cut 5. Hence, the silicon wafer 11 and die bonding resin layer 23 can be cut along the line along which the substrate should be cut 5 much more efficiently than in the case cut with a blade.

[0049] Also, since the cut surfaces 25a, 25a of the silicon wafer 11 cut along die line along which the substrate should be cut 5 are initially in close contact with each other, the cut individual pieces of the silicon wafer 11 and the cut individual pieces of the die bonding resin layer 23 have substantially the same outer shape, whereby the die bonding resin can be prevented from protruding from the cut surface 25 of each piece of the silicon wafer 11.

[0050] Further, before forming the starting point region for cutting 8 within the silicon wafer 11, the rear face 17 of the silicon wafer 11 is ground such that the silicon wafer 11 attains a predetermined thickness. When the silicon wafer 11 is thinned to a predetermined thickness as such, the silicon wafer 11 and die bonding resin 23 can be cut more accurately along the line along which the substrate should be cut 5. [0051] The above-mentioned method of cutting the silicon wafer 11 relates to a case where, as shown in Fig. 17(a), no fracture generated from the starting point region for cutting 8 acting as a start point occurs in the silicon wafer 11 until the expansion film 21 is expanded. However, as shown in Fig. 17(b), a fracture 28 may be generated from the starting point region for cutting 8 acting as a start point and caused 27 FP04 00 to reach the front face 3 and rear face 17 of the silicon wafer 11 before expanding the expansion film 21. Examples of the method of generating the fracture 28 include one in which stress applying means such as a knife edge is pressed against the rear face 17 of the silicon 5 wafer 11 along the starting point region for cutting 8, so as to generate a bending stress or shearing stress in the silicon wafer 11 along the starting point region for cutting 8; and one in which a temperature difference is imparted to the silicon wafer 11, so as to generate a thermal stress in the silicon wafer 11 along the starting point region for 10 cutting 8.

[0052] Stressing and cutting the silicon wafer 11 along the starting point region for cutting 8 as such before expanding the expansion film 21 can yield a semiconductor chip 25 which is cut with a very high accuracy. When the expansion film 21 attached to the silicon wafer 11 15 is expanded, the opposing cut surfaces 25a, 25a of the neighboring semiconductor chips 25, 25 are released from their close contact state in this case as well, whereby the die bonding resin layer 23 closely in contact with the rear face 17 of the silicon wafer 11 is cut along the cut surfaces 25a. Therefore, the silicon wafer 11 and die bonding resin 20 layer 23 can be cut along the starting point region for cutting 8 much more efficiently in this cutting method than in the case of cutting with a blade.

[0053] When the silicon wafer 11 is thin, the fracture 28 generated from the starting point region for cutting 8 acting as a start region may reach 25 the front face 3 and rear face 17 of the silicon wafer 11 as shown in Fig. 17(b) even if no stress is generated along the starting point region for 28 FP04 00 cutting 8.

[0054] When the starting point region for cutting 8 due to the molten processed region 13 is formed within.the silicon wafer 11 near the front face 3, and the fracture 28 is allowed to reach the front face 3 as shown in Fig. 18(a), the cutting accuracy can be made very high in the front face (i.e., the surface formed with the functional device) of the semiconductor chip 25 obtained by cutting. When the starting point region for cutting 8 due to the molten processed region 13 is formed within the silicon wafer 11 near the rear face 17, and the fracture 28 is allowed to reach the rear face 17 as shown in Fig. 18(b), on the other hand, the die bonding resin layer 23 can be cut accurately by expanding the expansion film 21.

[0055] The present invention is not limited to the above-mentioned embodiment. For example, though the above-mentioned embodiment relates to a case where the modified region 7 is formed by generating multiphoton absorption within the semiconductor substrate 1, there are cases where the modified region 7 can be formed by generating optical absorption equivalent to multiphoton absorption within the semiconductor substrate 1.

[0056] Though the above-mentioned method of cutting the silicon wafer 11 relates to a case where the molten processed region 13 is formed as a modified region, the molten processed region 13 and minute void 14 may be formed as a modified region. In this case, since the rear face 17 of the silicon wafer 11 is employed as the laser light incident face, the minute void 14 is formed on the opposite side of the molten processed region 13 from the laser light incident face, i.e., the 29 FP04 00 front face 3 side formed with the functional device 15. In cut surfaces, the part on the minute void 14 side tends to attain an accuracy higher than that in the part on the molten processed region 13 side, whereby the yield of the semiconductor chips 25 can further be improved when the 5 minute void 14 is formed on the front face 3 side formed with the functional device 15.

[0057] If the die bonding resin layer 23 is heated before expanding the expansion film 21 of the die bonding resin bearing film 20, the die bonding resin layer 23 can be cut more accurately and easily along the 10 line along which the substrate should be cut 5 simultaneously with the cutting of the silicon wafer 11 when expanding the expansion film 21. This seems to be because the die bonding resin layer 23 changes its physical property to one easy to tear apart upon heating. Specifically, when the die bonding resin layer 23 is heated for 1 to 30 minutes at a 15 temperature of 50°C to 120°C, the die bonding resin layer 23 changes its physical property to one easy to tear apart upon heating. In this regard, the die bonding resin layer 23 is less likely to change its physical property when the temperature is lower than 50°C, whereas there is a fear of the die bonding resin layer 23 softening such as to lose its 20 original shape if the temperature exceeds 120°C.

[0058] As a method of heating the die bonding resin layer 23 as mentioned above, the die bonding resin layer 23 as a whole may be heated, or a part of file die bonding resin layer 23 along the line along which the substrate should be cut 5 may selectively be heated. For 25 heating the die bonding resin layer 23 as a whole, the silicon wafer 11 and the die bonding resin bearing film 20 attached to the rear face 17 of 30 FP04 00 the silicon wafer 11 may be blown by warm air, put into a heating furnace, or mounted on a heating table in which a heater is embedded. For selectively heating a part of the die bonding resin layer 23 along the line along which the substrate should be cut 5, it will be sufficient if the line along which the substrate should be cut 5 is irradiated with laser light to which the die bonding resin layer 23 exhibits optical absorption, etc.

[0059] The die bonding resin layer 23 may be heated at any time from when the expansion film 21 is attached to the rear face 17 of the silicon wafer 11 by way of the die bonding resin layer 23 until the silicon wafer 11 and die bonding resin layer 23 are cut along the line along which the substrate should be cut 5 by expanding the expansion film 21. Before attaching the expansion film 21 to the rear face 17 of the silicon wafer 11 by way of the die bonding resin layer 23, the die bonding resin layer 23 may be heated while in the state of the die bonding resin bearing film 20, and then the expansion film 21 may be attached to the silicon wafer 11 by way of thus heated die bonding resin layer 23. In this case, the expansion film 21 may be attached to the silicon wafer 11 by way of the heated die bonding resin layer 23 immediately after heating the die bonding resin layer 23, or after a predetermined time from the heating of the die bonding resin layer 23. One of reasons why heating makes the die bonding resin layer 23 easier to divide as such seems to lie in that it reduces fracture elongation and increases tensile strength. Also, there are cases where the die bonding resin layer 23 can change its physical property to one easy to tear apart when irradiated with electromagnetic waves such as UV rays. 31 FP04 00 [0060] Here, specific examples of selectively heating the part of die bonding resin layer 23 along the line along which the substrate should be cut 5 will be explained. Among the drawings, parts identical or equivalent to each other will be referred to with numerals identical to 5 each other without repeating their overlapping descriptions.

[0061] First, as shown in Fig. 19(a), a protective film 18 is attached to the front face 3 side of the silicon wafer 11, so as to cover the functional devices 15, and then is secured by vacuum suction onto the mounting table 19 of the laser processing apparatus such that the rear face 17 of 10 the silicon wafer 11 faces up. After a line along which the substrate should be cut 5 is set like a grid running between neighboring functional devices 15,15, the silicon wafer 11 is irradiated with laser light L under the condition generating multiphoton absorption while positioning a light-converging point P within the silicon wafer 11 with the rear face 15 17 acting as a laser light incident face as shown in Fig. 19(b), and the mounting table 19 is moved such that the light-converging point P is relatively moved along the line along which the substrate should be cut 5. Consequently, as shown in Fig. 19(c), molten processed regions 13 form starting point regions for cutting 8 within the silicon wafer 11 20 along the line along which the substrate should be cut 5. In place of the protective film 18, a plate-like protective member made of glass or a resin may be attached to the front face 3 side of the silicon wafer 11. [0062] Subsequently, as shown in Fig. 20(a), a die bonding resin layer 23 is secured to the rear face 17 of the silicon wafer 11, and the 25 protective film 18 is secured by vacuum suction onto the mounting table 19 of the laser processing apparatus such that the rear face 17 of the 32 FP04 00 silicon wafer 11 faces up. Then, as shown in Fig. 20(b), the die bonding resin layer 23 is irradiated with laser light L having a predetermined wavelength (e.g., 808 nm) while positioning a light-converging point P therewithin, and the mounting table 19 is moved such that the light-converging point P is relatively moved along the line along which the substrate should be cut 5. Consequently, as shown in Fig. 20(c), a modified region 29 having such a property that it is easy to tear apart is formed in the die bonding resin layer 23 along the line along which the substrate should be cut 5. This modified region 29 is one having a physical property changed or weakened by a heating effect. The die bonding resin layer 23 may be irradiated along the line along which the substrate should be cut 5 with electron beams instead of the laser light L having a predetermined wavelength.

[0063] Subsequently, the silicon wafer 11 is removed from the mounting table 19 and, as shown in Fig. 21(a), an expansion film 21 is attached by way of an adhesive layer (an adhesive whose adhesion force weakens upon irradiation with UV rays and other energy beams) 31 to the die bonding resin layer 23 secured to the silicon wafer 11. The expansion film 21 with the adhesive layer 31 may be attached to the die bonding resin layer 23, or the expansion film 21 may be attached to the die bonding resin layer 23 after the adhesive layer 31 is laminated thereon.

[0064] Then, the protective film 18 is peeled off from the front face 3 side of the silicon wafer 11 as shown in Fig. 21(b), and marginal parts of the expansion film 21 are pulled outward as shown in Fig. 21(c), so as to expand the expansion film 21. As the expansion film 21 is 33 FP04 00 expanded, a fracture occurs in the thickness direction from the starting point region for cutting 8 acting as a start region, and reaches the front face 3 and rear face 17 of the silicon wafer 11. This cuts the silicon wafer 11 accurately along the line along which the substrate should be cut 5, thereby yielding a plurality of semiconductor chips 25 each having one functional device 15. Here, opposing cut surfaces 25a, 25a of neighboring semiconductor chips 25, 25 are released from their close contact state as the expansion film 21 expands, whereby the die bonding resin layer 23 closely in contact with the rear face 17 of the silicon wafer 11 is cut along the line along which the substrate should be cut 5 simultaneously with the cutting of the silicon wafer 11.

[0065] Subsequently, the adhesive layer 31 is irradiated with UV rays or other energy beams, so as to lower its adhesion force, and the semiconductor chips 25 with their corresponding cut pieces of the die bonding resin layer 23 closely in contact therewith are successively picked up.

[0066] Another specific example of selectively heating the part of die bonding resin layer 23 along the line along which the substrate should be cut 5 will now be explained. Among the drawings, parts identical or equivalent to each other will be referred to with numerals identical to each other without repeating their overlapping descriptions.

[0067] First, as in the specific example mentioned above, molten processed regions 13 form starting point regions for cutting 8 within the silicon wafer 11 along a line along which the substrate should be cut 5. Thereafter, as shown in Fig. 22(a), a die bonding resin bearing film 32 is attached to the rear face 17 of the silicon wafer 11, and the protective 34 FPQ4 00 film 18 is secured by vacuum suction onto the mounting table 19 of the laser processing apparatus such that the rear face 17 of the silicon wafer 11 faces up. The die bonding resin bearing film 32 is one in which a die bonding resin layer 23 is disposed by way of an adhesive layer 31 on an expansion film 21 which is made of a material transmitting laser light L having a predetermined wavelength (e.g., 808 nm). As the die bonding resin bearing film 32, one in which the die bonding resin layer 23 is directly disposed on the expansion film 21 made of a material transmitting laser light L having a predetermined wavelength may be used as well (see, for example, Japanese Patent Publication No. 1987034).

[0068] After being attached, the die bonding resin bearing film 32 is irradiated with the laser light L while positioning a light-converging point P within the die bonding resin layer 23 as shown in Fig. 22(b), and the mounting table 19 is moved such that the light-converging point P is relatively moved along the line along which the substrate should be cut 5. Consequently, as shown in Fig. 22(c), a modified region 29 having such a property that it is easy to tear apart is formed in the die bonding resin layer 23 along the line along which the substrate should be cut 5. [0069] Subsequently, the protective film 18 is peeled off from the front face 3 side of the silicon wafer 11 as shown in Figs. 23(a) and (b), and marginal parts of the expansion film 21 are pulled outward as shown in Fig. 23(c), so as to expand the expansion film 21. As the expansion film 21 is expanded, a fracture occurs in the thickness direction from the starting point region for cutting 8 acting as a start region, and reaches the front face 3 and rear face 17 of the silicon wafer 11. This cuts the 35 FP04 00 silicon wafer 11 accurately along the line along which the substrate should be cut 5, thereby yielding a plurality of semiconductor chips 25 each having one functional device 15. Here, opposing cut surfaces 25a, 25a of neighboring semiconductor chips 25, 25 are released from their close contact state as the expansion film 21 expands, whereby the die bonding resin layer 23 closely in contact with the rear face 17 of the silicon wafer 11 is cut along the line along which the substrate should be cut 5 simultaneously with the cutting of the silicon wafer 11.

[0070] Subsequently, the adhesive layer 31 is irradiated with UV rays or other energy beams, so as to lower its adhesion force, and the semiconductor chips 25 with their corresponding cut pieces of the die bonding resin layer 23 closely in contact therewith are successively picked up. The adhesive layer 31 may be irradiated with UV rays or other energy beams either before or after expanding the expansion film 21.

[0071] Though the die bonding resin layer 23 is irradiated with laser light having a predetermined wavelength along the line along which the substrate should be cut 5 in each of the above-mentioned specific examples, a mask formed with a light-transmitting part along the line along which the substrate should be cut 5 may be disposed on the die bonding resin layer 23 or die bonding resin bearing film 32, and totally irradiated with UV rays or other energy beams, so as to form a modified region 29 in the die bonding resin layer 23 along the line along which the substrate should be cut 5.

Industrial Applicability [0072] In the present invention, as explained in the foregoing, a 36 FP04 00 semiconductor substrate having a front face formed with a functional device can efficiently be cut together with a die bonding resin layer. 37