Substrate dividing method

The present invention relates to a substrate on which a cutting origin region used for dividing a substrate such as a semiconductor substrate is formed in a semiconductor device manufacturing process or the like.

With the recent miniaturization of semiconductor devices, the semiconductor substrate may be thinned to a thickness of about several tens of μm in the semiconductor device manufacturing process. When the thin semiconductor substrate is cut by a blade and divided, the occurrence of chipping and cracking increases as compared with the case where the semiconductor substrate is thick, and the yield of the semiconductor chip obtained by dividing the semiconductor substrate decreases. There is a problem.

As a method for dividing a semiconductor substrate that can solve such a problem, the methods described in Patent Document 1 and Patent Document 2 are known.

That is, in these methods described in Documents 1 and 2, a groove is formed from the surface side of a semiconductor substrate on which a functional element is formed by a blade, and then an adhesive sheet is applied to the surface. By pasting and holding the semiconductor substrate and polishing the back surface of the semiconductor substrate until it reaches a groove formed in advance, the semiconductor substrate is made thinner and the semiconductor substrate is divided.

<p><patcit num="1"><text>Japanese Unexamined Patent Publication No. 64-38209</text></patcit><patcit num="2"><text>Japanese Unexamined Patent Publication No. 62-4341</text></patcit></p>

<p> However, in the methods described in Documents 1 and 2, when the back surface of the semiconductor substrate is polished by surface grinding, when the surface ground surface reaches a groove formed in advance on the semiconductor substrate, when the surface ground surface reaches a groove formed in advance on the semiconductor substrate, Chipping or cracking may occur on the side surface of the groove.</p><p> Therefore, the present invention has been made in view of such circumstances, and a substrate having a cutting starting point region formed which can prevent the occurrence of chipping and cracking, reduce the thickness of the substrate, and divide the substrate. The purpose is to provide.</p>

<p> In order to achieve the above object, the substrate on which the cutting starting point region according to the present invention is formed is a substrate in which the modified region serving as the cutting starting point region by laser irradiation is formed in a grid pattern in a plan view. ,<u style="single">The modified region is formed so as to deviate from the center position in the thickness direction of the substrate toward one surface side of the substrate.</u>Cracks originating from the modified region have reached one surface of the substrate, but have reached the other surface.<u style="single">Not, and the cut surfaces due to cracks are in close contact with each other.</u>It is characterized by that.</p><p> In this substrate, a laser beam is irradiated by aligning a condensing point inside the substrate, and a modified region is formed inside the substrate in a grid pattern, and cracks starting from the modified region are formed on one side of the substrate. It has reached one surface, but not the other.</p><p> Then, this substrate can be divided in a grid pattern with such a modified region as a cutting starting point.</p><p> Here, the focusing point is a point where the laser beam is focused. Further, polishing means including cutting, grinding, chemical etching and the like. Further, the cutting starting point region means a region that becomes a cutting starting point when the substrate is cut. Therefore, the cutting starting point region is a planned cutting portion where cutting is scheduled on the substrate. The cutting origin region may be formed by continuously forming the modified region, or may be formed by intermittently forming the modified region.</p><p> Examples of the substrate include a semiconductor substrate such as a silicon substrate and a GaAs substrate, and an insulating substrate such as a sapphire substrate and an AlN substrate.</p>

<p> According to the present invention, it is possible to prevent the occurrence of chipping and cracking, to make the substrate thinner, and to provide a substrate having a cutting starting point region in which the substrate can be divided.</p>

<figref num="1">It is a top view of the processing object during laser processing by the laser processing method which concerns on this embodiment.</figref><figref num="2">It is sectional drawing along the line II-II of the processing object shown in FIG.</figref><figref num="3">It is a top view of the processing object after laser processing by the laser processing method which concerns on this embodiment.</figref><figref num="4">It is sectional drawing along the IV-IV line of the processing object shown in FIG.</figref><figref num="5">It is sectional drawing along the VV line of the processing object shown in FIG.</figref><figref num="6">It is a top view of the processing object cut by the laser processing method which concerns on this embodiment.</figref><figref num="7">It is a graph which shows the relationship between the electric field strength and the size of a crack spot in the laser processing method which concerns on this embodiment.</figref><figref num="8">It is sectional drawing of the processing object in the 1st process of the laser processing method which concerns on this embodiment.</figref><figref num="9">It is sectional drawing of the processing object in the 2nd step of the laser processing method which concerns on this embodiment.</figref><figref num="10">It is sectional drawing of the processing object in the 3rd process of the laser processing method which concerns on this embodiment.</figref><figref num="11">It is sectional drawing of the processing object in the 4th process of the laser processing method which concerns on this embodiment.</figref><figref num="12">It is a figure showing the photograph of the cross section of a part of the silicon wafer cut by the laser processing method which concerns on this embodiment.</figref><figref num="13">It is a graph which shows the relationship between the wavelength of the laser light and the transmittance inside the silicon substrate in the laser processing method which concerns on this embodiment.</figref><figref num="14">It is a schematic block diagram of the laser processing apparatus which concerns on Example 1. FIG.</figref><figref num="15">It is a flowchart for demonstrating the laser processing method which concerns on Example 1. FIG.</figref><figref num="16">It is a figure which shows the semiconductor substrate after the process of forming the cutting origin region which concerns on Example 1. FIG.</figref><figref num="17">It is a figure for demonstrating the process of pasting the protective film which concerns on Example 1. FIG.</figref><figref num="18">It is a figure for demonstrating the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="19">It is a figure for demonstrating the process of pasting the expansion film which concerns on Example 1. FIG.</figref><figref num="20">It is a figure for demonstrating the process of peeling off the protective film which concerns on Example 1. FIG.</figref><figref num="21">It is a figure for demonstrating the process of expanding the expansion film which concerns on Example 1 and picking up a semiconductor chip.</figref><figref num="22">It is a figure which shows the chamfer formed in the edge part on the back surface side of the cut surface of the semiconductor chip after the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="23">It is a figure for demonstrating the case where the melt processing region remains in the cut surface of the semiconductor chip after the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="24">It is a figure for demonstrating the case where the melt processing region does not remain in the cut surface of the semiconductor chip after the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="25">It is a figure for demonstrating the case where the melt processing region remains in the edge part on the back surface side of the cut surface of the semiconductor chip after the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="26">It is sectional drawing of the peripheral part of the semiconductor substrate before the process of polishing the semiconductor substrate which concerns on Example 1. FIG.</figref><figref num="27">It is a top view of the sapphire substrate which concerns on Example 2. FIG.</figref><figref num="28">It is sectional drawing for demonstrating the process of forming the cutting origin region which concerns on Example 2. FIG.</figref><figref num="29">It is sectional drawing for demonstrating the process of forming a functional element which concerns on Example 2. FIG.</figref><figref num="30">It is sectional drawing for demonstrating the process of pasting the protective film which concerns on Example 2. FIG.</figref><figref num="31">It is sectional drawing for demonstrating the process of polishing the sapphire substrate which concerns on Example 2. FIG.</figref><figref num="32">It is sectional drawing for demonstrating the process of pasting the expansion film which concerns on Example 2. FIG.</figref><figref num="33">It is sectional drawing for demonstrating the process of irradiating the protective film which concerns on Example 2 with ultraviolet rays.</figref><figref num="34">It is sectional drawing for demonstrating the process of peeling off the protective film which concerns on Example 2. FIG.</figref><figref num="35">It is sectional drawing for demonstrating the process of expanding the expansion film which concerns on Example 2 and separating a semiconductor chip.</figref>

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. The method of dividing a substrate according to the present embodiment is a step of forming a cutting starting point region by aligning a condensing point inside the substrate and irradiating it with laser light to form a modified region by absorbing multiple photons inside the substrate. After the step of forming the cutting starting point region, the step of polishing the substrate so that the substrate has a predetermined thickness is provided.

First, a laser processing method carried out in the step of forming a cutting starting point region, particularly multiphoton absorption will be described.

Material absorption bandgap E_GWhen the photon energy hν is smaller than that, it becomes optically transparent. Therefore, the condition for absorption in the material is hν&gt; E_GIs. However, even if it is optically transparent, if the intensity of the laser beam is made very large, nhν&gt; E_GAbsorption occurs in the material under the condition of (n = 2,3,4, ...). This phenomenon is called multiphoton absorption. In the case of a pulse wave, the intensity of the laser beam is the peak power density (W / cm) of the focusing point of the laser beam.²), For example, the peak power density is 1 × 10.⁸(W / cm²) Multiphoton absorption occurs under the above conditions. The peak power density is obtained by (energy per pulse of laser light at the focusing point) ÷ (beam spot cross-sectional area of laser light × pulse width). In the case of continuous waves, the intensity of the laser beam is the electric field intensity (W / cm) at the focusing point of the laser beam.²).

The principle of laser machining according to the present embodiment utilizing such multiphoton absorption will be described with reference to FIGS. 1 to 6. FIG. 1 is a plan view of the substrate 1 being laser-processed, FIG. 2 is a cross-sectional view of the substrate 1 shown in FIG. 1 along lines II-II, and FIG. 3 is a plan view of the substrate 1 after laser processing. Yes, FIG. 4 is a sectional view of the substrate 1 shown in FIG. 3 along the IV-IV line, FIG. 5 is a sectional view of the substrate 1 shown in FIG. 3 along the VV line, and FIG. 6 is cut. It is a top view of the substrate 1.

As shown in FIGS. 1 and 2, the surface 3 of the substrate 1 has a desired planned cutting line 5 from which the substrate 1 should be cut. The planned cutting line 5 is a virtual line extending in a straight line (a line may be actually drawn on the substrate 1 to be the planned cutting line 5). In the laser processing according to the present embodiment, the substrate 1 is irradiated with the laser beam L by aligning the condensing point P inside the substrate 1 under the condition that multiphoton absorption occurs to form the modified region 7. The focusing point is a point where the laser beam L is focused.

By moving the laser beam L relatively along the planned cutting line 5 (that is, along the direction of arrow A), the focusing point P is moved along the planned cutting line 5. As a result, as shown in FIGS. 3 to 5, the modified region 7 is formed only inside the substrate 1 along the scheduled cutting line 5, and the modified region 7 forms the cutting starting region (scheduled cutting portion) 8. It is formed. In the laser processing method according to the present embodiment, the substrate 1 does not generate heat by absorbing the laser beam L to form the modified region 7. The modified region 7 is formed by transmitting the laser beam L through the substrate 1 and generating multiphoton absorption inside the substrate 1. Therefore, since the laser beam L is hardly absorbed by the surface 3 of the substrate 1, the surface 3 of the substrate 1 does not melt.

In cutting the substrate 1, if there is a starting point at the cutting point, the substrate 1 is cracked from the starting point, so that the substrate 1 can be cut with a relatively small force as shown in FIG. Therefore, the substrate 1 can be cut without causing unnecessary cracks on the surface 3 of the substrate 1.

The following two methods can be considered for cutting the substrate starting from the cutting starting point region. One is a case where the substrate is cracked and the substrate is cut from the cutting origin region by applying an artificial force to the substrate after the cutting origin region is formed. This is, for example, cutting when the thickness of the substrate is large. When an artificial force is applied, for example, bending stress or shear stress is applied to the substrate along the cutting starting point region of the substrate, or thermal stress is generated by giving a temperature difference to the substrate. .. The other is a case where the cutting starting point region is formed to spontaneously crack in the cross-sectional direction (thickness direction) of the substrate starting from the cutting starting point region, and as a result, the substrate is cut. This is possible, for example, when the thickness of the substrate is small, a cutting starting point region is formed by one row of modified regions, and when the thickness of the substrate is large, multiple rows are formed in the thickness direction. This is possible by forming a cutting starting point region with the modified region. Even in the case of this spontaneous cracking, the crack does not advance to the surface of the portion corresponding to the portion where the cutting origin region is not formed at the cutting portion, and the portion corresponding to the portion where the cutting origin region is formed. Since only can be divided, the division can be controlled well. In recent years, the thickness of substrates such as silicon wafers has tended to decrease, so such a cutting method with good controllability is very effective.

The modified regions formed by multiphoton absorption in this embodiment include the following (1) to (3).

(1) When the modified region is a crack region containing one or more cracks Substrate (for example, glass or LiTaO)₃The electric field strength at the condensing point is 1 × 10 by aligning the condensing point inside the piezoelectric material consisting of⁸(W / cm²) Or more and the pulse width is 1 μs or less. The magnitude of this pulse width is a condition in which a crack region can be formed only inside the substrate without causing extra damage to the surface of the substrate while causing multiphoton absorption. As a result, a phenomenon called optical damage due to multiphoton absorption occurs inside the substrate. This optical damage induces thermal strain inside the substrate, which forms crack regions inside the substrate. The upper limit of the electric field strength is, for example, 1 × 10.¹²(W / cm²). The pulse width is preferably 1 ns to 200 ns, for example. The formation of crack regions by multiphoton absorption is described in, for example, "Inside the glass substrate by solid-state laser harmonics" on pages 23 to 28 of the 45th Laser Thermal Processing Workshop Proceedings (December 1998). It is described in "Marking".

The present inventor experimentally determined the relationship between the electric field strength and the size of cracks. The experimental conditions are as follows.

(A) Substrate: Pyrex® glass (thickness 700 μm) (B) Laser Light source: Semiconductor laser excitation Nd: YAG laser Wavelength: 1064nm Laser light spot cross-sectional area: 3.14 × 10^-8cm² Oscillation form: Q-switched pulse Repeat frequency: 100kHz Pulse width: 30ns Output: Output &lt;1mJ / pulse Laser light quality: TEM₀₀ Polarization characteristics: Linearly polarized light (C) Condensing lens Transmittance to laser wavelength: 60 percent (D) Moving speed of the mounting table on which the board is mounted: 100 mm / sec

The laser light quality is TEM.₀₀Means that the light collecting property is high and the light can be collected up to the wavelength of the laser beam.

FIG. 7 is a graph showing the results of the above experiment. The horizontal axis is the peak power density, and since the laser light is a pulsed laser light, the electric field strength is represented by the peak power density. The vertical axis shows the size of the crack portion (crack spot) formed inside the substrate by the 1-pulse laser beam. Crack spots gather to form a crack area. The size of the crack spot is the size of the portion having the maximum length in the shape of the crack spot. The data indicated by the black circles in the graph is when the magnification of the condensing lens (C) is 100 times and the numerical aperture (NA) is 0.80. On the other hand, the data indicated by white circles in the graph is when the magnification of the condensing lens (C) is 50 times and the numerical aperture (NA) is 0.55. Peak power density is 10¹¹(W / cm²It can be seen that crack spots are generated inside the substrate from the degree of), and the crack spots also increase as the peak power density increases.

Next, in the laser processing according to the present embodiment, the mechanism of cutting the substrate by forming the crack region will be described with reference to FIGS. 8 to 11. As shown in FIG. 8, the condensing point P is aligned with the inside of the substrate 1 under the condition that multiphoton absorption occurs, and the substrate 1 is irradiated with the laser beam L to form a crack region 9 inside along the planned cutting line. .. The crack region 9 is a region containing one or a plurality of cracks. A cutting starting point region is formed by the crack region 9. As shown in FIG. 9, the crack grows further starting from the crack region 9 (that is, starting from the cutting starting region), and as shown in FIG. 10, the crack reaches the front surface 3 and the back surface 21 of the substrate 1, and FIG. 11 As shown in, the substrate 1 is cut by cracking the substrate 1. The cracks that reach the front and back surfaces of the substrate may grow naturally, or may grow when a force is applied to the substrate.

(2) When the modified region is a melt processing region By aligning the condensing point inside the substrate (for example, a semiconductor material such as silicon), the electric field strength at the condensing point is 1 × 10.⁸(W / cm²) Or more and the pulse width is 1 μs or less. As a result, the inside of the substrate is locally heated by multiphoton absorption. By this heating, a melt processing region is formed inside the substrate. The melting treatment region is a region that has been once melted and then resolidified, a region that is exactly in a molten state, or a region that is in a state of being resolidified from a molten state, and can also be said to be a region in which the phase has changed or a region in which the crystal structure has changed. Further, the melt-treated region can be said to be a region in which one structure is changed to another in a single crystal structure, an amorphous structure, or a polycrystalline structure. That is, for example, it means a region changed from a single crystal structure to an amorphous structure, a region changed from a single crystal structure to a polycrystalline structure, and a region changed from a single crystal structure to a structure including an amorphous structure and a polycrystalline structure. To do. When the substrate has a silicon single crystal structure, the melt processing region has, for example, an amorphous silicon structure. The upper limit of the electric field strength is, for example, 1 × 10.¹²(W / cm²). The pulse width is preferably 1 ns to 200 ns, for example.

The present inventor has experimentally confirmed that a melt processing region is formed inside the silicon wafer. The experimental conditions are as follows.

(A) Substrate: Silicon wafer (thickness 350 μm, outer diameter 4 inches) (B) Laser Light source: Semiconductor laser excitation Nd: YAG laser Wavelength: 1064nm Laser light spot cross-sectional area: 3.14 × 10^-8cm² Oscillation form: Q-switched pulse Repeat frequency: 100kHz Pulse width: 30ns Output: 20 μJ / pulse Laser light quality: TEM₀₀ Polarization characteristics: Linearly polarized light (C) Condensing lens Magnification: 50x NA: 0.55 Transmittance to laser wavelength: 60 percent (D) Moving speed of the mounting table on which the board is mounted: 100 mm / sec

FIG. 12 is a diagram showing a photograph of a cross section of a part of a silicon wafer cut by laser processing under the above conditions. A melt processing region 13 is formed inside the silicon wafer 11. The size of the melt processing region 13 formed under the above conditions in the thickness direction is about 100 μm.

It will be described that the melt processing region 13 is formed by multiphoton absorption. FIG. 13 is a graph showing the relationship between the wavelength of the laser beam and the transmittance inside the silicon substrate. However, the reflective components on the front surface side and the back surface side of the silicon substrate are removed, and the transmittance is shown only inside. The above relationship was shown for each of the thickness t of the silicon substrate of 50 μm, 100 μm, 200 μm, 500 μm, and 1000 μm.

For example, at 1064 nm, which is the wavelength of the Nd: YAG laser, when the thickness of the silicon substrate is 500 μm or less, it can be seen that 80% or more of the laser light is transmitted inside the silicon substrate. Since the thickness of the silicon wafer 11 shown in FIG. 12 is 350 μm, the melt processing region 13 by multiphoton absorption is formed near the center of the silicon wafer, that is, at a portion of 175 μm from the surface. In this case, the transmittance is 90% or more with reference to a silicon wafer having a thickness of 200 μm, so that the laser beam is hardly absorbed inside the silicon wafer 11 and most of it is transmitted. This does not mean that the laser beam is absorbed inside the silicon wafer 11 and the melt processing region 13 is formed inside the silicon wafer 11 (that is, the melt processing region is formed by normal heating by the laser light). It means that the melt processing region 13 was formed by multiphoton absorption. The formation of the melt processing region by multiphoton absorption can be found in, for example, "Evaluation of Silicon Machining Characteristics by Picosecond Pulsed Laser" on pages 72-73 of the 66th Annual Meeting of the Welding Society (April 2000). Have been described.

In the silicon wafer, cracks are generated in the cross-sectional direction starting from the cutting starting point region formed by the melting processing region, and the cracks reach the front surface and the back surface of the silicon wafer, resulting in cutting. Will be done. These cracks that reach the front and back surfaces of the silicon wafer may grow naturally, or may grow when a force is applied to the silicon wafer. When cracks naturally grow from the cutting starting point region to the front surface and the back surface of the silicon wafer, cracks grow from the state where the melting treatment region forming the cutting starting point region is melted, and the cutting starting point region. In both cases, cracks grow when the melt-treated region forming the wafer is resolidified from the molten state. However, in both cases, the melt processing region is formed only inside the silicon wafer, and the melt treatment region is formed only inside as shown in FIG. 12 on the cut surface after cutting. When the cutting starting point region is formed inside the substrate by the melting treatment region, unnecessary cracking off the cutting starting point region line is unlikely to occur at the time of cutting, so that cutting control becomes easy.

(3) When the modified region is the refractive index change region Align the focusing point inside the substrate (for example, glass), and the electric field strength at the focusing point is 1 x 10⁸(W / cm²) Or more and the pulse width is 1 ns or less to irradiate the laser beam. When the pulse width is made extremely short and multiphoton absorption occurs inside the substrate, the energy due to multiphoton absorption is not converted to thermal energy, and the ion valence changes, crystallization, polarization orientation, etc. inside the substrate. Permanent structural change is induced to form a refractive index change region. The upper limit of the electric field strength is, for example, 1 × 10.¹²(W / cm²). The pulse width is preferably 1 ns or less, more preferably 1 ps or less, for example. The formation of the refractive index change region by multiphoton absorption is described, for example, in "Proceedings of the 42nd Laser Thermal Processing Study Group (November 1997)" on pages 105 to 111, "Inside the glass by femtosecond laser irradiation. Photoinduced structure formation .

The cases (1) to (3) have been described above as the modified region formed by multiphoton absorption, but the cutting starting region should be formed as follows in consideration of the crystal structure of the substrate and its cleavage. For example, it is possible to cut the substrate with even smaller force and with high accuracy starting from the cutting starting point region.

That is, in the case of a substrate made of a diamond-structured single crystal semiconductor such as silicon, a cutting origin region is formed in the direction along the (111) plane (first cleavage plane) and the (110) plane (second cleavage plane). Is preferable. Further, in the case of a substrate made of a group III-V compound semiconductor having a zinc blende type structure such as GaAs, it is preferable to form a cutting origin region in the direction along the (110) plane. In addition, sapphire (Al₂O₃In the case of a substrate having a hexagonal crystal structure such as), the (0001) plane (C plane) is the main plane and the direction is along the (1120) plane (A plane) or the (1100) plane (M plane). It is preferable to form a cutting origin region.

The orientation flat on the substrate along the direction in which the cutting origin region should be formed (for example, the direction along the (111) plane in the single crystal silicon substrate) or the direction orthogonal to the direction in which the cutting origin region should be formed. Then, by using the orientation flat as a reference, it becomes possible to easily and accurately form the cutting origin region along the direction in which the cutting origin region should be formed on the substrate.

Hereinafter, the present invention will be described in more detail with reference to Examples.

[Example 1] Example 1 of the substrate dividing method according to the present invention will be described. In the first embodiment, the substrate 1 is a silicon wafer (thickness 350 μm, outer diameter 4 inches) (hereinafter, substrate 1 is referred to as semiconductor substrate 1 in the first embodiment), and the surface of the semiconductor substrate 1 is used in the device manufacturing process. In 3, the target is one in which a plurality of functional elements are formed in a matrix. Here, the functional element means a light receiving element such as a photodiode, a light emitting element such as a laser diode, or a circuit element formed as a circuit.

First, the step of forming the cutting starting point region inside the semiconductor substrate 1 will be described. Prior to the description thereof, the laser processing apparatus used in the step of forming the cutting starting point region will be described with reference to FIG. .. FIG. 14 is a schematic configuration diagram of the laser processing apparatus 100.

The laser processing apparatus 100 includes a laser light source 101 that generates a laser beam L, a laser light source control unit 102 that controls the laser light source 101 to adjust the output of the laser beam L, a pulse width, and the like, and a reflection function of the laser beam L. A dichroic mirror 103 that has a light source L and is arranged so as to change the direction of the optical axis of the laser light L by 90 °, a light collecting lens 105 that collects the laser light L reflected by the dichroic mirror 103, and a light collecting lens 105 The mounting table 107 on which the semiconductor substrate 1 irradiated with the laser beam L focused by the lens 105 is mounted, the X-axis stage 109 for moving the mounting table 107 in the X-axis direction, and the mounting table 107 are mounted on X. A Y-axis stage 111 for moving in the Y-axis direction orthogonal to the axial direction, a Z-axis stage 113 for moving the mounting table 107 in the Z-axis direction orthogonal to the X-axis and the Y-axis direction, and these three stages. It is provided with a stage control unit 115 that controls the movement of 109,111,113.

Since the Z-axis direction is orthogonal to the surface 3 of the semiconductor substrate 1, it is the direction of the depth of focus of the laser beam L incident on the semiconductor substrate 1. Therefore, by moving the Z-axis stage 113 in the Z-axis direction, the condensing point P of the laser beam L can be aligned with the inside of the semiconductor substrate 1. Further, the movement of the condensing point P in the X (Y) axis direction is performed by moving the semiconductor substrate 1 in the X (Y) axis direction by the X (Y) axis stage 109 (111).

The laser light source 101 is an Nd: YAG laser that generates pulsed laser light. Other lasers that can be used for the laser light source 101, Nd: YVO₄There are lasers, Nd: YLF lasers and titanium sapphire lasers. When forming a melt processing region, Nd: YAG laser, Nd: YVO₄It is preferable to use a laser or Nd: YLF laser. In the first embodiment, the pulsed laser beam is used for processing the semiconductor substrate 1, but continuous wave laser beam may be used as long as it can cause multiphoton absorption.

The laser processing apparatus 100 further includes an observation light source 117 that generates visible light for illuminating the semiconductor substrate 1 mounted on the mounting table 107 with visible light, and the same optical axis as the dichroic mirror 103 and the condensing lens 105. It is equipped with a beam splitter 119 for visible light arranged above. A dichroic mirror 103 is arranged between the beam splitter 119 and the condensing lens 105. The beam splitter 119 has a function of reflecting about half of visible light and transmitting the other half, and is arranged so as to change the direction of the optical axis of visible light by 90 °. Approximately half of the visible light generated from the observation light source 117 is reflected by the beam splitter 119, and this reflected visible light passes through the dichroic mirror 103 and the condensing lens 105, and cuts the semiconductor substrate 1 such as the line 5 scheduled to be cut. Illuminate the including surface 3.

The laser processing apparatus 100 further includes a beam splitter 119, a dichroic mirror 103, and an image pickup element 121 and an imaging lens 123 arranged on the same optical axis as the focusing lens 105. As the image sensor 121, for example, there is a CCD camera. The reflected light of visible light that illuminates the surface 3 including the line 5 to be cut passes through the focusing lens 105, the dichroic mirror 103, and the beam splitter 119, is imaged by the imaging lens 123, and is imaged by the imaging element 121. It becomes the imaging data.

The laser processing apparatus 100 further includes an imaging data processing unit 125 into which imaging data output from the imaging element 121 is input, an overall control unit 127 that controls the entire laser processing apparatus 100, and a monitor 129. The image pickup data processing unit 125 calculates focus data for focusing the visible light generated by the observation light source 117 on the surface 3 based on the image pickup data. Based on this focus data, the stage control unit 115 moves and controls the Z-axis stage 113 so that the visible light is focused on the surface 3. Therefore, the image pickup data processing unit 125 functions as an autofocus unit. Further, the image pickup data processing unit 125 calculates image data such as an enlarged image of the surface 3 based on the image pickup data. This image data is sent to the overall control unit 127, various processes are performed by the overall control unit, and the image data is sent to the monitor 129. As a result, an enlarged image or the like is displayed on the monitor 129.

Data from the stage control unit 115, image data from the imaging data processing unit 125, and the like are input to the overall control unit 127, and the laser light source control unit 102, the observation light source 117, and the stage control unit are also based on these data. By controlling 115, the entire laser processing apparatus 100 is controlled. Therefore, the overall control unit 127 functions as a computer unit.

Subsequently, the step of forming the cutting starting point region when the above-mentioned laser processing apparatus 100 is used will be described with reference to FIGS. 14 and 15. FIG. 15 is a flowchart for explaining a process of forming a cutting starting point region.

The light absorption characteristics of the semiconductor substrate 1 are measured by a spectrophotometer or the like (not shown). Based on this measurement result, a laser light source 101 that generates laser light L having a transparent wavelength or a wavelength with little absorption with respect to the semiconductor substrate 1 is selected (S101). Subsequently, the thickness of the semiconductor substrate 1 is measured. The amount of movement of the semiconductor substrate 1 in the Z-axis direction is determined based on the measurement result of the thickness and the refractive index of the semiconductor substrate 1 (S103). This is because the focusing point P of the laser beam L is located inside the semiconductor substrate 1, so that the Z axis of the semiconductor substrate 1 is based on the focusing point P of the laser beam L located on the surface 3 of the semiconductor substrate 1. The amount of movement in the direction. This movement amount is input to the overall control unit 127.

The semiconductor substrate 1 is placed on the mounting table 107 of the laser processing apparatus 100. Then, visible light is generated from the observation light source 117 to illuminate the semiconductor substrate 1 (S105). The surface 3 of the semiconductor substrate 1 including the illuminated line 5 to be cut is imaged by the image sensor 121. The planned cutting line 5 is a desired virtual line on which the semiconductor substrate 1 should be cut. Here, in order to obtain a semiconductor chip by dividing the semiconductor substrate 1 for each functional element formed on its surface 3, the scheduled cutting lines 5 are set in a grid pattern so as to run between adjacent functional elements. The image pickup data captured by the image sensor 121 is sent to the image pickup data processing unit 125. Based on this image pickup data, the image pickup data processing unit 125 calculates focus data such that the focus of visible light of the observation light source 117 is located on the surface 3 (S107).

This focus data is sent to the stage control unit 115. The stage control unit 115 moves the Z-axis stage 113 in the Z-axis direction based on the focus data (S109). As a result, the focus of visible light of the observation light source 117 is located on the surface 3 of the semiconductor substrate 1. The image pickup data processing unit 125 calculates the enlarged image data of the surface 3 of the semiconductor substrate 1 including the scheduled cutting line 5 based on the image pickup data. This enlarged image data is sent to the monitor 129 via the overall control unit 127, so that the enlarged image near the line 5 scheduled to be cut is displayed on the monitor 129.

The movement amount data determined in advance in step S103 is input to the overall control unit 127, and this movement amount data is sent to the stage control unit 115. Based on this movement amount data, the stage control unit 115 moves the semiconductor substrate 1 in the Z-axis direction by the Z-axis stage 113 to a position where the condensing point P of the laser beam L is inside the semiconductor substrate 1 (S111). ..

Subsequently, the laser light L is generated from the laser light source 101, and the laser light L is applied to the planned cutting line 5 on the surface 3 of the semiconductor substrate 1. Since the condensing point P of the laser beam L is located inside the semiconductor substrate 1, the melt processing region is formed only inside the semiconductor substrate 1. Then, the X-axis stage 109 and the Y-axis stage 111 are moved along the planned cutting line 5, and the cutting starting point region along the planned cutting line 5 is set as a semiconductor by the melting processing region formed along the planned cutting line 5. It is formed inside the substrate 1 (S113).

As described above, the step of forming the cutting starting point region is completed, and the cutting starting point region is formed inside the semiconductor substrate 1. When the cutting starting point region is formed inside the semiconductor substrate 1, cracks occur naturally or by a relatively small force in the thickness direction of the semiconductor substrate 1 starting from the cutting starting point region.

In the first embodiment, in the step of forming the cutting starting point region described above, the cutting starting point region is formed at a position close to the surface 3 side inside the semiconductor substrate 1, and the cutting starting point region is used as the starting point in the thickness direction of the semiconductor substrate 1. Is cracked. FIG. 16 is a diagram showing the semiconductor substrate 1 after the cutting starting point region is formed. As shown in FIG. 16, the cracks 15 generated from the cutting starting point region in the semiconductor substrate 1 are formed in a grid pattern along the planned cutting line, reach only the front surface 3 of the semiconductor substrate 1, and reach the back surface 21 on the back surface 21. Not reached. That is, the crack 15 generated in the semiconductor substrate 1 individually divides a plurality of functional elements 19 formed in a matrix on the surface of the semiconductor substrate 1. Further, the cut surfaces of the semiconductor substrate 1 cut by the crack 15 are in close contact with each other.

In addition, "the cutting starting point region is formed at a position close to the surface 3 side inside the semiconductor substrate 1" means that the modified region such as the melt processing region constituting the cutting starting point region is in the thickness direction of the semiconductor substrate 1. It means that it is formed deviated from the center position (half the thickness) to the surface 3 side. That is, it means that the center position of the width of the modified region in the thickness direction of the semiconductor substrate 1 is deviated from the center position in the thickness direction of the semiconductor substrate 1 toward the surface 3 side, and the modified region. It does not mean that all the parts of the semiconductor substrate 1 are located on the surface 3 side with respect to the center position in the thickness direction of the semiconductor substrate 1.

Next, the process of polishing the semiconductor substrate 1 will be described with reference to FIGS. 17 to 21. 17 to 21 are diagrams for explaining each step including the step of polishing the semiconductor substrate. In Example 1, the semiconductor substrate 1 is thinned from a thickness of 350 μm to a thickness of 50 μm.

As shown in FIG. 17, the protective film 20 is attached to the surface 3 of the semiconductor substrate 1 after the cutting starting point region is formed. The protective film 20 is for protecting the functional element 19 formed on the surface 3 of the semiconductor substrate 1 and for holding the semiconductor substrate 1. Subsequently, as shown in FIG. 18, the back surface 21 of the semiconductor substrate 1 is surface-ground, and after this surface grinding, the back surface 21 is chemically etched to reduce the thickness of the semiconductor substrate 1 to 50 μm. As a result, that is, by polishing the back surface 21 of the semiconductor substrate 1, the back surface 21 reaches the cracks 15 generated from the cutting starting point region, and the semiconductor substrate 1 is divided into semiconductor chips 25 having each of the functional elements 19. The chemical etching described above includes wet etching (HF / HNO).₃) And plasma etching (HBr / Cl)₂) Etc. can be mentioned.

Then, as shown in FIG. 19, the expansion film 23 is attached so as to cover the back surface of all the semiconductor chips 25, and then, as shown in FIG. 20, it is attached so as to cover the functional elements 19 of all the semiconductor chips 25. The protective film 20 that had been removed is peeled off. Subsequently, as shown in FIG. 21, the expansion film 23 is expanded to separate the semiconductor chips 25 from each other, and the semiconductor chip 25 is picked up by the adsorption collet 27.

As described above, according to the substrate dividing method according to the first embodiment, the back surface 21 of the semiconductor substrate 1 can be polished after the functional element 19 is formed on the front surface 3 of the semiconductor substrate 1 in the device manufacturing process. .. Then, due to the following effects of each of the steps of forming the cutting starting point region and the step of polishing the semiconductor substrate, it is possible to obtain a thin semiconductor chip 25 with a high yield so as to correspond to the miniaturization of the semiconductor device. ..

That is, according to the step of forming the cutting starting point region, it is possible to prevent unnecessary cracking or melting that deviates from the desired cutting scheduled line for cutting the semiconductor substrate 1 on the surface 3 of the semiconductor substrate 1. It is possible to prevent unnecessary cracking and melting of the semiconductor chip 25 obtained by separating the semiconductor substrate 1.

Further, according to the step of forming the cutting starting point region, since the surface 3 of the semiconductor substrate 1 along the planned cutting line does not melt, the distance between the adjacent functional elements 19 can be narrowed, and the distance between the adjacent functional elements 19 can be narrowed. It is possible to increase the number of semiconductor chips 25 to be separated.

On the other hand, in the step of polishing the semiconductor substrate, the back surface 21 of the semiconductor substrate 1 is surface-ground so that the semiconductor substrate 1 has a predetermined thickness after the cutting starting point region is formed inside the semiconductor substrate 1. At this time, the back surface is ground. Even if 21 reaches the crack 15 generated from the cutting origin region, the cut surfaces of the semiconductor substrate 1 cut by the crack 15 are in close contact with each other, so that chipping or cracking of the semiconductor substrate 1 by surface polishing can be performed. Can be prevented. Therefore, it is possible to prevent the occurrence of chipping and cracking, reduce the thickness of the semiconductor substrate 1, and divide the semiconductor substrate 1.

The adhesion of the cut surface on the semiconductor substrate 1 described above prevents the grinding debris from entering the crack 15 caused by surface grinding, and prevents the grinding debris contamination of the semiconductor chip 25 obtained by dividing the semiconductor substrate 1 from being contaminated. It also has an effect. Similarly, the close contact of the cut surfaces on the semiconductor substrate 1 also has the effect of reducing chip skipping of the semiconductor chips 25 due to surface grinding as compared with the case where the semiconductor chips 25 are separated from each other. That is, a protective film 20 having a reduced holding force can be used.

Further, in the step of polishing the semiconductor substrate, since the back surface 21 of the semiconductor substrate 1 is chemically etched, the back surface of the semiconductor chip 25 obtained by dividing the semiconductor substrate 1 can be further smoothed. Further, since the cut surfaces of the semiconductor substrate 1 due to the cracks 15 generated from the cut starting point region are in close contact with each other, as shown in FIG. 22, only the edge portion on the back surface side of the cut surface is selectively etched. A chamfer 29 is formed. Therefore, the bending strength of the semiconductor chip 25 obtained by dividing the semiconductor substrate 1 can be improved, and the occurrence of chipping and cracking in the semiconductor chip 25 can be prevented.

The relationship between the semiconductor chip 25 and the melt processing region 13 after the step of polishing the semiconductor substrate is shown in FIGS. 23 (a) to 25 (b). Since the semiconductor chip 25 shown in each figure has each effect described later, it can be used properly according to various purposes. Here, FIGS. 23 (a), 24 (a), and 25 (a) show a case where the crack 15 reaches the surface 3 of the semiconductor substrate 1 before the process of polishing the semiconductor substrate, and FIG. 23 (a) b), FIG. 24 (b) and FIG. 25 (b) are cases where the crack 15 does not reach the surface 3 of the semiconductor substrate 1 before the step of polishing the semiconductor substrate. Also in the cases of FIGS. 23 (b), 24 (b) and 25 (b), the crack 15 reaches the surface 3 of the semiconductor substrate 15 after the step of polishing the semiconductor substrate.

As shown in FIGS. 23 (a) and 23 (b), in the semiconductor chip 25 in which the melt processing region 13 remains in the cut surface, the cut surface is protected by the melt treatment region 13, and the semiconductor chip 25 Improves anti-folding strength.

As shown in FIGS. 24 (a) and 24 (b), the semiconductor chip 25 in which the melt processing region 13 does not remain in the cut surface is effective when the melt treatment region 13 does not have a positive effect on the semiconductor device. Is.

As shown in FIGS. 25 (a) and 25 (b), in the semiconductor chip 25 in which the melt processing region 13 remains on the edge portion on the back surface side of the cut surface, the edge portion is protected by the melt treatment region 13. Therefore, it is possible to prevent the occurrence of chipping and cracking at the edge portion, as in the case where the edge portion of the semiconductor chip 25 is chamfered.

Further, as shown in FIGS. 23 (a), 24 (a), and 25 (a), the crack 15 reaches the surface 3 of the semiconductor substrate 1 before the step of polishing the semiconductor substrate. As shown in 23 (b), FIG. 24 (b) and FIG. 25 (b), the semiconductor substrate is polished when the crack 15 does not reach the surface 3 of the semiconductor substrate 1 before the process of polishing the semiconductor substrate. The straightness of the cut surface of the semiconductor chip 25 obtained after the process is further improved.

By the way, whether or not the crack 15 reaches the surface 3 of the semiconductor substrate 1 before the process of polishing the semiconductor substrate is, of course, related to the depth of the melt processing region 13 from the surface 3, but the melt treatment It is also related to the size of region 13. That is, if the size of the melt processing region 13 is reduced, the crack 15 does not reach the surface 3 of the semiconductor substrate 1 even if the depth from the surface 3 of the melt treatment region 13 is shallow. The size of the melting processing region 13 can be controlled, for example, by the output of the pulsed laser beam in the process of forming the cutting starting point region, and increases when the output of the pulsed laser beam is increased, and decreases when the output of the pulsed laser beam is decreased. Become.

Further, in consideration of a predetermined thickness of the semiconductor substrate 1 to be thinned in the step of polishing the semiconductor substrate (for example, before the step of forming the cutting starting point region), the semiconductor is formed by at least the predetermined thickness. It is preferable that the peripheral edge portion (outer peripheral portion) of the substrate 1 is rounded by chamfering. 26 (a) and 26 (b) are cross-sectional views of the peripheral edge of the semiconductor substrate 1 before and after the step of polishing the semiconductor substrate according to the first embodiment. The thickness of the semiconductor substrate 1 shown in FIG. 26 (a) before the process of polishing the semiconductor substrate is 350 μm, and the thickness of the semiconductor substrate 1 shown in FIG. 26 (b) after the process of polishing the semiconductor substrate is 50 μm. is there. As shown in FIG. 26 (a), a plurality of chamfered roundnesses (7 in this case) are formed in advance on the peripheral edge of the semiconductor substrate 1 at intervals of 50 μm, that is, a cross section of the peripheral edge of the semiconductor substrate 1. The shape is wavy. As a result, as shown in FIG. 26 (b), the peripheral edge portion of the semiconductor substrate 1 after the step of polishing the semiconductor substrate 1 is in a rounded state by chamfering, so that chipping and cracking occur in the peripheral edge portion. This can be prevented, and by extension, the handling can be facilitated by improving the mechanical strength.

[Example 2] Example 2 of the substrate dividing method according to the present invention will be described with reference to FIGS. 27 to 35. In Example 2, the substrate 1 is a sapphire substrate (thickness 450 μm, outer diameter 2 inches) which is an insulating substrate (hereinafter, in Example 2, board 1 is referred to as sapphire substrate 1), and a semiconductor which becomes a light emitting diode. When getting a chip. 28 to 35 are cross-sectional views of the sapphire substrate 1 shown in FIG. 27 along XX-XX.

First, as shown in FIG. 28, the condensing point P is aligned with the inside of the sapphire substrate 1 and the laser beam L is irradiated to form the modified region 7 inside the sapphire substrate 1. A plurality of functional elements 19 are formed in a matrix on the surface 3 of the sapphire substrate 1 in a later process, and the sapphire substrate 1 is divided for each of the functional elements 19. Therefore, the planned cutting lines are set in a grid pattern when viewed from the surface 3 side according to the size of each functional element 19, the modified region 7 is formed along the planned cutting line, and the modified region 7 is cut. The starting point area.

The peak power density at the condensing point P is 1 × 10.⁸(W / cm²) Or more and the pulse width is 1 μs or less, when the sapphire substrate 1 is irradiated with laser light, a crack region is formed as the modified region 7 (a melt processing region may be formed). Further, if the modified region 7 is formed with the (0001) plane of the sapphire substrate 1 as the surface 3 in the direction along the (1120) plane and in the direction orthogonal to the direction, this modified region 7 can be formed in a later step. Starting from the cutting starting point region according to 7, it is possible to cut the substrate with even smaller force and with high accuracy. This is the same even if the modified region 7 is formed in the direction along the (1100) plane and the direction orthogonal to the direction.

After the formation of the cutting starting point region by the modified region 7, as shown in FIG. 29, an n-type gallium nitride based compound semiconductor layer (hereinafter referred to as n-type layer) 31 is placed on the surface 3 of the sapphire substrate 1 with a thickness of 6 μm. The crystal is grown until the thickness becomes 1 μm, and the p-type gallium nitride based compound semiconductor layer (hereinafter referred to as p-type layer) 32 is further grown on the n-type layer 31 until the thickness becomes 1 μm. Then, by etching the n-type layer 31 and the p-type layer 32 halfway along the modified region 7 formed in a grid pattern, a plurality of n-type layers 31 and p-type layers 32 are formed. The functional elements 19 are formed in a matrix.

After forming the n-type layer 31 and the p-type layer 32 on the surface 3 of the sapphire substrate 1, the condensing point P is aligned with the inside of the sapphire substrate 1 and the laser beam L is irradiated to the inside of the sapphire substrate 1. The modified region 7 may be formed. Further, the laser beam L may be irradiated from the front surface 3 side of the sapphire substrate 1 or from the back surface 21 side. Even when the laser beam L is irradiated from the surface 3 side after the formation of the n-type layer 31 and the p-type layer 32, the laser beam L is light-transmitting to the sapphire substrate 1, the n-type layer 31 and the p-type layer 32. Therefore, it is possible to prevent the n-type layer 31 and the p-type layer 32 from melting.

After forming the functional element 19 composed of the n-type layer 31 and the p-type layer 32, the protective film 20 is attached to the surface 3 side of the sapphire substrate 1 as shown in FIG. The protective film 20 is for protecting the functional element 19 formed on the surface 3 of the sapphire substrate 1 and for holding the sapphire substrate 1. Subsequently, as shown in FIG. 31, the back surface 21 of the sapphire substrate 1 is surface-ground to reduce the thickness of the sapphire substrate 1 to a thickness of 150 μm. By polishing the back surface 21 of the sapphire substrate 1, cracks 15 are generated starting from the cutting starting point region of the modified region 7, and the cracks 15 reach the front surface 3 and the back surface 21 of the sapphire substrate 1 to reach the n-type layer 31. The sapphire substrate 1 is divided into semiconductor chips 25 having each of the functional elements 19 composed of the p-type layer 32 and the p-type layer 32.

Then, as shown in FIG. 32, the expandable expansion film 23 is attached so as to cover the back surface of all the semiconductor chips 25, and then, as shown in FIG. 33, the protective film 20 is irradiated with ultraviolet rays. The UV curable resin, which is the adhesive layer of the protective film 20, is cured, and the protective film 20 is peeled off as shown in FIG. 34. Subsequently, as shown in FIG. 35, the expansion film 23 is expanded outward to separate the semiconductor chips 25 from each other, and the semiconductor chips 25 are picked up by an adsorption collet or the like. After that, electrodes are attached to the n-type layer 31 and the p-type layer 32 of the semiconductor chip 25 to manufacture a light emitting diode.

As described above, according to the method of dividing the substrate according to the second embodiment, in the step of forming the cutting starting point region, the condensing point P is aligned with the inside of the sapphire substrate 1 and the laser beam L is irradiated to sapphire. Since the modified region 7 is formed by generating a phenomenon called multiphoton absorption inside the substrate 1, the modified region 7 of the sapphire substrate 1 is aligned with the desired cutting line in which the sapphire substrate 1 should be cut. A cutting origin region can be formed inside. When the cutting starting point region is formed inside the sapphire substrate 1, cracks 15 are generated naturally or by a relatively small force in the thickness direction of the sapphire substrate 1 starting from the cutting starting point region.

Then, in the step of polishing the sapphire substrate 1, after forming the cutting starting point region inside the sapphire substrate 1, the sapphire substrate 1 is polished so that the sapphire substrate 1 has a predetermined thickness. Even if the crack 15 generated from the cutting starting point region is reached, the cut surfaces of the sapphire substrate 1 cut by the crack 15 are in close contact with each other, so that chipping and cracking of the sapphire substrate 1 due to polishing are prevented. be able to.

Therefore, it is possible to prevent the occurrence of chipping and cracking, make the sapphire substrate 1 thinner, and divide the sapphire substrate 1, so that the semiconductor chip 25 having the thinner sapphire substrate 1 can be obtained with a high yield.

The same effect as described above can be obtained even when the substrate is divided when an AlN substrate or a GaAs substrate is used instead of the sapphire substrate 1.

1 ... substrate, 3 ... front surface, 5 ... planned cutting line, 7 ... modified area, 8 ... cutting starting area, 20 ... protective film, 21 ... back surface, 23 ... expansion film, 25 ... semiconductor chip (chip), L ... laser light, P ... focusing point.