JP6324796B2 - Single crystal substrate processing method - Google Patents

Description

The present invention relates to a single crystal substrate processing method for processing a single crystal substrate such as a sapphire (Al ₂ O ₃ ) substrate, a silicon carbide (SiC) substrate, or a gallium nitride (GaN) substrate.

In the optical device manufacturing process, an optical device layer comprising an n-type nitride semiconductor layer and a p-type nitride semiconductor layer on the surface of a sapphire (Al ₂ O ₃ ) substrate, silicon carbide (SiC) substrate, gallium nitride (GaN) substrate An optical device wafer is formed by forming optical devices such as light-emitting diodes and laser diodes in a plurality of regions partitioned by a plurality of division lines formed in a lattice pattern. Then, each optical device is manufactured by dividing the region where the optical device is formed by irradiating the optical device wafer with a laser beam along the planned division line and cutting it.

  As a method of dividing a wafer such as the above-mentioned optical device wafer, a laser that uses a pulse laser beam having a wavelength that is transparent to the workpiece, and irradiates the pulse laser beam with a focusing point inside the region to be divided. Processing methods have also been attempted. The division method using this laser processing method is to divide the inside of the work piece by irradiating a pulse laser beam having a wavelength that is transparent to the wafer by aligning the condensing point from one side of the wafer. This is a technique for dividing a wafer by continuously forming a modified layer along a line and applying an external force along a street whose strength is reduced by the formation of the modified layer (for example, Patent Documents). 1).

  In addition, as a method of dividing a wafer such as a semiconductor wafer or an optical device wafer along a planned division line, ablation processing is performed by irradiating the wafer with a pulsed laser beam having a wavelength having an absorptivity to the wafer. Thus, a technique of cleaving by forming a laser processing groove and applying an external force along the planned dividing line where the laser processing groove serving as the starting point of breakage has been put into practical use (see, for example, Patent Document 2).

Japanese Patent No. 3408805 JP-A-10-305420

Thus, even in the optical device divided by any of the above processing methods, there is a problem that sludge such as a modified layer or debris remains on the outer peripheral surface to lower the luminance of the optical device.
Single crystal substrates such as sapphire (Al ₂ O ₃ ), silicon carbide (SiC), and gallium nitride (GaN) are difficult-to-cut materials, and the upper surface of the single crystal substrate is polished to a desired thickness. However, there is a problem that it is difficult to form a plurality of recesses on the upper surface of the single crystal substrate in order to improve the luminance of the optical device.

The present invention has been made in view of the above-mentioned facts, and a main technical problem thereof is to provide a method for processing a single crystal substrate that can efficiently form a desired thickness by polishing the upper surface of the single crystal substrate. It is.
Another technical problem is to provide a method for processing a single crystal substrate, which can efficiently form a plurality of recesses on the surface of the single crystal substrate.

In order to solve the main technical problem, according to the present invention, a method for processing a single crystal substrate,
A numerical aperture setting step for setting a numerical aperture (NA) of a condensing lens for condensing a pulsed laser beam to a predetermined value for a single crystal substrate;
The focused point of the pulsed laser beam is positioned at a predetermined position from the upper surface of the single crystal substrate and irradiated with the pulsed laser beam to grow a pore and an amorphous material that shields the pore from the upper surface of the single crystal substrate to form a shield tunnel. Forming a shield tunnel to be formed;
And amorphous removing step of removing the amorphous by polishing the shield tunnel formed in the single crystal substrate by abrasive, only including,
The numerical aperture (NA) of the condenser lens set to a predetermined value in the numerical aperture setting step is such that the value divided by the refractive index (N) of the single crystal substrate is in the range of 0.05 to 0.2. Set to
A method for processing a single crystal substrate is provided.

Abrasives used in the upper Kihi amorphous removal step is less of a single crystal substrate hardness.
The single crystal substrate is either a sapphire (Al2O3) substrate, a silicon carbide (SiC) substrate, or a gallium nitride (GaN) substrate, and the abrasive is sapphire (Al2O3), silicon carbide (SiC), gallium nitride (GaN). It is desirable that the abrasive grain is made of any one of silicate and quartz.
In the shield tunnel forming step, shield tunnels are formed in a continuous manner along the outline dividing the single crystal substrate into chips, and in the amorphous removing step, the outer periphery of the chip is polished.
Further, the shield tunnel forming step forms a shield tunnel connected to the upper surface of the single crystal substrate at a predetermined depth, and the amorphous removal step polishes the upper surface of the single crystal substrate to form a predetermined single crystal substrate. Form to thickness.
Further, the shield tunnel forming step forms shield tunnels scattered at desired positions on the upper surface of the single crystal substrate, and the amorphous removal step polishes the single crystal substrate to form concave portions on the upper surface of the single crystal substrate. To do.

In the method for processing a single crystal substrate according to the present invention, a numerical aperture (NA) of a condensing lens for condensing a pulse laser beam is set to a predetermined value with respect to the single crystal substrate, and the collection of the pulse laser beam is performed. Shield tunnel formation in which a light tunnel is positioned at a predetermined position from the upper surface of a single crystal substrate and a pulse laser beam is irradiated to grow pores and amorphous that shield the pores from the upper surface of the single crystal substrate to form a shield tunnel process and the shield tunnel formed in the single crystal substrate was polished with abrasive saw including a amorphous removing step of removing the amorphous, the light collection is set to a predetermined value in the open talkative setting step The numerical aperture (NA) of the lens is set so that the value divided by the refractive index (N) of the single crystal substrate is in the range of 0.05 to 0.2. Formed Since the amorphous material forming the tunnel is fragile, the polishing material used in the amorphous removal process is polished with abrasive grains made of a material having a hardness lower than that of the single crystal substrate, so that only the amorphous material is removed. It can be easily removed. Therefore, the divided surface of the chip divided along the shield tunnel formed on the single crystal substrate, and the shield tunnel layer formed on the upper surface side of the single crystal substrate in order to form the single crystal substrate with a predetermined thickness And shield tunnels formed scattered on the upper surface side of the single crystal substrate can be efficiently performed.

The perspective view of the optical device wafer as a single crystal substrate processed by the processing method of the single crystal substrate by this invention. The perspective view which shows the state which affixed the optical device wafer shown in FIG. 1 on the dicing tape with which the cyclic | annular flame | frame was mounted | worn. The principal part perspective view of the laser processing apparatus for implementing the shield tunnel formation process in the processing method of the single crystal substrate by this invention. Explanatory drawing which shows 1st Embodiment of the shield tunnel formation process in the processing method of the single crystal substrate by this invention. The figure which shows the relationship between the numerical aperture (NA) of a condensing lens, the refractive index (N) of an optical device wafer, and the value (S = NA / N) which remove | divided the numerical aperture (NA) by the refractive index (N). 1 is a perspective view of a dividing apparatus for dividing an optical device wafer in which a shield tunnel is formed by a method for processing a single crystal substrate according to the present invention into individual optical devices. Explanatory drawing of the wafer division | segmentation process implemented with the division | segmentation apparatus shown in FIG. FIG. 8 is a perspective view of an optical device that is individually divided by the wafer dividing step shown in FIG. 7. Explanatory drawing which shows 1st Embodiment of the amorphous | non-crystalline removal process in the processing method of the single crystal substrate by this invention. The perspective view of the sapphire substrate as a single crystal substrate processed by the processing method of the single crystal substrate by this invention. Explanatory drawing which shows 2nd Embodiment of the shield tunnel formation process in the processing method of the single crystal substrate by this invention. Explanatory drawing which shows 2nd Embodiment of the amorphous | non-crystalline removal process in the processing method of the single crystal substrate by this invention. Explanatory drawing which shows 3rd Embodiment of the shield tunnel formation process in the processing method of the single crystal substrate by this invention.

  Hereinafter, a method for processing a single crystal substrate according to the present invention will be described in more detail with reference to the accompanying drawings.

  FIG. 1 is a perspective view of an optical device wafer as a single crystal substrate processed by the laser processing method according to the present invention. In the optical device wafer 2 shown in FIG. 1, optical devices 21 such as light emitting diodes and laser diodes are formed in a matrix on the surface 2a of a sapphire substrate having a thickness of 300 μm. Each optical device 21 is partitioned by division lines 22 formed in a lattice shape.

A first embodiment of a method for processing a single crystal substrate for processing the above-described optical device wafer 2 as a single crystal substrate will be described with reference to FIGS.
First, a wafer support process is performed in which the optical device wafer 2 is attached to the surface of a dicing tape mounted on an annular frame. That is, as shown in FIG. 2, the back surface 2b of the optical device wafer 2 is attached to the surface of the dicing tape 30 with the outer periphery mounted so as to cover the inner opening of the annular frame 3. Therefore, the surface 2a of the optical device wafer 2 attached to the surface of the dicing tape 30 is on the upper side.

  FIG. 3 shows a laser processing apparatus that performs laser processing along the planned division line 22 of the optical device wafer 2 on which the wafer support step described above is performed. A laser processing apparatus 4 shown in FIG. 3 includes a chuck table 41 that holds a workpiece, laser beam irradiation means 42 that irradiates a workpiece held on the chuck table 41 with a laser beam, and a chuck table 41 that holds the workpiece. An image pickup means 43 for picking up an image of the processed workpiece is provided. The chuck table 41 is configured to suck and hold the workpiece. The chuck table 41 is moved in a processing feed direction indicated by an arrow X in FIG. 3 by a processing feed means (not shown) and is also shown in FIG. 3 by an index feed means (not shown). It can be moved in the index feed direction indicated by the arrow Y.

  The laser beam application means 42 includes a cylindrical casing 421 arranged substantially horizontally. In the casing 421, pulse laser beam oscillation means including a pulse laser beam oscillator and repetition frequency setting means (not shown) are arranged. A condenser 422 equipped with a condenser lens 422a for condensing the pulse laser beam oscillated from the pulse laser beam oscillation means is attached to the tip of the casing 421. The condenser lens 422a of the condenser 422 has a numerical aperture (NA) set as follows. That is, the numerical aperture (NA) of the condenser lens 422a is set to a value obtained by dividing the numerical aperture (NA) by the refractive index (N) of the single crystal substrate in the range of 0.05 to 0.2 (numerical aperture). Setting process). The laser beam irradiation unit 42 includes a condensing point position adjusting unit (not shown) for adjusting the condensing point position of the pulsed laser beam collected by the condensing lens 422a of the condenser 422.

  The imaging means 43 attached to the tip of the casing 421 constituting the laser beam irradiation means 42 is an infrared illumination means for irradiating a workpiece with infrared rays in addition to a normal imaging device (CCD) for imaging with visible light. And an optical system that captures infrared rays irradiated by the infrared illumination means, and an image sensor (infrared CCD) that outputs an electrical signal corresponding to the infrared rays captured by the optical system. It sends to the control means which is not illustrated.

In order to perform laser processing along the planned division line 22 of the optical device wafer 2 on which the wafer support process described above is performed using the laser processing apparatus 4 described above, the condensing point of the pulse laser beam is used as a single crystal substrate. A positioning step of positioning the condenser lens and the single crystal substrate relatively in the optical axis direction so as to be positioned at a desired position in the thickness direction of the optical device wafer 2 is performed.
First, the dicing tape 30 side on which the optical device wafer 2 is adhered is placed on the chuck table 41 of the laser processing apparatus 4 shown in FIG. 3 described above. Then, the optical device wafer 2 is held on the chuck table 41 via the dicing tape 30 by operating a suction means (not shown) (wafer holding process). Therefore, the surface 2a of the optical device wafer 2 held by the chuck table 41 is on the upper side. In FIG. 3, the annular frame 3 to which the dicing tape 30 is attached is omitted, but the annular frame 3 is held by an appropriate frame holding means provided on the chuck table 41. In this way, the chuck table 41 that sucks and holds the optical device wafer 2 is positioned directly below the imaging unit 43 by a processing feed unit (not shown).

  When the chuck table 41 is positioned immediately below the image pickup means 43, an alignment operation for detecting a processing region to be laser processed of the optical device wafer 2 is executed by the image pickup means 43 and a control means (not shown). That is, the image pickup unit 43 and the control unit (not shown) are divided division lines 22 formed in a predetermined direction of the optical device wafer 2, and a condenser 422 of the laser beam irradiation unit 42 that irradiates laser beams along the division division lines 22. Image processing such as pattern matching for alignment with the laser beam is performed to align the laser beam irradiation position (alignment process). In addition, the alignment of the laser beam irradiation position is similarly performed on the division line 22 formed on the optical device wafer 2 in the direction orthogonal to the predetermined direction.

  When the alignment step described above is performed, the chuck table 41 is moved to the laser beam irradiation region where the condenser 422 of the laser beam irradiation means 42 for irradiating the laser beam as shown in FIG. It is positioned directly below the condenser 422. At this time, as shown in FIG. 4A, the optical device wafer 2 is positioned such that one end of the planned dividing line 22 (the left end in FIG. 4A) is located directly below the condenser 422. Then, a condensing (not shown) is performed so that the condensing point P of the pulsed laser beam LB condensed by the condensing lens 422a of the condenser 422 is positioned at a desired position in the thickness direction of the optical device wafer 2 as a single crystal substrate. The point position adjusting means is operated to move the condenser 422 in the optical axis direction (positioning step). In the illustrated embodiment, the condensing point P of the pulse laser beam is a desired position (for example, 5-10 μm from the surface 2a to the back surface 2b side) from the upper surface (front surface 2a side) on which the pulse laser beam is incident on the optical device wafer 2. Position).

  When the positioning step is carried out as described above, the laser beam irradiation means 42 is operated to irradiate the pulse laser beam LB from the collector 422 and near the condensing point P positioned on the optical device wafer 2 (upper surface (surface 2a )) From the bottom surface (rear surface 2b) to form a shield tunnel by forming pores and amorphous shielding the pores. That is, the chuck table 41 is set in the direction indicated by the arrow X1 in FIG. 4A while irradiating the sapphire substrate constituting the optical device wafer 2 from the condenser 422 with a pulse laser beam LB having a wavelength having transparency. It is moved at a feed rate (shield tunnel forming process). Then, as shown in FIG. 4B, when the other end of the planned dividing line 22 (the right end in FIG. 4A) reaches the irradiation position of the condenser 422 of the laser beam irradiation means 42, irradiation with a pulse laser beam is performed. And the movement of the chuck table 41 is stopped.

  By performing the above-described shield tunnel formation process, the optical device wafer 2 is moved from the vicinity of the focal point P of the pulse laser beam LB (upper surface (surface 2a)) to the lower surface (surface 2a) as shown in FIG. The pores 231 and the amorphous 232 formed around the pores 231 grow toward the back surface 2b), and a predetermined interval (in the embodiment shown, an interval of 16 μm (processing) Amorphous shield tunnel 23 is formed at a feed rate of 800 mm / second) / (repetition frequency: 50 kHz)). As shown in FIGS. 4D and 4E, the shield tunnel 23 includes a pore 231 having a diameter of about φ1 μm formed at the center and an amorphous material having a diameter of φ16 μm formed around the pore 231. In the illustrated embodiment, the amorphous 232 adjacent to each other is connected to each other. The amorphous shield tunnel 23 formed in the above-described shield tunnel forming step can be formed from the upper surface (front surface 2a) to the lower surface (back surface 2b) of the optical device wafer 2, and therefore the thickness of the wafer. Even if it is thick, it is sufficient to irradiate the pulse laser beam once, so that the productivity is extremely good. Further, since the debris does not scatter in the shield tunnel forming process, the problem of degrading the quality of the device is also solved.

As described above, when the shield tunnel forming process is performed along the predetermined division line 22, the chuck table 41 is indexed and moved in the direction indicated by the arrow Y by the interval of the division line 22 formed on the optical device wafer 2. (Indexing step), the shield tunnel forming step is performed. If the shield tunnel forming process is performed along all the division lines 22 formed in the predetermined direction in this way, the chuck table 41 is rotated 90 degrees to form the division division formed in the predetermined direction. The shield tunnel forming step is performed along the planned dividing line 22 extending in a direction orthogonal to the line 22.
In the above-described embodiment, the optical device wafer 2 is held on the chuck table 41 with the surface 2a facing upward, and shielded by irradiating a pulsed laser beam along the division line 22 from the surface 2a side of the optical device wafer 2 Although the example of forming the tunnel 23 has been shown, the optical device wafer 2 is held on the chuck table 41 with the back surface of the optical device wafer 2 facing up, and shielded by irradiating a pulse laser beam along the planned division line 22 from the back surface side of the optical device wafer 2 A tunnel 23 may be formed.

In order to form a good shield tunnel 23 in the above-described shield tunnel forming step, as described above, the numerical aperture (NA) of the condenser lens 422a is set to the refractive index (N) of the single crystal substrate. It is important that the value (S) divided by is set in the range of 0.05 to 0.2.
Here, the relationship between the numerical aperture (NA), the refractive index (N), and the value obtained by dividing the numerical aperture (NA) by the refractive index (N) (S = NA / N) will be described with reference to FIG. . In FIG. 5, the pulse laser beam LB incident on the condenser lens 422a is condensed at an angle (α) with respect to the optical axis. At this time, sin α is the numerical aperture (NA) of the condenser lens 422a (NA = sin θ). When the pulsed laser beam LB condensed by the condensing lens 422a is applied to the optical device wafer 2 composed of a single crystal substrate, the single crystal substrate constituting the optical device wafer 2 has a higher density than air, so the pulse laser beam LB is angled. Refraction from (α) to angle (β). At this time, the angle (β) with respect to the optical axis varies depending on the refractive index (N) of the single crystal substrate constituting the optical device wafer 2. Since the refractive index (N) is (N = sin α / sin β), the value (S = NA / N) obtained by dividing the numerical aperture (NA) by the refractive index (N) of the single crystal substrate is sin β. Therefore, it is important to set sin β in the range of 0.05 to 0.2 (0.05 ≦ sin β ≦ 0.2).
The reason why the value (S = NA / N) obtained by dividing the numerical aperture (NA) of the condenser lens 422a by the refractive index (N) of the single crystal substrate is set in the range of 0.05 to 0.2 will be described below. To do.

[Experiment 1-1]
A shield tunnel was formed on a sapphire (Al ₂ O ₃ ) substrate (refractive index: 1.7) having a thickness of 1000 μm under the following processing conditions, and the quality of the shield tunnel was judged.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Processing feed rate: 800 mm / sec

Condenser lens numerical aperture (NA) Shield tunnel good / bad S = NA / N
0.05 None
0.1 Somewhat good 0.058
0.15 Good 0.088
0.2 Good 0.117
0.25 Good 0.147
0.3 Good 0.176
0.35 Slightly good 0.205
0.4 Defect
0.45 Defect: Void is generated 0.5 Defect: Void is formed 0.55 Defect: Void is formed 0.6 Defect: Void is formed

As described above, in the sapphire substrate (refractive index: 1.7), the numerical aperture (NA) of the condenser lens 422a that condenses the pulsed laser beam is equal to the refractive index (N) of the single crystal substrate. By setting the value divided by (S = NA / N) in the range of 0.05 to 0.2, a shield tunnel is formed. Therefore, in the sapphire substrate (refractive index: 1.7), it is important to set the numerical aperture (NA) of the condenser lens 422a that condenses the pulse laser beam to 0.1 to 0.35.

[Experiment 1-2]
A shield tunnel was formed on a silicon carbide (SiC) substrate (refractive index: 2.63) having a thickness of 1000 μm under the following processing conditions, and the quality of the shield tunnel was determined.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Processing feed rate: 800 mm / sec

Condenser lens numerical aperture (NA) Shield tunnel good / bad S = NA / N
0.05 None
0.1 None
0.15 Somewhat good 0.057
0.2 Good 0.076
0.25 Good 0.095
0.3 Good 0.114
0.35 Good 0.133
0.4 Good 0.153
0.45 Good 0.171
0.5 Good 0.19
0.55 Slightly good 0.209
0.6 Defect: A void is formed

As described above, in the silicon carbide (SiC) substrate (refractive index: 2.63), the numerical aperture (NA) of the condenser lens 422a that condenses the pulse laser beam is divided by the refractive index (N) of the single crystal substrate. By setting the value (S = NA / N) in the range of 0.05 to 0.2, a shield tunnel is formed. Therefore, in the silicon carbide (SiC) substrate, it is important to set the numerical aperture (NA) of the condenser lens 422a that condenses the pulse laser beam to 0.15 to 0.55.

[Experiment 1-3]
A shield tunnel was formed on a gallium nitride (GaN) substrate (refractive index: 2.3) having a thickness of 1000 μm under the following processing conditions, and the quality of the shield tunnel was determined.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Processing feed rate: 800 mm / sec

Condenser lens numerical aperture (NA) Shield tunnel good / bad S = NA / N
0.05 None
0.1 Slightly good 0.043
0.15 Good 0.065
0.2 Good 0.086
0.25 Good 0.108
0.3 Good 0.130
0.35 Good 0.152
0.4 Good 0.173
0.45 Good 0.195
0.5 Slightly good 0.217
0.55 Defect: A void is formed 0.6 Defect: A void is formed

As described above, in a gallium nitride (GaN) substrate, a value obtained by dividing the numerical aperture (NA) of the condenser lens 422a that condenses the pulsed laser beam by the refractive index (N) of the single crystal substrate (S = NA / N). Is set in the range of 0.05 to 0.2, a shield tunnel is formed. Therefore, in the gallium nitride (GaN) substrate, it is important to set the numerical aperture (NA) of the condenser lens 422a for condensing the pulse laser beam to 0.1 to 0.5.
Since the shield tunnel is formed on the side irradiated with the laser beam from the condensing point P, the condensing point of the pulse laser beam needs to be positioned on the inner side adjacent to the surface opposite to the side on which the pulse laser beam is incident. There is.

  From the above-described Experiment 1-1, Experiment 1-2, and Experiment 1-3, the value obtained by dividing the numerical aperture (NA) of the condenser lens 422a that condenses the pulse laser beam by the refractive index (N) of the single crystal substrate (S = NA / N) was set in a range of 0.05 to 0.2, it was confirmed that a shield tunnel was formed.

  Next, the correlation between the energy of the pulse laser beam and the length of the shield tunnel is examined.

[Experiment 2]
A sapphire (Al ₂ O ₃ ) substrate, silicon carbide (SiC) substrate, gallium nitride (GaN) substrate with a thickness of 1000 μm is irradiated with a pulsed laser beam under the following processing conditions, and the energy of the pulsed laser beam (μJ / 1 pulse) and shielding The relationship with the tunnel length (μm) was obtained.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Processing feed rate: 800 mm / sec

The average output is increased until the shield tunnel is formed at intervals of 0.05 W (1 μJ / 1 pulse), and after the shield tunnel is formed, the average output is 10 W (200 μJ / pulse) at intervals of 0.5 W (10 μJ / 1 pulse). The average output was increased to 1 pulse), and the length of the shield tunnel (μm) was measured.

Pulse energy (μJ / 1 pulse) Shield tunnel length (μm)
Sapphire Silicon carbide Gallium nitride 1 None None None 2 None None None 3 None None None 4 None None None 5 65 65 70
10 75 85 85
20 125 115 125
30 150 155 170
40 175 185 205
50 190 230 250
60 210 265 295
70 245 290 330
80 260 330 365
90 315 370 415
100 340 395 450
110 365 430 485
120 400 470 530
130 425 500 565
140 455 535 610
150 490 570 650
160 525 610 685
170 550 640 735
180 575 675 770
190 610 715 815
200 640 740 850

To form a shield tunnel from (upper surface (front surface 2a)) to lower surface (back surface 2b) of the optical device wafer 2 made of a sapphire (Al ₂ O ₃ ) substrate having a thickness of 300 μm under the above processing conditions from Experiment 2 The pulse energy of the pulse laser beam may be set to 90 μJ / 1 pulse. In the case of a silicon carbide (SiC) substrate having a thickness of 300 μm, the pulse energy of the pulse laser beam may be set to 80 μJ / 1 pulse, and in the case of a gallium nitride (GaN) substrate having a thickness of 300 μm, the pulse energy of the pulse laser beam may be set. What is necessary is just to set to 70 microJ / 1 pulse.

  Next, we examine the wavelength of the pulsed laser beam and the formation of shield tunnels.

[Experiment 3-1]
A sapphire substrate with a thickness of 1000 μm is reduced to 2940 nm, 1550 nm, 1030 nm, 515 nm, 343 nm, 257 nm, and 151 nm under the following processing conditions to form a sapphire substrate with a band gap of 8.0 eV (wavelength conversion: 155 nm). It was verified whether a shield tunnel could be formed.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Processing feed rate: 800 mm / sec

Wavelength (nm) Shield tunnel good or bad 2940 good 1550 good 1030 good 515 good 343 good 257 bad 151 ablation bad on the incident surface

As described above, in the sapphire substrate, it was confirmed that a shield tunnel was formed when the wavelength of the pulse laser beam was set to be twice or more the wavelength corresponding to the band gap of 8.0 eV (wavelength conversion: 155 nm).

[Experiment 3-2]
With a silicon carbide (SiC) substrate having a thickness of 1000 μm, the wavelength of the pulse laser beam is lowered to 2940 nm, 1550 nm, 1030 nm, 515 nm, and 257 nm under the following processing conditions, and a silicon carbide having a band gap of 2.9 eV (wavelength conversion: 425 nm). We verified whether a shield tunnel could be formed on the (SiC) substrate.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Processing feed rate: 800 mm / sec

Wavelength (nm) Shield tunnel good or bad 2940 Good 1550 Good 1030 Good 515 Ablation failure on the incident surface 257 Ablation failure on the incident surface

As described above, in the silicon carbide (SiC) substrate, it can be confirmed that the shield tunnel is formed when the wavelength of the pulse laser beam is set to more than twice the wavelength corresponding to the band gap of 2.9 eV (wavelength conversion: 425 nm). It was.

[Experiment 3-3]
A gallium nitride (GaN) substrate having a thickness of 1000 μm is subjected to the following processing conditions, and the wavelength of the pulse laser beam is decreased to 2940 nm, 1550 nm, 1030 nm, 515 nm, and 257 nm, and the band gap is 3.4 eV (wavelength conversion: 365 nm) gallium nitride ( It was verified whether a shield tunnel could be formed on the (GaN) substrate.

Processing condition wavelength: 1030 nm pulse laser Repetition frequency: 50 kHz
Pulse width: 10 ps
Average output: 3W
Condensing spot diameter: φ10μm
Processing feed rate: 500 mm / sec

Wavelength (nm) Shield tunnel good / bad 2940 good 1550 good 1030 good 515 bad 257 Ablation bad at the incident surface

As described above, in the gallium nitride (GaN) substrate, it can be confirmed that a shield tunnel is formed when the wavelength of the pulse laser beam is set to more than twice the wavelength corresponding to the band gap of 3.4 eV (wavelength conversion: 365 nm). It was.

  From Experiment 3-1, Experiment 3-2, and Experiment 3-3 described above, it can be confirmed that a shield tunnel is formed when the wavelength of the pulsed laser beam is set to more than twice the wavelength corresponding to the band gap of the single crystal substrate. It was.

The sapphire (Al ₂ O ₃ ) substrate, silicon carbide (SiC) substrate, and gallium nitride (GaN) substrate have been described above, but the present invention relates to a quartz (SiO 2) substrate, a lithium tantalate (LT) substrate, a lithium niobate ( The present invention can also be applied to single crystal substrates such as LN) substrates and langasite (La3Ga5SiO14) substrates.

  When the shield tunnel forming process described above is performed, an external force is applied to the optical device wafer 2 to continuously form the shield tunnel 23 composed of the pores 231 and the amorphous 232 formed around the pores 231. A wafer dividing step of dividing the optical device wafer 2 into individual optical devices 21 along the planned division line 22 is performed. The wafer dividing step is performed using a dividing apparatus 5 shown in FIG. 6 includes a frame holding means 51 for holding the annular frame 3 and a tape extending means for extending the optical device wafer 2 mounted on the annular frame 3 held by the frame holding means 51. 52 and a pickup collet 53. The frame holding means 51 includes an annular frame holding member 511 and a plurality of clamps 512 as fixing means arranged on the outer periphery of the frame holding member 511. The upper surface of the frame holding member 511 forms a mounting surface 511a on which the annular frame 3 is placed, and the annular frame 3 is placed on the mounting surface 511a. The annular frame 3 placed on the placement surface 511 a is fixed to the frame holding member 511 by the clamp 512. The frame holding means 51 configured as described above is supported by the tape extending means 52 so as to be able to advance and retreat in the vertical direction.

  The tape expansion means 52 includes an expansion drum 521 disposed inside the annular frame holding member 511. The expansion drum 521 has an inner diameter and an outer diameter that are smaller than the inner diameter of the annular frame 3 and larger than the outer diameter of the optical device wafer 2 attached to the dicing tape 30 attached to the annular frame 3. The expansion drum 521 includes a support flange 522 at the lower end. The tape expansion means 52 in the illustrated embodiment includes support means 523 that can advance and retract the annular frame holding member 511 in the vertical direction. The support means 523 includes a plurality of air cylinders 523 a disposed on the support flange 522, and the piston rod 523 b is connected to the lower surface of the annular frame holding member 511. As described above, the support means 523 including the plurality of air cylinders 523a is configured such that the mounting position of the annular frame holding member 511 is substantially the same as the upper end of the expansion drum 521 as shown in FIG. Then, as shown in FIG. 7 (b), it is moved in the vertical direction between the extended positions below the upper end of the expansion drum 521 by a predetermined amount.

  A wafer dividing process performed using the dividing apparatus 5 configured as described above will be described with reference to FIG. That is, the annular frame 3 on which the dicing tape 30 to which the optical device wafer 2 is attached is attached to the mounting surface of the frame holding member 511 constituting the frame holding means 51 as shown in FIG. It is placed on 511a and fixed to the frame holding member 511 by the clamp 512 (frame holding step). At this time, the frame holding member 511 is positioned at the reference position shown in FIG. Next, the plurality of air cylinders 523a as the support means 523 constituting the tape expansion means 52 are operated to lower the annular frame holding member 511 to the expansion position shown in FIG. Accordingly, since the annular frame 3 fixed on the mounting surface 511a of the frame holding member 511 is also lowered, the dicing tape 30 attached to the annular frame 3 is an expansion drum as shown in FIG. It expands in contact with the upper edge of 521 (tape expansion process). As a result, since a tensile force acts radially on the optical device wafer 2 adhered to the dicing tape 30, the shield tunnel 23 described above is continuously formed, and along the scheduled division line 22 in which the strength is reduced. Thus, the optical devices 21 are separated from each other, and a space S is formed between the optical devices 21.

Next, as shown in FIG. 7C, the pick-up collet 53 is operated to adsorb the optical device 21, peeled off from the dicing tape 30, picked up, and transported to a tray or die bonding process (not shown). In the pickup process, as described above, the gap S between the individual optical devices 21 attached to the dicing tape 30 is widened, so that the pickup can be easily performed without contacting the adjacent optical devices 21. be able to.
In this manner, the optical device 21 picked up has amorphous 232 remaining on the outer peripheral surface as shown in FIG.

If the above-described pickup process is performed, an amorphous removal process is performed in which the amorphous 232 remaining on the outer peripheral surface of the optical device 21 is polished with an abrasive to remove the amorphous.
In this amorphous removal step, as shown in FIG. 9A, the outer peripheral surface of the optical device 21 is polished by using the sand paper 6, so that the amorphous 232 remaining on the outer peripheral surface of the optical device 21. Remove. As a result, as shown in FIG. 9B, the outer peripheral surface of the optical device 21 is removed from the amorphous surface, and the sapphire (Al ₂ O ₃ ) substrate is exposed. Therefore, the brightness of the optical device 21 can be improved.
The amorphous 232 constituting the shield tunnel 23 formed on the optical device wafer 2 made of a sapphire (Al ₂ O ₃ ) substrate as a single crystal substrate in the shield tunnel forming step is fragile, and thus is amorphous. The polishing material used in the removal step can be easily removed only by removing the amorphous 232 by polishing with abrasive grains made of a material having a hardness lower than that of the single crystal substrate. In the above-described embodiment, since the single crystal substrate constituting the optical device wafer 2 is a sapphire (Al ₂ O ₃ ) substrate, the hardness of sapphire (Al ₂ O ₃ ) as a polishing material (new Mohs hardness No. 1). 12) Use abrasive grains made of the following materials. Therefore, when the shield tunnel 23 is formed on the silicon carbide (SiC) substrate, a material having a silicon carbide (SiC) substrate hardness (new Mohs hardness No. 13) or less as an abrasive, such as silicon carbide (SiC), gallium nitride. Abrasive grains made of (GaN), silicate, and quartz are used.

Next, a second embodiment of the method for processing a single crystal substrate according to the present invention will be described with reference to FIGS.
FIG. 10 shows a sapphire substrate 10 having a thickness of, for example, 300 μm as a single crystal substrate. A processing method for forming the thickness of the sapphire substrate 10 to 150 μm will be described.
In order to form the sapphire substrate 10 having a thickness of 300 μm so as to have a thickness of 150 μm, first, as described above, the numerical aperture (NA) of the condenser lens for condensing the pulse laser beam is set to the sapphire substrate 10 which is a single crystal substrate. A numerical aperture setting step for setting to a predetermined value is performed.
Then, as shown in FIG. 11, the condensing point of the pulsed laser beam is positioned at a desired position from the upper surface of the single crystal substrate and irradiated with the pulsed laser beam. A shield tunnel forming step is performed in which the amorphous to be shielded is grown and the shield tunnel 23 is formed to be connected at a depth of 150 μm from the upper surface of the sapphire substrate 10. This shield tunnel forming step is performed on the entire surface of the sapphire substrate 10 based on the above-described processing conditions using the laser processing apparatus 3 shown in FIG. 3 so that the shield tunnel 23 layer is formed at a depth of 150 μm from the upper surface. It is formed. At this time, in order to form the shield tunnel 23 at a depth of 150 μm from the upper surface of the sapphire substrate 10, the pulse energy of the pulse laser beam is set to 30 μJ / 1 pulse based on the result of Experiment 2 above.

  Next, an amorphous removal step is performed in which the upper surface of the sapphire substrate 10 as a single crystal substrate on which the shield tunnel formation step has been performed is polished to form the sapphire substrate 10 to a predetermined thickness (for example, 150 μm). This amorphous removal step is performed using a polishing apparatus 6 shown in FIG. A polishing apparatus 6 shown in FIG. 12A includes a chuck table 61 that holds a workpiece, and a polishing means 62 that grinds the workpiece held on the chuck table 61. The chuck table 61 is configured to suck and hold a workpiece on the upper surface, and is rotated in a direction indicated by an arrow 61a in FIG. The polishing means 62 includes a spindle housing 621, a rotating spindle 622 that is rotatably supported by the spindle housing 621 and rotated by a rotation driving mechanism (not shown), a mounter 623 attached to the lower end of the rotating spindle 622, and the mounter And a polishing tool 624 attached to the lower surface of 623. The polishing tool 624 includes a circular base 625 and a polishing pad 626 attached to the lower surface of the base 625, and the base 625 is attached to the lower surface of the mounter 623 with fastening bolts 627. . In the illustrated embodiment, the polishing pad 626 is mixed with abrasive grains made of silica as an abrasive in the felt.

In order to perform the amorphous removal process using the polishing apparatus 6 described above, sapphire having the shield tunnel forming process performed on the upper surface (holding surface) of the chuck table 61 as shown in FIG. The surface of the substrate 10 opposite to the surface on which the shield tunnel 23 layer is formed is placed. Then, the sapphire substrate 10 is sucked and held on the chuck table 61 by a suction means (not shown) (wafer holding step). Therefore, the surface of the sapphire substrate 10 held on the chuck table 61 is the upper side where the shield tunnel 23 layer is formed. When the sapphire substrate 10 is sucked and held on the chuck table 61 in this way, the polishing tool of the polishing means 62 is rotated while rotating the chuck table 61 in the direction indicated by the arrow 61a in FIG. 624 is rotated at a predetermined rotational speed in a direction indicated by an arrow 624a in FIG. 12A, and the polishing pad 626 is brought into contact with the upper surface of the sapphire substrate 10 which is a processing surface as shown in FIG. 12B. As shown in FIG. 12A and FIG. 12B, the polishing tool 624 is ground downward (in a direction perpendicular to the holding surface of the chuck table 61) by a predetermined amount as indicated by an arrow 624b in FIGS. To send. As a result, as shown in FIG. 12C, the shield tunnel 23 layer formed on the upper surface side of the sapphire substrate 10 is removed, and the sapphire (Al ₂ O ₃ ) substrate is exposed. In addition, since the amorphous 232 constituting the shield tunnel 23 layer formed on the sapphire substrate 10 as the single crystal substrate in the shield tunnel forming process is fragile as described above, the polishing used in the amorphous removing process is used. By polishing the material with abrasive grains made of silica having a hardness of a single crystal substrate or less, only the shield tunnel 23 layer can be easily removed. In the embodiment described above, since the single crystal substrate is made of a sapphire substrate, abrasive grains made of a material having a sapphire (Al ₂ O ₃ ) hardness (new Mohs hardness No. 12) or less as an abrasive, such as sapphire ( Abrasive grains made of Al ₂ O ₃ ), gallium nitride (GaN), or silicate can be used.

As described above, since the amorphous 232 constituting the shield tunnel 23 layer formed on the sapphire substrate 10 as the single crystal substrate in the shield tunnel forming process is fragile as described above, sapphire ( By polishing with abrasive grains made of silica or the like having a hardness of Al ₂ O ₃ ) or less, only the shield tunnel 23 layer can be easily removed, so that the sapphire substrate 10 as a single crystal substrate can be efficiently determined. Can be formed.

Next, a third embodiment of the method for processing a single crystal substrate according to the present invention will be described with reference to FIG. In the third embodiment, a method of forming recesses on the surface of the sapphire substrate 10 having a thickness of, for example, 300 μm as the single crystal substrate shown in FIG. 10 will be described.
In order to form, for example, 75 μm deep concave portions on the surface of the sapphire substrate 10 having a thickness of 300 μm, first, as described above, the numerical aperture (NA) of the condenser lens for condensing the pulsed laser beam is simply set. A numerical aperture setting step for setting a predetermined value is performed on the sapphire substrate 10 which is a crystal substrate.
Then, as shown in FIG. 13, the pulse laser beam is focused at a desired position from the upper surface of the single crystal substrate and irradiated with the pulse laser beam. A shield tunnel forming step is carried out in which an amorphous material to be shielded is grown and the shield tunnels 23 are scattered and formed at a depth of 75 μm from the upper surface of the sapphire substrate 10. This shield tunnel forming step is performed on the entire surface of the sapphire substrate 10 based on the above-described processing conditions using the laser processing apparatus 3 shown in FIG. 3 so that the shield tunnel 23 layer is formed at a depth of 150 μm from the upper surface. It is formed. At this time, in order to form the shield tunnel 23 at a depth of 75 μm from the upper surface of the sapphire substrate 10, the pulse energy of the pulse laser beam is set to 10 μJ / 1 pulse based on the result of Experiment 2 above.

Next, an amorphous removal step is performed in which the upper surface of the sapphire substrate 10 as the single crystal substrate on which the shield tunnel formation step has been performed is polished to form the upper surface of the sapphire substrate 10 with recesses. This amorphous removal step is performed in the same manner as the amorphous removal process shown in FIGS. 12A and 12B using the polishing apparatus 6 shown in FIG. . As a result, since the region where the shield tunnels 23 are scattered on the upper surface of the sapphire substrate 10 is fragile as described above, abrasive grains made of silica or the like having a sapphire (Al ₂ O ₃ ) hardness or less are used. By polishing, only the region where the shield tunnel 23 is formed can be easily removed, so that the surface of the sapphire substrate 10 as a single crystal substrate shown in FIG. In the embodiment, the concave portion 101 having a depth of 75 μm) can be efficiently formed to a predetermined thickness.

Above, and mainly sapphire (Al ₂ O ₃₎ processing the substrate in the above-described embodiment, sapphire (Al ₂ O ₃₎ substrate, silicon carbide (SiC) substrate has been described experimental example of the gallium nitride (GaN) substrate However, the present invention can also be applied to single crystal substrates such as a quartz (SiO2) substrate, a lithium tantalate (LT) substrate, a lithium niobate (LN) substrate, and a langasite (La3Ga5SiO14) substrate.

2: Optical device wafer 21: Optical device 22: Planned division line 23: Shield tunnel 3: Ring frame 30: Dicing tape 4: Laser processing device 41: Chuck table 42 of laser processing device: Laser beam irradiation means 422: Condenser 5: Splitting device 10: Sapphire substrate

Claims (6)

A method of processing a single crystal substrate,
A numerical aperture setting step for setting a numerical aperture (NA) of a condensing lens for condensing a pulsed laser beam to a predetermined value for a single crystal substrate;
The focused point of the pulse laser beam is positioned at the desired position from the upper surface of the single crystal substrate, and the pulse laser beam is irradiated to grow the pores and the amorphous material that shields the pores from the upper surface of the single crystal substrate. Forming a shield tunnel to be formed;
And amorphous removing step of removing the amorphous by polishing the shield tunnel formed in the single crystal substrate by abrasive, only including,
The numerical aperture (NA) of the condenser lens set to a predetermined value in the numerical aperture setting step is such that the value divided by the refractive index (N) of the single crystal substrate is in the range of 0.05 to 0.2. Set to
A method for processing a single crystal substrate, comprising: Abrasives used in amorphous removing step is less of a single crystal substrate hardness, machining method of a single crystal substrate according to claim 1 Symbol placement. The single crystal substrate is a sapphire (Al ₂ O ₃ ) substrate, a silicon carbide (SiC) substrate, or a gallium nitride (GaN) substrate,
Abrasive, sapphire (Al ₂ O _3), silicon carbide (SiC), gallium nitride (GaN), silicates, is any abrasive grains made of quartz, according to claim 1 or 2, wherein the single crystal substrate Processing method. The shield tunnel forming step is formed by connecting shield tunnels along a contour dividing a single crystal substrate into chips,
Amorphous removal step of polishing the outer periphery of the chip, the processing method of the single crystal substrate according to any one of claims 1-3. The shield tunnel forming step is formed by connecting shield tunnels with a predetermined depth on the upper surface of the single crystal substrate,
Amorphous removal step, by polishing the upper surface of the single crystal substrate to form a single crystal substrate to a predetermined thickness, the processing method of the single crystal substrate according to any one of claims 1-3. In the shield tunnel forming step, shield tunnels are formed at desired positions on the upper surface of the single crystal substrate,
The method for processing a single crystal substrate according to any one of claims 1 to 3 , wherein the amorphous removal step comprises polishing the single crystal substrate to form a recess on the upper surface of the single crystal substrate.

Images