JP2014033024A - Laser dicing device and method and wafer processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform laser dicing with high accuracy by preventing ablation even when performing laser dicing of a wafer, especially a thin wafer of 50 μm or less, thereby keeping the work distance constant and enhancing the irradiation efficiency.SOLUTION: A laser dicing device 10 for forming a modified layer in a wafer W having a plurality of devices formed on the surface by irradiating a wafer W with laser light includes a wafer table 20 having a flat holding surface formed to transmit laser light and holding the back side of the wafer W tightly, and a laser irradiation device 60 for irradiating the back side of the wafer W with laser light via the wafer table 20. In the gap between the condenser lens 66 surface of the laser irradiation device 60 and the laser light incident surface 22B of the wafer table 20, a liquid film 26A of a refraction liquid A having a refraction index larger than that of the air is held by interface tension.

Description

  The present invention relates to a laser dicing apparatus and method and a wafer processing method, and more particularly to a laser dicing apparatus and method and a wafer processing method for processing a wafer having a plurality of devices formed on the surface.

  As a method of dividing a wafer having a plurality of devices (semiconductor elements) formed on the surface into individual chips, a dicing method using a laser (laser dicing) is known. Laser dicing is a method in which a laser beam is incident on a wafer along a street (division planned line) to form a modified layer by multiphoton absorption inside the wafer. The laser-diced wafer is then divided into individual chips starting from the modified layer by applying external stress to the wafer. This laser dicing has the advantage that it can be divided with little chipping.

  Examples of methods and apparatuses for performing such laser dicing include Patent Documents 1 to 3.

  Patent Document 1 describes that a wafer is bonded and supported by a dicing tape and a ring frame, and laser dicing is performed by injecting laser light from the dicing tape side to the wafer back side.

  In Patent Document 2, it is proposed that a wafer is sandwiched and supported by a pair of transparent glass plates so that laser light can be incident from both front and back sides of the wafer.

  In Patent Document 3, a holding means having an annular holding surface that holds and contacts a region where a device is not formed is provided on the peripheral portion of the wafer surface, holds the wafer without touching the device formation region, It has been proposed to form a modified layer inside the wafer by applying laser light from the back side through a dicing tape.

  As can be seen from Patent Documents 1 to 3, generally, a circuit pattern or the like is formed on the surface side of the wafer, and if a metal film such as a circuit pattern is formed near the street, the laser beam is reflected. Therefore, laser light is incident from the back surface of the wafer.

  In Patent Document 4, laser dicing is performed in a state where a wafer is placed on a stage provided at the bottom of an immersion tank filled with a refractive liquid (for example, water) and a condenser lens of a laser head is immersed in the refractive liquid. Processing is described. Further, it is described that the refracting liquid in the immersion tank is caused to flow by jetting water. Thereby, since reflection of laser light can be suppressed, it is said that the irradiation efficiency of laser light can be improved.

  In any of these Patent Documents 1 to 4, laser dicing irradiates a laser beam efficiently at a certain condensing point depth (also referred to as a focal depth) position of the wafer thickness, and has a constant depth along the street. It is important to continuously form the modified layer at the position.

  By the way, in recent years, it is required to perform laser dicing on a wafer having a very thin thickness, for example, 50 μm or less, such as a wafer for TSV (through silicon via).

JP 2007-123404 A JP 2005-109045 A JP 2010-029927 A JP 2007-136482 A

  However, the methods and apparatuses of Patent Documents 1 to 4 have a drawback that the following problems that are likely to occur when laser dicing is performed on a thin wafer, for example, a wafer of 50 μm or less, cannot be solved.

  (1) Thin wafers are prone to ablation.

  Ablation refers to a phenomenon in which continuous holes are formed from the inside of the wafer to the laser light incident surface of the wafer (or the wafer back surface when irradiated from the back surface of the wafer) due to melting by the laser light applied to the wafer. When ablation occurs, the molten material scatters and adheres to the wafer, destroying the attached portion of the element.

  (2) The irradiation efficiency of a laser beam becomes worse for a thin wafer.

  When the thickness of the wafer is thin, it is necessary to make the laser beam incident from a wide angle since the depth of the condensing point of the laser beam inside the wafer must be reduced. However, when laser light is incident from a wide angle, the difference in refractive index between air and the wafer is large, so that the laser light is reflected at the interface (wafer surface), and the irradiation efficiency of the laser light is reduced accordingly. Thereby, sufficient modification cannot be performed. That said, if the laser power is increased too much, ablation tends to occur.

  (3) An extremely high precision work distance is required for a thin wafer.

    Here, the work distance refers to the distance between the tip of the condensing lens and the condensing point position set inside the wafer. If the work distance is not constant, the laser light focusing point depth changes when the laser light is irradiated along the wafer street (division line), so a modified layer can be formed at a fixed depth position. Disappear. As a result, not only the wafer splitting performance is degraded, but also the vicinity of the laser light incident surface of the wafer is melted when the work distance varies in the direction of shortening, which causes ablation. Therefore, when the wafer is thin, ablation occurs only when the work distance fluctuates in a slightly shorter direction.

  Furthermore, the warpage and bending unique to thin wafers hinder the constant work distance. The problem of warping or bending of the wafer can be corrected by holding the wafer on the table and applying laser light.

  However, in Patent Document 1, in order to allow laser light to enter the back surface of the wafer, the front surface side of the wafer is sucked and held by a table. If the surface side of the wafer is sucked and held by the table in this manner, for example, in the case of a wafer having a fine structure such as a MEMS element (Micro Electro Mechanical System), the element is destroyed by the suction. There is. Such a problem also occurs in the laser dicing apparatus of Patent Document 2 in which a wafer is sandwiched between transparent glass plates.

  Further, as described in Patent Document 3, it is conceivable to support only the peripheral edge of the wafer with a holding means having an annular holding surface, but if only the peripheral edge of the wafer is supported, the wafer is bent, There is a drawback that a modified layer cannot be formed in a predetermined region.

  In addition, the surface (tape surface) of a dicing tape (particularly in the case of a thin tape) that is generally adhered to the back surface of a wafer is less likely to form a flat surface such as a hard body such as glass, and generally undulates. Often have. In addition, there are many cases where the object is slightly rough. For this reason, when laser light is incident from the back surface of the wafer through the dicing tape, there is a problem that the laser light is not efficiently incident due to the influence of scattering on the tape surface.

  Further, in Patent Document 3, air is introduced from a blow unit into a space formed by a concave portion provided in a holding unit and a wafer, and the back surface of the wafer is pressed by a warp correction unit, thereby warping the wafer. I am trying to correct the swell. However, in the method disclosed in Patent Document 3 as described above, it is theoretically difficult to completely correct the warpage and waviness of the wafer. For this reason, when the laser beam is incident from the back surface of the wafer, the focusing point of the laser beam cannot be adjusted to a predetermined position inside the wafer, and the modified layer by the laser beam is accurately placed inside the wafer. Cannot be formed.

  In addition, when laser light is incident from the back surface of the wafer through the glass plate while supporting the back surface of the wafer with a glass plate, an air layer is formed between the glass surface and the back surface of the wafer (the surface on which the dicing tape is adhered). When interposed, reflection occurs repeatedly at the interface between them, and the laser light is scattered outside the desired portion of the wafer, causing laser burn. The loss due to such scattered light consumes energy in the vicinity of the loss, and the laser light in the vicinity, such as surface deterioration on the glass surface and adhesion of debris to the dicing tape due to deterioration, will be consumed. You will suffer from the harmful effects of scattering.

  Furthermore, if the glass surface and the back surface of the wafer are not in close contact, even if the laser power is increased and the laser is turned on, due to the scattering effect between the glass surface and the back surface of the wafer, the inside of the wafer can be efficiently processed. However, there is a problem that a modified layer is not formed.

  Another factor that makes the work distance difficult to keep constant is a change in the refractive index in the laser beam path between the condenser lens and the laser beam incident surface of the wafer. In the case of Patent Document 4, laser irradiation is performed in a state where the condenser lens is immersed in the refractive liquid in the immersion tank. Therefore, the refractive index of the refractive liquid is likely to fluctuate, and even if antireflection can be suppressed, the work distance is constant. Hard to become. That is, the stage is moved for laser dicing along the street, or the refractive liquid in the immersion tank is made to flow by jetting the refractive liquid to flow the refractive liquid in the immersion tank. Vortices and bubbles are easily generated. Moreover, since patent document 4 performs laser irradiation in the state which made the condensing lens immersed in the refractive liquid in an immersion tank, the liquid temperature of the vicinity of a condensing lens becomes high compared with another part, and it is in an immersion tank. Convection occurs in Because of these factors, the refractive index tends to fluctuate on the optical path between the condenser lens and the laser light incident surface of the wafer, so that the work distance is not constant.

  That is, the refracting liquid in Patent Document 4 suppresses reflection and contributes to an increase in irradiation efficiency, but acts badly for a constant work distance. Thus, by simply using the technique of Patent Document 4, it is not possible to laser dice a thin wafer, particularly a wafer of 50 μm or less, with high accuracy.

  In particular, when the work distance is not constant, it is not possible to cope with the problem of forming the modified layer with high precision only in the wafer without causing ablation even for a thin wafer. That is, it is difficult to form a modified layer constantly from the surface at a constant depth while scanning at high speed, even if it is adjusted to a constant depth from the surface while following the surface shape by autofocusing. The height position at which the modified layer is formed also varies due to a change in the refractive index due to the fluctuation of the liquid interposed therebetween. In particular, when a modified layer is accurately formed at a certain depth position for a very thin wafer, if the depth position where the modified layer is formed is shifted near the surface on the laser irradiation side, Not only the modification but also ablation is caused from the wafer surface. In such a case, the surface of the wafer is intact and does not meet the purpose of the present application for forming the modified layer only inside the wafer.

  The present invention has been made in view of such circumstances, and even when laser dicing is performed on a wafer, particularly a thin wafer of 50 μm or less, ablation can be prevented, a work distance can be made constant, and irradiation can be performed. Since efficiency can be improved, a laser dicing apparatus and method capable of performing highly accurate laser dicing and a wafer processing method are provided.

  Furthermore, the present invention can correct the warpage and deflection of the wafer without destroying the finely structured device formed on the wafer surface, and can make the work distance constant, so that it is a wafer, particularly a thin wafer of 50 μm or less. However, it is an object of the present invention to provide a laser dicing apparatus and method and a wafer processing method capable of forming a modified layer with high accuracy inside a wafer.

  In order to achieve the above object, a laser dicing apparatus according to the present invention irradiates a wafer having a plurality of devices on its surface with laser light and forms a modified layer inside the wafer. A wafer table having a flat holding surface formed so as to be able to transmit the laser beam and tightly holding the back side of the wafer, and a laser that irradiates the laser beam from the back side of the wafer through the wafer table A liquid film of the refractive liquid A having a refractive index larger than that of air due to the interfacial tension in the gap between the condensing lens surface of the laser irradiation means and the laser light incident surface of the wafer table. It is retained.

  In order to achieve the above object, the laser dicing method of the present invention irradiates a wafer having a plurality of devices formed on its surface with laser light from a laser irradiation means, and modifies the modified layer inside the wafer. In the laser dicing method for forming the wafer, a wafer holding step of holding the wafer by a wafer table formed so as to transmit the laser light and having a flat holding surface for tightly holding the back side of the wafer, and the laser A moving step of moving the laser irradiation means and / or the wafer table closer so that a gap of a predetermined distance is formed between the condenser lens surface of the irradiation means and the laser light incident surface of the wafer table; and air in the gap A liquid film forming step of forming a liquid film by supplying a refractive liquid A having a higher refractive index than the laser beam; Characterized in that from the back side and a irradiation step of irradiating via the wafer table.

  The laser dicing apparatus and method of the present invention can be applied regardless of the thickness of the wafer, but is particularly effective for thin wafers of 50 μm or less. Although the lower limit of the wafer thickness was not provided, it is preferably 30 μm or more.

  Further, “flatness on the“ flat holding surface ”does not mean perfect flatness, but refers to flatness that allows the wafer to be in close contact to correct warpage or deflection of the wafer. Furthermore, one surface of the wafer table is a laser light incident surface, and the other surface is a holding surface.

  According to the laser dicing apparatus and method of the present invention, the liquid film of the refractive liquid A having a refractive index larger than that of air is formed in the gap between the condensing lens surface of the laser irradiation means and the laser light incident surface of the wafer table. It is held by interfacial tension.

  Thus, the numerical aperture can be made larger than the liquid film of refractive liquid A having a refractive index larger than that of air exists in the gap between the condenser lens surface and the laser light incident surface of the wafer table. The incident angle of the laser beam incident on the wafer table can be increased, and the refraction angle of the laser beam incident on the wafer table can be increased.

  By reducing the incident angle of the laser light, the laser light emitted from both ends of the condenser lens can be prevented from being totally reflected, and can be made incident inside the wafer via the wafer table, so that the irradiation efficiency can be improved. . By improving the irradiation efficiency, the inside of the wafer can be modified with the energy of a small laser beam, so that ablation hardly occurs.

  In addition, since the refraction angle of the laser beam incident on the wafer table is increased, the laser beam can be collected from a wide range of angles and the focal point can be connected. It becomes easy to tie. Thereby, it can respond to the laser dicing of a thin wafer. Furthermore, by condensing the laser beam from a wide range of angles and connecting the focal points, the irradiation area on the laser beam incident surface of the wafer is increased to reduce the energy per unit area. Melting on the surface can be prevented. On the other hand, the laser beam at the condensing point position inside the wafer can be narrowed down, the irradiation area in the condensing point region can be reduced, and the energy per unit area can be increased. Thereby, melting in the light condensing point region can be promoted. As a result, even a thin wafer, particularly a wafer having a thickness of 50 μm or less, melts only the condensing point region inside the wafer, that is, the region to be modified, so that ablation does not occur.

  Further, the liquid film of the refractive liquid A is held by the interfacial tension only in the gap between the condenser lens surface and the laser light incident surface of the wafer table. As a result, in order to continuously form a modified layer with high precision at a certain depth inside the wafer along the street, the liquid film does not move even if the condenser lens and the wafer table move relatively fast. There is no standing or flirting. Thereby, since there is no change in the refractive index due to the refractive liquid A, it contributes to making the work distance constant.

  Furthermore, according to the laser dicing apparatus and method of the present invention, the wafer to be laser diced is held by a wafer table having a flat holding surface that is formed so as to be able to transmit laser light and that holds the back side of the wafer in close contact. As a result, the warpage and deflection of the wafer are corrected.

  Further, according to the laser dicing apparatus and method of the present invention, the refractive liquid A is made a thin film out of the laser light path from the condenser lens to the wafer, and most of the laser light path is made a solid medium like a wafer table. Therefore, it is possible to extremely reduce the change in the refractive index in the middle of the optical path.

  As mentioned above, by increasing the numerical aperture and increasing the refraction angle, suppressing the refractive index fluctuation in the liquid film, correcting the warpage and deflection of the wafer, and suppressing the refractive index fluctuation in the entire laser beam path Since the laser beam can be irradiated with a constant work distance to the shallow focal point depth position of the wafer with a small laser power, the modified layer can be continuously formed with high accuracy at a constant wafer depth.

  As a result, the laser dicing apparatus and method of the present invention can prevent ablation and make the work distance constant and improve the irradiation efficiency even when laser dicing is performed on a wafer, particularly a thin wafer of 50 μm or less. Therefore, accurate laser dicing can be performed.

  In addition, since the work distance can be made constant by correcting the warpage and deflection of the wafer without destroying the fine-structured device formed on the wafer surface, even a wafer, particularly a thin wafer of 50 μm or less, can be used. The modified layer can be formed accurately inside.

  In the present invention, the shortest distance between the condensing lens surface and the laser light incident surface of the wafer table is preferably in the range of 0.01 to 2 mm. More preferably, although it depends on the operation accuracy of the apparatus, it is preferably about 0.2 mm or less as long as the condensing lens surface and the laser light incident surface of the wafer table do not contact each other. This is because, in this range, the interfacial tension tends to act effectively between the condenser lens, the wafer table, and the refractive liquid A. That is, even if it operates relatively at a high speed, the refractive liquid A follows the condenser lens by the interfacial tension. Here, the “shortest distance” is used because the condensing lens has a convex shape.

  Since the interfacial tension varies depending on the surface tension and viscosity of the refractive liquid A, the material of the condensing lens and the wafer table, and the surface condition, the shortest distance is not limited to 0.01 to 2 mm. It is important to set the gap distance to be generated automatically.

  However, the distance between the condenser lens and the wafer table itself needs to be set constant. The setting of the work distance varies slightly depending on the wavelength of the laser, the refractive index of the material, and the position of the focal point. However, the setting of the thickness of the refracting liquid A should not change greatly. The thickness of the refracting liquid A needs to increase the lubricity while maintaining transparency and follow the mutual movement by interfacial tension even if the condenser lens and the wafer table move at high speed. It is.

  Also, the surface state of the wafer table needs to be kept constant in order to stabilize the movement of the refractive liquid A due to the interfacial tension of the liquid.

  For example, in the technique described in the cited document 4, there are no particular restrictions on a substrate material such as a wafer to be processed by a laser or a film material thereon. It is necessary to perform laser processing on various substrate materials in the same manner. Therefore, under conditions where the refractive liquid is inserted between the condenser lens and the substrate and the substrate is filled with the refractive liquid, the substrate surface is made of a hydrophilic material and the substrate surface is made of a hydrophobic material. In such a case, under the condition that the condenser lens and the substrate operate at a high speed, the tracking ability of the refractive liquid may change greatly, and the refractive index fluctuation may also change greatly. In some cases, when having a hydrophobic surface, there are problems such as repelling the refraction liquid, which may not make sense for the idea of passing light through a stable highly refractive medium. In the case of the present invention, a very small amount of the refracting liquid A is interposed in a state where the condensing lens and the wafer table are made of a transparent solid and a solid space so as to be close to each other. The liquid A is supported on both surfaces, and the stable interfacial tension of the refractive liquid A is skillfully used to ensure a stable light path while lubricating while preventing direct contact between the solids. Further, the wafer table and the condensing lens are made of the same glass and are made of a material having a refractive index close to 1.5, and the refractive liquid A filling the space between them is also close to 1.5. The refractive index can be almost the same from the lens to the wafer table. This constant refractive index ensures a high numerical aperture NA and prevents a loss of light introduction due to reflection in the middle, and increases the prospective angle at the laser focusing point, thereby reducing the thickness of the substrate material such as a wafer. Even if there is, it contributes greatly to making it a condition not to ablate. The “expected angle” is described in the description of the numerical aperture NA, which will be described later.

  Further, it is preferable that the laser light incident surface of the wafer table is roughened. The rough surface treatment increases the substantial surface area of the laser light incident surface of the wafer table, and the interfacial tension between the individual liquids acting on the large surface area also increases. As a result, the liquid film of the refractive liquid A is easily held in the gap between the condensing lens and the laser light incident surface of the wafer table. The refractive liquid A is preferably a liquid containing isopropyl alcohol. Since isopropyl alcohol (IPA) has a lower surface tension than water, it enters all the gaps between the condenser lens and the wafer table, making it easier to hold the refractive liquid A between the condenser lens and the wafer table. .

  In the present invention, the wafer table has a table plate that can transmit laser light and has flat front and back surfaces, and the table plate is formed of any one of quartz glass, germanium, silicon, and plastic resin. Is preferred. In particular, it is preferable to form the wafer table with a material whose refractive index of the laser light path is close to the refractive index of the wafer, such as germanium or silicon. In this case, it is important to select germanium and silicon that can transmit laser light.

  In the present invention, it is preferable that the back side of the wafer is closely held on the holding surface of the wafer table with the refractive liquid B having a refractive index larger than that of air interposed.

  Thus, the wafer is held in close contact with the wafer table by the refractive liquid B, so that the wafer is held in a flat state without bending. Thereby, the work distance can be made more constant. Then, by irradiating the laser beam from the back side of the wafer through the wafer table, the modified layer can be accurately formed inside the wafer. Further, the dicing process can be performed without touching the surface of the wafer at all. Therefore, the dicing process can be performed without destroying the device formed on the surface of the wafer.

  In particular, by introducing and interposing the refractive liquid B in the minute gap between the wafer and the wafer table, the refractive liquid B automatically spreads uniformly in the minute gap due to capillary action. Also, the interface tension acting between the interface between the refractive liquid B and the holding surface of the wafer table, and the interface acting between the refractive liquid and the wafer (the surface on which the tape is adhered when the tape is adhered). Due to the influence of tension, the wafer comes into close contact with the wafer table uniformly without any gaps.

  As a result, when a modified layer is formed inside the wafer by irradiating the laser beam through the wafer table, the laser beam is efficiently imaged and modified inside the wafer without significant energy loss along the path. A layer can be formed.

  Moreover, in this invention, the tape which can permeate | transmit a laser beam may be stuck to the back surface of a wafer, and does not need to be stuck. In the former case, the surface of the tape is generally rough, and scattering tends to occur when air is present. However, when a liquid (refractive liquid B) is interposed as in the present invention, the surface having a rough surface is changed, and the interfacial tension is increased due to the effect of increasing the surface area, and the wettability is increased. Also, since the liquid fills up the unevenness of the fine roughness, the liquid spreads uniformly over the roughened tape surface, and the refractive index distribution is further reduced and scattered. The laser beam can be transmitted more efficiently.

  Further, as the refractive liquid A and the refractive liquid B, water, ethanol, isopropyl alcohol (IPA), or the like can be used. In the present invention, ethanol or IPA is preferably used, but even water (refractive index is 1.3) generally has a higher refractive index than air (refractive index 1) and is used as a refractive liquid. be able to.

  The types of the refractive liquid A and the refractive liquid B may be different, but are preferably the same.

  Although the wafer table transmits laser light, for example, when quartz glass is used as the wafer table, the refractive index is 1.45. Therefore, liquid rather than gas is used for these wafer tables and tapes. On the other hand, the refractive index has a close value. When the refractive index is close, reflection and scattering are reduced at that interface, and energy loss when irradiating laser light can be greatly reduced.

  In addition, if a liquid having a refractive index equivalent to that of a wafer table or a tape is used as the refractive liquid A and the refractive liquid B, the loss of the laser beam at the interface is almost eliminated, and the laser beam is efficiently transmitted to the inside of the wafer. A quality layer can be formed.

  As described above, according to the present invention, a modified layer is accurately formed at a certain depth inside the wafer while irradiating the laser beam from the back surface of the wafer without destroying the device formed on the front surface of the wafer. In addition, an improved and stable modified layer is formed inside the wafer by reducing scattering in the middle path of the laser beam, and laser burning due to scattering of the laser beam is eliminated by scattering in the middle path. It becomes possible to prevent deterioration of components and performance.

  The refractive liquid B is preferably a liquid containing isopropyl alcohol. As a result, it is possible to suppress the generation of a watermark on the wafer, and for the same reason as described in the refractive liquid A, it is possible to make the back surface of the wafer and the wafer table more evenly adhered. .

  In addition to the uniform contact between the back surface of the wafer and the wafer table, and the uniform liquid between them, the refractive index variation in the laser beam path can be reduced. There is almost no in-plane uniform state.

  In the present invention, the holding surface of the wafer table is preferably roughened. As a result, the wafer can be held in close contact with a large contact force. That is, the rough surface treatment of the wafer table increases the substantial surface area of the wafer table and increases the interfacial tension between the individual liquids that act on the large surface area. Therefore, the back surface of the wafer (the surface to which the tape is applied when the tape is applied) and the wafer table are more closely adhered to each other, and there is no difference in refractive index between the back surface of the wafer and the wafer table. It becomes possible to form a modified layer by light inside the wafer.

  In the present invention, the wafer table is installed horizontally with the holding surface facing down, and the laser irradiation means irradiates the wafer held on the wafer table with laser light from above the wafer table. Is preferred.

  In the present invention, since the refractive liquid A needs to be held by the interfacial tension in the gap between the condenser lens and the wafer table, a wafer table having a larger area than the condenser lens is disposed below the condenser lens. It is easier to hold the refractive liquid A.

  In the present invention, it is preferable that a laser antireflection plate is installed at a position separated from the holding surface.

  As a result, unnecessary reflection of the laser light transmitted through the wafer can be prevented, and problems such as reflection burn can be effectively prevented.

  In order to achieve the above object, the wafer processing method of the present invention has a plurality of devices formed on the front surface and a laser beam applied to the wafer mounted on the frame via a tape attached to the back surface. A laser dicing step of irradiating and forming a modified layer inside the wafer, and an expanding step of dividing the tape of the wafer subjected to the laser dicing treatment into individual chips on the tape. In the wafer processing method which has, in the said laser dicing process, the laser dicing method of any one of Claim 10 to 17 is used.

  According to the present invention, since the laser dicing method according to any one of claims 10 to 17 is used in the laser dicing process, the laser dicing is performed on a wafer, particularly a thin wafer of 50 μm or less. Even so, accurate laser dicing can be performed. Therefore, even when the wafer is divided into individual chips in the subsequent expanding process, the wafer can be divided with high accuracy.

  According to the present invention, even when laser dicing is performed on a wafer, particularly a thin wafer having a thickness of 50 μm or less, ablation can be prevented, irradiation efficiency can be improved, and the work distance can be made constant. Good laser dicing can be performed.

  Further, according to the present invention, since the work distance can be made constant by correcting the warpage and deflection of the wafer without destroying the fine-structured device formed on the wafer surface, the wafer, particularly a thin film having a thickness of 50 μm or less. Even in the case of a wafer, the modified layer can be accurately formed inside the wafer.

1 is a schematic configuration diagram showing an embodiment of a laser dicing apparatus according to the present invention. Sectional drawing in which the modified layer is formed on the wafer by the laser dicing apparatus according to the present invention Enlarged view of the main part provided with a surrounding member for confining refractive liquid around the condenser lens A perspective view showing a wafer mounted on a dicing frame Illustration that thin wafers are likely to ablate Schematic diagram explaining the relationship between ablation and angle of view Explanatory drawing explaining the difference of the laser optical path by the refractive liquid A and air Schematic configuration diagram showing another embodiment of a laser dicing apparatus Explanatory drawing which shows one aspect | mode of the rough surface process of the holding surface of a wafer table Explanatory diagram of expanding process How to create a wafer unit Wafer unit peeling method

  Preferred embodiments of a laser dicing apparatus and method and a wafer processing method according to the present invention will be described below with reference to the accompanying drawings.

[Laser dicing equipment]
FIG. 1 is a schematic configuration diagram showing an embodiment of a laser dicing apparatus according to the present invention. FIG. 2 is a sectional view in which a modified layer is formed on the wafer by the laser dicing apparatus according to the present invention.

  As shown in FIGS. 1 and 2, the laser dicing apparatus 10 according to the present embodiment mainly includes a wafer table 20 that holds a wafer W and laser irradiation that makes laser light incident on the wafer W held on the wafer table 20. Interfacial tension is applied only to the gap between the device 60 and the lens surface 66A (laser beam emitting surface) of the condenser lens 66 of the laser irradiation device 60 and the table surface 22B (laser beam incident surface) of the table plate 22 of the wafer table 20. And a liquid film 26A formed of a refractive liquid A having a refractive index larger than that of air, and a laser antireflection plate 50.

  First, the wafer W to be processed by the laser dicing apparatus 10 of the present embodiment will be described.

  The wafer W to be processed by the laser dicing apparatus 10 according to the present embodiment is a wafer W having a plurality of devices (for example, MEMS elements) formed on the surface, and is processed while mounted on the dicing frame F. The

  FIG. 3 is a perspective view showing the wafer W mounted on the dicing frame F. FIG.

  As shown in FIG. 3, the wafer W is mounted on the dicing frame F via the dicing tape T. The dicing frame F is formed in a frame shape, and a dicing tape T is attached to the inside of the dicing frame F. The wafer W is mounted on a dicing frame F with its back surface attached to a dicing tape T.

  Here, the dicing tape T attached to the dicing frame F is formed of a material having ductility, and is formed of a material capable of transmitting laser light emitted from the laser irradiation device 60. In this example, it is made of a transparent material having ductility. Further, it is more preferable to have air permeability.

  The wafer W mounted on the dicing frame F is held by the wafer table 20 on the surface (back surface) to which the dicing tape T is attached.

  As shown in FIG. 1, the wafer table 20 mainly includes a table plate 22 and a table plate holding frame 24 that holds the table plate 22.

  The table plate 22 is formed in a disk shape corresponding to the wafer W to be processed, and both the front surface and the back surface thereof are formed flat. The table plate 22 is formed in a disk shape having a larger diameter than the wafer W to be processed so that the entire surface of the wafer W to be processed can be supported. The table plate 22 is formed of a material that can transmit laser light emitted from the laser irradiation device 60. As an example, in the present embodiment, it is made of transparent quartz glass. The table plate 22 is formed with a necessary and sufficient thickness so that the wafer W to be processed can be held without bending.

  The table plate 22 has a lower surface in FIG. 1 as a holding surface 22A for holding the wafer W. Between the holding surface 22A and the surface (rear surface) of the wafer W on which the dicing tape T is adhered, a uniform and thin film liquid film 26B of the refractive liquid B is interposed, and the wafer W passes through the liquid film 26B. Are held in close contact with the table plate 22. As will be described in detail later, examples of the refractive liquid B include water, ethanol, isopropyl alcohol (IPA), and the like.

  The table plate holding frame 24 is formed in an annular shape. The table plate 22 is held on the inner periphery of the table plate holding frame 24.

  The table plate holding frame 24 is rotated about its axis by a rotation drive mechanism (not shown), and is moved up and down (Z direction) by a lift drive mechanism (not shown). Moreover, while moving in the front-rear direction (Y direction) on the horizontal plane by a front-rear drive mechanism (not shown), it moves in the left-right direction (X direction) on the horizontal plane by a left-right drive mechanism (not shown).

  Thereby, the table board 22 rotates around an axis | shaft, when the table board holding frame 24 rotates. Further, the table plate holding frame 24 moves up and down as it moves up and down. Further, the table plate holding frame 24 moves back and forth when moving in the front-rear direction, and moves left and right when moved in the left-right direction.

  The laser irradiation device 60 is installed above the wafer table 20 and emits laser light vertically toward the wafer table 20.

  As shown in FIG. 2, the laser irradiation device 60 mainly includes a laser oscillation device 62, a collimator lens 64, a condenser lens (condenser lens) 66, and an actuator 68.

  The gap between the lens surface 66A (laser beam emitting surface) of the condenser lens 66 and the table surface 22B (laser beam incident surface) of the table plate 22 is held by the interfacial tension and has a refractive index higher than that of air. A liquid film 26A formed of a large refractive liquid A is interposed.

  The table surface 22B of the table plate 22 is preferably roughened. As the rough surface treatment, for example, GC abrasive grain # 2000 may be used, and the table surface 22B may be lapped for about 20 minutes. Moreover, you may use the abrasive grain about # 500. Other than GC, the table surface 22B may be roughened using abrasive grains such as WA. By doing so, the table surface 22B becomes ground glass and is scattered in the air due to surface roughness and becomes cloudy. The surface roughness may be 0.1 mm or less in terms of Ra, but is not limited to this, and it is better that the liquid is buried in the roughness gap so that the interfacial tension is increased and the surface area is increased.

  By subjecting the table surface 22B of the table plate 22 to the rough surface in this way, a minute gap space is formed between the condenser lens 66 and the table plate 22, and the refraction liquid A is caused by capillary action. It spreads uniformly and efficiently in the gap. As a result, it becomes easy to hold the refractive liquid A firmly in the gap between the condenser lens 66 and the table plate 22 with interfacial tension. Although the refractive liquid A will be described in detail later, water, ethanol, isopropyl alcohol (IPA) or the like can be used in the same manner as the refractive liquid B described above.

  In addition, the refracting liquid A and the refracting liquid B can be of different types, but are preferably the same.

  The laser oscillator 62 oscillates laser light in accordance with the processing conditions of the wafer W. The laser light oscillated from the laser oscillation device 62 is collimated by the collimator lens 64 and then condensed at the condensing point P by the condenser lens 66. As a result, when the condensing point P is set inside the wafer W and laser light is incident on the wafer W, a modified region is formed inside the wafer W. When the wafer W is moved horizontally in this state, the modified region is continuously formed along the movement locus of the condensing point P, and the modified layer L is formed. At the time of dividing the wafer, this modified layer is formed along the street (division line).

  The actuator 68 slightly moves the condenser lens 66 in the optical axis direction (Z-axis direction). That is, the condensing lens 66 is held by a lens frame (not shown) and is supported so as to be movable in the optical axis direction. The condenser lens 66 is driven by the actuator 68 and moves minutely in the optical axis direction.

  By driving the actuator 68 and moving the condensing lens 66 in the optical axis direction, the position of the condensing point P of the laser light is displaced in the Z direction. Thereby, the position (position in the Z direction) where the modified layer L is formed can be adjusted. In addition, a plurality of modified layers L can be formed inside the wafer W by changing the position of the condensing point P in the Z direction and making the laser light enter the wafer W a plurality of times.

  In addition to adjusting the position of the condensing point P in the Z direction, the distance of the gap between the lens surface 66A of the condensing lens 66 and the table surface 22B of the table plate 22 is adjusted. Adjust so that it is held in the gap by tension. That is, the surface tension of the condensing lens 66, the table plate 22, and the refracting liquid A, the interfacial tension between the condensing lens 66 and the refracting liquid A, the table plate 22 and the refracting liquid A, the viscosity of the refraction liquid A, and the like. Although it depends on the physical properties, the shortest distance D (see FIG. 2) between the lens surface 66A of the condenser lens 66 and the table surface 22B of the table plate 22 is set to be in the range of 0.01 to 2 mm.

  When the position of the condensing lens 66 is adjusted so that the condensing point P is at a predetermined position, if the gap distance exceeds 1 mm, the adjustment can be made by increasing the thickness of the table plate 22. . A thickness of the table plate 22 may be determined in advance by performing a preliminary test.

  The laser antireflection plate 50 is installed below the wafer table 20. This laser antireflection plate 50 is formed in a flat plate shape having an antireflection treatment (for example, black treatment) on the upper surface portion, and is horizontally installed at a position away from the holding surface 22A of the wafer table 20 by a predetermined distance. Is done. That is, a predetermined space is formed between the holding surface 22 </ b> A of the wafer table 20 and the laser antireflection plate 50.

  The laser light emitted from the laser irradiation device 60 and transmitted through the wafer W enters the laser antireflection plate 50. Thereby, unnecessary reflection can be prevented and occurrence of reflection burn can be prevented.

  The laser dicing apparatus 10 of the present embodiment is configured as described above.

  The operation of the laser dicing apparatus 10 is controlled by a control device (not shown). The control device executes a predetermined control program, controls the operation of each unit, and executes the processing of the wafer W.

[Laser dicing method]
Next, a laser dicing method using the laser dicing apparatus 10 configured as described above will be described.

  As described above, the wafer W is processed while mounted on the dicing frame F. The wafer W is mounted on the dicing frame F by attaching the back surface (the surface on which no device is formed) to the dicing tape T.

  The wafer W mounted on the dicing frame F is transferred to the lower part of the wafer table 20 by a transfer device (not shown) (for example, an arm of a robot). At this time, the wafer W is transported to the lower position of the wafer table 20 with the surface on which the dicing tape T is adhered facing upward, and the surface is in a non-contact state.

  The wafer W transferred to the lower position of the wafer table 20 is delivered from the transfer device to the wafer table 20. The delivery is performed by tightly holding the back surface of the wafer W with the holding surface 22A of the wafer table 20. Specifically, this is performed as follows.

  First, the wafer W is positioned. That is, the wafer W is positioned so that the center of the wafer W coincides with the center of the wafer table 20. Thereafter, the refracting liquid B is supplied uniformly to the back surface of the wafer W using a refracting liquid B supplying means (not shown) for dropping or applying the refracting liquid B to form a liquid film 26B. The wafer W may be positioned after the refractive liquid B is supplied to form the liquid film 26B. Further, the refractive liquid B may be supplied to the holding surface 22A of the wafer table 20 in place of the back surface of the wafer W or together with the back surface of the wafer W to form a liquid film 26B on the holding surface 22A. Thereafter, the holding surface 22A of the wafer table 20 is pressed against the back surface of the wafer W with a predetermined pressure. As a result, the refractive liquid B is uniformly and thinly spread between the back surface of the wafer W and the holding surface 22A of the wafer table 20, and the wafer W is held in close contact with the wafer table 20 via the liquid film 26B. Is done.

  In this way, the holding surface 22A of the wafer table 20 is held in close contact with the surface of the wafer W on which the dicing tape T is adhered with the liquid film 26B of the refractive liquid B interposed. Thereby, the wafer W can be held in a flat state without being bent. Further, since the back surface of the wafer W is held in close contact with the dicing tape T, the front surface of the wafer W can be held in a non-contact state with a solid object (other than air). Thereby, a plurality of devices (for example, MEMS elements) formed on the wafer surface are not damaged.

  The transfer device that has delivered the wafer W retracts from the lower portion of the wafer table 20. Thereafter, predetermined preprocessing and alignment processing are performed, and laser dicing is started.

  Laser dicing is performed by making the laser beam emitted from the laser irradiation device 60 enter the wafer W along the street.

  However, the conventional laser dicing has a problem that ablation is likely to occur when the wafer W is thin, particularly when the wafer W is 50 μm or less.

  Here, the reason why ablation is likely to occur in the case of a thin wafer W will be described with reference to FIG. FIG. 4 shows two relations between the condenser lens 66 and the wafer W, and the wafer table 20 is omitted.

  4A shows laser dicing in an ideal state, and FIG. 4B shows actual laser dicing. FIG. 4 (i) is a cross-sectional view showing the laser beam path, (ii) is a three-dimensional schematic diagram of the focused state of the laser beam, and (iii) is a cross-section showing the melted state of the wafer by the laser beam. FIG.

  As shown in (i) and (ii) of FIG. 4A, the laser dicing in the ideal state is such that the laser light emitted from the condensing lens 66 is condensed into a pinpoint (dot) inside the wafer W. Thus, the condensing point P is connected. Thereby, as shown to (iii) of FIG. 4 (A), the position of the condensing point P is pinpointed and it will be in a modification | reformation state.

  However, the actual laser dicing does not become an ideal state, and as shown in (i) of FIG. 4 (B), the laser optical path proceeds linearly to the vicinity of the set condensing point P, A curved optical path is formed in the vicinity of the condensing point position, and proceeds linearly again. As a result, as shown in (ii) of FIG. 4B, the laser light is condensed with an elliptical area without being focused at the focal point position. As a result, the area difference between the irradiation area S1 on the laser light incident surface (wafer back surface in this embodiment) of the wafer W and the irradiation area S2 in the condensing point region inside the wafer becomes small. When the area difference is small as described above, when melting occurs in the condensing point region inside the wafer, energy absorption of the laser light occurs in a chain manner up to the laser light incident surface a of the wafer W. As a result, as shown in (iii) of FIG. 4B, so-called ablation in which a continuous hole is formed in the wafer W from the focal point region to the laser beam incident surface a occurs. When ablation occurs, the molten material scatters around and adheres to the wafer W. Thereby, the adhering wafer portion becomes a defective product.

  By the way, the length of the curved laser beam path is called the Rayleigh length (L). Even in the case of a YAG laser beam having a wavelength of 1064 mm which is a high quality laser beam, the Rayleigh length (L) is about 20 μm. become.

  Thereby, when the thickness of the wafer W is thin, the difference between the wafer thickness and the Rayleigh length (L) becomes small, so the area difference between the irradiation area S1 and the irradiation area S2 becomes small. As a result, the difference between the melting energy at the laser light incident surface a of the wafer W and the melting energy at the condensing point region inside the wafer becomes small, so that ablation is likely to occur. For this reason, it is said that the limit of the thickness of a wafer that can be laser-diced is about 30 μm.

  Therefore, in order to prevent ablation and perform laser dicing with high accuracy on a thin wafer W, particularly a wafer W of 50 μm or less, it is necessary to reduce the influence of the Rayleigh length (L) and bring it closer to the ideal state. The following three conditions are important in order to approach the ideal state.

  (Condition A) By converging the laser light from a wide angle by converging the converging point by increasing the prospective angle, the light is collected from the laser light incident surface a (wafer back surface in this embodiment) of the wafer W into the wafer. The light spot depth is made shallower, and the area difference between the irradiation area S1 and the irradiation area S2 is increased. For this purpose, a large refraction angle is required to rapidly narrow down from a wide area in the process of converging (condensing) from when the laser beam is emitted until it reaches the condensing point in the wafer. A number NA is required.

  (Condition B) The work distance is made constant so that the focal point depth does not fluctuate. In the case of a thin wafer, variation in the focal point depth due to slight variation in work distance causes ablation. In other words, in the case of a thin wafer, not only a large refraction angle is required, but also the refractive index is required not to fluctuate in the optical path from when the laser beam is emitted until it reaches the condensing point in the wafer. The

  (Condition C) When the thickness of the wafer is thin, the depth of the condensing point of the laser beam inside the wafer has to be reduced, and therefore it is necessary to make the laser beam incident from a wide range of angles. Therefore, even if the laser beam is condensed from a wide angle, the laser beam emitted from the end of the condenser lens is prevented from being reflected at the interface.

  Here, as a further detailed description of the conditions A to C described above, the relationship between the occurrence of ablation and the prospective angle related to the aperture ratio NA will be described.

  That is, the influence of the numerical aperture NA is an important factor for preventing ablation.

  Here, the numerical aperture NA is expressed by the following equation, where θ is the maximum angle with respect to the optical axis of the light beam emitted from the condenser lens 66 to the object, and n is the refractive index n of the medium between the object and the condenser lens. Is done.

NA = n * sinθ
Usually, in the case of air, since the refractive index is 1, the numerical aperture NA is theoretically 1 at the maximum, but in reality, the limit is about 0.95. That is, if the light medium from the laser light condensing lens is air, the numerical aperture NA will not be 1 or more. Conventionally, in a condensing laser beam, the medium through which the laser beam passes is through air, so that laser dicing is naturally performed under the condition of a low numerical aperture NA.

  As shown in FIG. 5A, as an item related to the numerical aperture NA, 2θ obtained by doubling θ is called a prospective angle, and is an elevation angle viewed from the focal point P. When the prospective angle is large, the laser beam is narrowed down from a wide angle, and the area of the condensing point can be made as small as possible relative to the irradiation area on the wafer surface (on the laser beam incident surface of the wafer) that irradiates the laser beam. .

  On the other hand, when the viewing angle is small, the area of the condensing point may not change so much with respect to the irradiation area on the wafer surface irradiated with laser light.

  In relation to the numerical aperture NA as described above, a large numerical aperture NA corresponds to a large prospective angle.

  Here, according to the mechanism in which ablation occurs, a substrate material such as silicon has a temperature dependency in the absorption coefficient of laser light.

  FIGS. 5B and 5C are diagrams illustrating how the ablation differs due to the temperature dependence of the absorption coefficient of the laser light in relation to the prospective angle.

  FIG. 5B shows an event that occurs in the wafer W when the prospective angle is small. In laser dicing, the irradiation time with respect to the wafer W by a pulse laser is usually about 100 ns.

  In the irradiation time within 100 ns, the energy of the laser beam first condensed inside the wafer is first absorbed most at the condensing point P. Since the energy of the laser beam is locally consumed at the most concentrated point, the internal reforming occurs explosively at that point.

At the next moment, when the prospective angle is small, a large amount of energy is consumed in the upper part of the condensing point P in the vicinity of the condensing point and the temperature rises rapidly. As a result, the energy absorption of the laser beam increases in the vicinity immediately above the condensing point P, and a little time is delayed from the internal reforming of the condensing point P, so that the internal reforming occurs immediately above the condensing point P. Subsequently, the internal reforming occurs in sequence toward the upper part in sequence so that the temperature of the upper part rises and the laser energy is absorbed and the internal reforming continues. That is, since rapid laser energy absorption accompanying a rapid temperature rise occurs in a chain from the initial light condensing point P toward the upper part, the chain reaction finally occurs on the surface of the wafer W (laser light incident surface). When it reaches the hole, the surface portion of the wafer blows away. This is the ablation mechanism. (Reference: Junji Omura, Japan Society for the Promotion of Science 145th Committee, 125th Research Meeting P.9)
What is important for preventing ablation due to the mechanism described above is the prospective angle at the condensing point P of the laser beam. That is, when the prospective angle is small, the area of the internal reforming area of the condensing point P is not much different from the laser irradiation area immediately above it. As a result, the temperature immediately above the condensing point P immediately rises and the laser energy is absorbed suddenly, which promotes energy absorption in the upper part in a chain. The reforming proceeds.

  On the other hand, FIG. 5C shows an event that occurs in the wafer W when the prospective angle is large.

  When the prospective angle is large at the condensing point P of the laser beam, even if internal reforming occurs at the condensing point P, the laser irradiation area immediately above it suddenly increases, so the temperature does not increase so rapidly. Therefore, rapid energy absorption in the vicinity of the condensing point P does not occur, and only the vicinity of the condensing point P is modified. As a result, it is possible to prevent chain internal reforming from the focal point P of the wafer toward the wafer surface.

  According to such an ablation generation mechanism, when a thinner wafer W is processed without being ablated, it is important to increase a prospective angle for laser incidence on the wafer W. If the prospective angle is increased as much as possible, it becomes possible to prevent the ablation phenomenon that occurs when chain laser absorption reaches the wafer surface.

  Further, the focal point depth (d) from the wafer surface where the laser beam is focused is proportional to the wavelength λ and inversely proportional to the square of the numerical aperture NA as shown in the following equation.

d = λ / NA ²
Accordingly, it is important to increase the numerical aperture NA in order to reduce the focal point depth d.

  The size of the focal point depth d also affects ablation. In other words, when the elliptical sphere portion that is the range connecting the condensing points has an elongated shape in the depth direction, modification is likely to occur in a chain, but when the elliptical sphere is compact in the depth direction, The region is limited to a part of the local area, and only local modification is required.

  Also, resolution is important in order to concentrate large energy only in a local area inside the wafer. The resolution δ is expressed by the following equation.

δ = 0.61 * λ / NA
The resolution δ is proportional to the wavelength λ and inversely proportional to the numerical aperture NA. In order to reduce the resolution δ, it is better to increase the numerical aperture NA. Also in this respect, it is possible to increase the numerical aperture NA by using a lens having a high numerical aperture NA by interposing a medium having a high refractive index instead of air. Contribute to concentrating on.

  Therefore, in the laser dicing method of the present embodiment, the liquid film 26A of the refractive liquid A having a refractive index larger than that of air is interposed between the lens surface 66A of the condenser lens 66 and the table surface 22B of the table plate 22. In this way, the numerical aperture (NA) was increased.

  That is, the control device (not shown) drives the actuator 68 to adjust the position of the condensing lens 66 so that the condensing point P is set at a predetermined position inside the wafer W, and after the adjustment. The distance between the lens surface on the output side of the condenser lens 66 and the surface of the table plate 22 is set to be in the range of 0.01 to 2 mm.

  Then, the refracting liquid A is supplied into the gap of 0.01 to 2 mm by capillary action, and the liquid film 26A of the refracting liquid A is formed only in the gap. In this case, after the droplet of the refractive liquid A is dropped on the table plate 22 just below the condensing lens, the condensing lens 66 and the table plate 22 are moved by moving the condensing lens 66 closer to the table plate 22 side. A liquid film 26A may be formed between the two.

  After the above preparation is completed, a laser beam is emitted, and the wafer table 20 is moved so that the emitted laser beam is incident on the wafer W along the street.

  FIG. 6 shows, as a comparison, a case in which a liquid film 26A of refractive liquid A is interposed in the gap between the lens surface 66A of the condenser lens 66 and the table surface 22B of the table plate 22 (right side of the imaginary line in the center of FIG. 6). 7 is a diagram comparing laser light paths when air is interposed between a lens surface 66A of a condensing lens 66 and a table surface 22B of a table plate 22 (left side of an imaginary line in the center of FIG. 6).

  In FIG. 6, the refractive liquid A having a refractive index of 1.5 is used. In addition, the condenser lens 66 (refractive index 1.5), table plate 22 (refractive index 1.5), refractive liquid B (refractive index 1.5), dicing tape T (refractive index 1.5), and wafer W (Refractive index 3.6) is common to the left and right sides of the imaginary line in the center of FIG.

  The refraction angle at the interface having a different refractive index is expressed by the following equation.

n ₁ * sin θ1 = n ₂ * sin θ2
In order for light to reach the inside of the material by refraction, the influence of the refractive index is large.

As can be seen from FIG. 6, when air is interposed between the condenser lens 66 and the table plate 22, the laser light emitted from the end of the condenser lens 66 is large at the interface between the condenser lens 66 and air. Refract. As a result, the incident angle θ ₁ to the table plate 22 is increased, so that it is totally reflected at the interface between the air and the table plate 22 and is not incident on the table plate 22.

  On the other hand, when the refractive liquid A is interposed between the condenser lens 66 and the table plate 22, the laser light emitted from the end of the condenser lens 66 is the interface between the condenser lens 66 and the refractive liquid A. Refracts smaller than air. As a result, the incident angle θ <b> 1 to the table plate 22 is reduced, so that the incident light enters the table plate 22 without being totally reflected at the interface between the air and the table plate 22.

  When air is interposed between the condenser lens 66 and the table plate 22, the refraction angle θ2 of the laser light incident on the table plate 22 is smaller than the incident angle θ1 on the table plate 22. Therefore, the laser beam enters the wafer W while standing. As a result, the laser beam cannot be collected from a wide angle with respect to the wafer W.

  On the other hand, when the refracting liquid A is interposed between the condenser lens 66 and the table plate 22, the refraction angle θ2 of the laser light incident on the table plate 22 is the incident angle on the table plate 22. Since it becomes larger than θ1, the laser beam enters the wafer W in a state where it lies down. In the case of FIG. 6, since the refractive indexes of the condenser lens 66, the table plate 22, the refracting liquid B, and the dicing tape T are all the same at 1.5, the laser light emitted from the condenser lens 66 is linear. A laser beam path is formed.

  As a result, as can be seen from FIG. 6, when the refractive liquid A is interposed between the condensing lens 66 and the table plate 22, the depth of the condensing point P1 where the laser light is condensed inside the wafer is It can be made shallower than the depth of the condensing point P <b> 2 when air is interposed between the condensing lens 66 and the table plate 22.

  As described above, even if the laser light is emitted from the converging lens 66 at the same angle, the effective angle is limited due to the influence of total reflection in the midway path when air is interposed midway, and the narrow area Only laser light enters the material. Therefore, the prospective angle becomes small.

  On the other hand, when the light passes through the medium having the same refractive index from the condenser lens 66 and finally enters the material, the influence of total reflection can be eliminated, and the laser beam can be narrowed down from a wide angle. The corner gets bigger. As a result, even a thin wafer W does not ablate.

  In this embodiment, not only the refractive liquid A having a refractive index larger than that of air is interposed between the condenser lens 66 and the table plate 22, but also the table plate 22, the dicing tape T, and the wafer W. A medium having a high refractive index was used for all of the refracting liquid B for closely contacting the back surface of the substrate to the table plate 22. Thus, a high numerical aperture is ensured for all of the laser light paths from the condenser lens 66 to the wafer W.

  As described above, the optical path from when the laser beam emitted from the condensing lens 66 reaches a set condensing point in the wafer is a medium having a high refractive index higher than that of air. Since the laser beam from the angle can be condensed, the focal point can be connected to the shallow focal point depth position from the laser beam incident surface a of the wafer W.

  A condensing lens that achieves a high magnification and a high numerical aperture in such a high refractive medium is also called a solid immersion lens.

  For example, as such a lens, an immersion condensing lens as disclosed in Japanese Patent Application Laid-Open No. 2004-061589 is used as a lens for a microscope, but can be applied in terms of achieving a high numerical aperture. In addition, Olympus silicon immersion condenser lens UPLSAPO30X and the like are condenser lenses for biological microscopes, but can be applied as the condenser lens of the present application in that a high numerical aperture can be secured.

  In the conventional condensing lens in which air is the medium, it is theoretically impossible to take a numerical value of 1 or more as the numerical aperture of the lens, but the medium has a refractive index equivalent to that of the glass of the condensing lens 66. By doing so, a large numerical aperture NA of 1 or more can be obtained.

  In the present embodiment, since the thin liquid film 26A of the refractive liquid A is held only by the interfacial tension only in the gap between the condenser lens 66 and the table plate 22, as described in Patent Document 4. There is no change in the work distance accompanying the change in the refractive index of the refractive liquid A.

  That is, since the upper surface and the lower surface of the thin liquid film 26A of the refractive liquid A are held in contact with the two solid media of the condenser lens 66 and the table plate 22, the refractive liquid A is separated from the condenser lens 66. It is stably held between the table plate 22.

  In this case, as shown in FIG. 2A, a cylindrical enclosing member 69 that surrounds the vicinity of the condenser lens 66 and extends to a position where the tip of the condensing lens 66 is close to the table plate 22 (does not contact the table plate). It is preferable to provide it. Accordingly, the refraction liquid A can be reliably held between the condensing lens 66 and the table plate 22 in combination with the above-described interfacial tension regardless of the shape of the condensing lens 66. Therefore, even if the wafer table 20 is moved in the horizontal direction at high speed in order to irradiate the laser along the street, the refractive liquid A can be confined in the narrow space of the enclosing member 69. There is nothing that can be left behind.

  Further, as described above, the liquid film 26A of the refractive liquid A is held in the state where it is in contact with the two solid media of the condenser lens 66 and the table plate 22, and thus is stored in the immersion tank of Patent Document 4. It does not have a free liquid surface like refractive liquid. Therefore, even if the wafer table 20 is moved at a high speed in the horizontal direction in order to irradiate the laser along the street, the liquid film 26A will not be waved or swayed. Further, the liquid film 26 </ b> A of the refractive liquid A also serves as a lubricant that smoothes the relative movement between the condenser lens 66 and the table plate 22.

  As a result, even when the wafer table 20 is moved at high speed in the horizontal direction, the refracting liquid A does not swell and fluctuations of the refracting liquid A are also suppressed in the gap between the condenser lens 66 and the table plate 22. It remains stably held by the acting interfacial tension.

  Here, in Japanese Patent Application Laid-Open No. 2007-136482 of Patent Document 4, in order to improve the laser irradiation efficiency, that is, to reduce the reflectance, water is put between the laser emitting portion and the substrate surface (wafer surface). Yes. However, the purpose of the problem is only to increase the irradiation efficiency of laser light. The problem of processing without causing ablation in laser dicing of a thin wafer is a problem unique to the present application, and Patent Document 4 does not set such a problem.

  In the present invention, all of the refraction liquid A is in contact with the condensing lens 66, and even if the refraction liquid A is heated by the heat of the condensing lens 66 due to laser light irradiation, the refraction liquid A is refracted as in Patent Document 4. A temperature difference hardly occurs in the liquid A.

  As described above, the thin liquid film 26A of the refractive liquid A is held only by the interfacial tension only in the gap between the condenser lens 66 and the table plate 22, so that the refractive liquid A is waved, fluctuated, or convected. This can be effectively suppressed. Thereby, since the refractive index of the refractive liquid A does not fluctuate, the work distance can be made constant.

  The gap between the condenser lens 66 and the table plate 22 is preferably as narrow as possible, but the condenser lens 66 and the table plate 22 move relatively fast to irradiate the laser along the street. Therefore, a gap distance of 0.1 mm or more where the condensing lens 66 and the table plate 22 do not contact each other is necessary, and it is preferably 1 mm or less so that the refractive liquid A is held in the gap with interfacial tension. The condensing lens 66 and the table plate 22 are arranged so as to have a gap of 0.1 mm or more and 1 mm or less, and a minute amount of the refractive liquid A is inserted into the gap to form a liquid film. Contact with the table plate 22 during high-speed movement can be prevented, and the numerical aperture for laser irradiation can be increased.

  Further, in order to make the work distance constant, it is important that there is no change in the refractive index in the laser light path from the condenser lens 66 to the wafer W as well as the liquid film 26A of the refractive liquid A. In this respect, in the embodiment of the present invention, most of the laser light path is formed by the table plate 22 (for example, quartz glass) which is a solid medium, so that the work distance can be made more constant. That is, the solid medium can extremely reduce the fluctuation of the refractive index in the middle of the laser optical path, and can ensure a laser transmission state having a stable refractive index with good reproducibility.

  As described above, in the embodiment of the present invention, since the refractive index variation is prevented as much as possible in the laser light path from the condenser lens 66 to the wafer W and the work distance is made constant, Even when the thickness is 50 μm or less, laser dicing can be performed with high accuracy while suppressing ablation.

  As described above, the variation of the work distance is caused not only by the variation of the refractive index in the laser optical path but also when the wafer W is thinly warped or bent.

  However, in the present embodiment, the wafer W is in close contact with the holding surface 22A of the wafer table 20 by interposing the liquid film 26B of the refractive liquid B, and is held on the holding surface 22A without warping or bending. , Work distance can be fixed. Thereby, the modified layer L can be accurately formed along the street.

  In addition, since the wafer W is held on the wafer table 20 without touching the surface, it can be processed without destroying a device (such as a MEMS element) formed on the surface.

  In addition, if the surface of the wafer W is held tightly, the holding surface 22A is burned by the laser light transmitted through the wafer W, or a melt adheres, and the smoothness of the holding surface 22A cannot be maintained. By holding the wafer W in close contact without touching the surface, it is possible to prevent such a problem from occurring. Thereby, even if it processes continuously, the wafer W can always be closely_contact | adhered and hold | maintained.

  Furthermore, since the laser light is incident on the back surface of the wafer W, the modified layer L can be accurately and reliably formed at a predetermined position without being affected by the device formed on the front surface.

  When the laser beam is incident on the wafer W along the street and the processing is completed, the wafer W is transferred from the wafer table 20 to the transfer device, and is transferred to the next process by the transfer device.

<First and refractive liquid B>
Next, the refractive liquid A and the refractive liquid B used in the present embodiment will be described.

  In the present embodiment, as described above, the refractive liquid A is held as the liquid film 26 </ b> A in the gap between the condenser lens 66 and the table plate 22 of the wafer table 20.

  Further, the wafer W is held in close contact with the holding surface 22A of the wafer table 20 with the liquid film 26B of the refractive liquid B interposed on the surface (back surface) to which the dicing tape T is adhered, and the laser beam is emitted from the wafer W. The light enters from the back side of W through the refractive liquid B.

  For this reason, the liquid used as the refracting liquid A and the refracting liquid B may be any liquid that can transmit at least laser light. For example, water, ethanol, isopropyl alcohol (IPA), or the like can be used. Of these liquids, ethanol and IPA are preferred. Ethanol and IPA have a lower surface tension than water and high wettability with respect to the condenser lens 66, the table plate 22, and the dicing tape T. For this reason, these liquids (that is, ethanol and IPA) are not formed into droplets when they are interposed in the gap between the condensing lens 66 and the table plate 22, but spread in a thin film shape, and the condensing lens 66 and the table plate are spread. 22 is held. Thereby, the refracting liquid A can be held between the condenser lens 66 and the table plate 22 in a highly adhesive state.

  Further, these liquids (that is, ethanol and IPA) are uniformly and thinly formed between the surface (rear surface) of the wafer W on which the dicing tape T is adhered and the holding surface 22A of the wafer table 20 as a whole. Spread out wet. As a result, the refractive liquid B can hold the wafer W in a tightly adhered state.

  In particular, in the present embodiment, an embodiment using IPA as the refractive liquid B is suitable. IPA has higher volatility than other liquids, and can suppress the occurrence of a watermark on the wafer W. It is also possible to prevent organic contamination of the wafer table 20 and its peripheral part. In addition, the aspect using the liquid mixture (For example, the liquid mixture of IPA, water, or ethanol) of IPA and another liquid instead of IPA is also preferable.

  In the present embodiment, it is preferable that the refractive liquid A uses a liquid having a refractive index comparable to that of the condenser lens 66 and the table plate 22. According to this aspect, the difference in refractive index at the interface with the condenser lens 66 and the table plate 22 can be eliminated, and in addition to improving the transmittance of the laser beam, there is no change in the refractive angle of the laser beam. It becomes possible to do.

  Further, when the numerical aperture is increased, the condenser lens 66 is used from the center to the end. When the usage range of the condenser lens 66 is widened, the problem of spherical aberration increases. As a result, the spherical aberration leads to the variation of the modified layer at the focal point P in the depth direction. In particular, when it is desired to focus the laser beam on the shallow focal point P while preventing ablation on a thin wafer W, it is necessary to reduce the spherical aberration as much as possible. As a method of reducing this spherical aberration, it is desirable to reduce the refractive index difference from the condenser lens 66. In particular, spherical aberration is further increased by changing the refraction angle due to the refractive index difference at the interface in the medium. If there is no difference in refractive index, it behaves so as to travel in the same medium optically so that there is no interface in the medium.

  As such a refractive liquid A having a small refractive index difference, for example, Cargill standard refractive liquid A sold by Moritex Co., Ltd. as a commercially available one is preferably used. This is also called immersion oil and is used as a liquid for mounting the glasses together. For example, Cargill's standard refractive liquid Type A or Type B has a refractive index of 1.51. Since the refractive index of the glass used for the standard condensing lens 66 etc. is also about 1.5, there is almost no refractive index difference. Therefore, the change in the refraction angle immediately after coming out of the condensing lens 66 is small, and is hardly affected by the spherical aberration of the condensing lens 66.

  Further, when quartz glass is also used as the material of the wafer table 20 that supports the wafer W, the refractive index is about 1.46, which is almost the same as that of the refractive liquid A. Therefore, there is almost no change in the refraction angle, and the variation in the depth direction of the condensing point P due to the influence of spherical aberration can be reduced.

  If the substrate (wafer) itself is immersed in immersion oil suitably used as the refractive liquid A as in Patent Document 4, the substrate is contaminated with the immersion oil, and the substrate cannot be used. The target substrate may be not only silicon but also various cases such as sapphire and SiC, but the familiarity and contamination of the immersion oil are different from those substrates. Further, the direct contact of the refractive liquid A with the substrate has many problems from the viewpoint of stabilizing the refractive index.

  On the other hand, when the condensing lens 66 is glass, the wafer table 20 that supports the wafer W is also glass, and a refractive liquid A having a refractive index equivalent to the refractive index of the same glass material is used between them. The refraction angle hardly changes, and the influence of the spherical aberration of the condenser lens 66 can be reduced. Moreover, since the change in refractive index can be minimized by the transparent liquid that constantly lubricates the mutual movement of the same material between the glasses and moves by the interfacial tension, it is possible to ensure a stable optical path. As a result, laser light can be condensed with a high aperture, and ablation caused by the formation of a chain-modified layer on the substrate surface (wafer surface) due to absorption of laser energy can be prevented.

  For the same reason as the above-described refractive liquid A, a mode in which a liquid having a refractive index comparable to that of the dicing tape T is used as the refractive liquid B is preferable. According to this aspect, the difference in refractive index at the interface with the dicing tape T can be eliminated, and the transmittance of laser light can be improved. For example, when the dicing tape T is made of a polyolefin-based polyethylene film (refractive index: about 1.54), for example, dichlorotoluene (refractive index: 1.546) can be preferably used as the refractive liquid B26.

  As the liquid having a predetermined refractive index, for example, a refractive index standard solution manufactured by Kyoto Electronics Industry, a contact liquid (refractive liquid) manufactured by Shimadzu Corporation, a Cargill standard refractive solution manufactured by Moritex, etc. can be used.

<Wafer table>
In the present embodiment, as described above, the table plate 22 of the wafer table 20 is made of transparent quartz glass. However, the present invention is not limited to this. For example, the table plate 22 is made of a germanium plate, a silicon plate, a plastic resin plate, or the like. It may be. Moreover, according to the aspect in which the wafer W and the table plate 22 are made of the same material, the laser light can be transmitted without being affected by refraction by using the refracting liquid 26 having the same refractive index as that of the material. The rate can be further improved.

  Further, quartz glass or the like used as a material for the table plate 22 has an absorption band in the infrared light region, and the transmittance may be reduced. In such a case, the table plate 22 may not be made of quartz glass but may be made of a transparent resin material such as acrylic.

  As described above, the material of the table plate 22 is preferably a transmittance at a wavelength selected by the laser beam.

  In the present embodiment, as described above, the table plate 22 is formed with a necessary and sufficient thickness so that the wafer W to be processed can be held without bending. The table plate 22 is also required to have a uniform thickness.

  In this manner, the thickness of the table plate 22 of the wafer table 20 is made uniform and the bending is eliminated, so that the wafer plate 20 is pasted from the position of the condenser lens 66 of the laser irradiation device 60 via the wafer table 20 and the dicing tape T. The distance to the wafer surface is almost constant. Thereby, since the work distance can be made constant, it is possible to accurately form the modified layer at a certain focal point depth position even with a thin wafer.

  Since the thickness accuracy of the glass forming the table plate 22 greatly affects the variation in work distance, even if it is 200 mm quartz glass, the TTV (Total Thickness Variation) is usually 1 μm or less, actually 0.5 μm or less. It is polished with high precision.

  Further, the relative height accuracy with which the table plate 22 is operated with respect to the condenser lens 66 of the laser light source is also suppressed to about 0.5 μm. As a result, laser light can be condensed at a certain depth position with high accuracy without using autofocus, and laser dicing without ablation becomes possible.

  When the table plate 22 is bent alone, for example, as shown in FIG. 7, the laser beam incident surface (surface opposite to the holding surface 22A) 22B of the table plate 22 is formed on the outer peripheral portion. An annular rim member 30 may be provided along the same. In this way, by reinforcing the outer peripheral portion of the table plate 22 with a reinforcing member such as the rim member 30, it is possible to suppress the bending of the table plate 22.

  In the configuration example shown in FIG. 7, a plurality of suction holes (not shown) are formed on the outer peripheral portion of the lower surface of the table plate holding frame 24 (the surface on the dicing frame F side). By being vacuum-sucked through the suction holes, it is held in close contact with the table plate holding frame 24.

  In the present embodiment, the holding surface 22A of the wafer table 20 is preferably roughened. Further, instead of the holding surface 22A of the wafer table 20, or together with the holding surface 22A of the wafer table 20, the surface to be held of the dicing tape T attached to the wafer W (the surface opposite to the wafer W) is rough. Surface treatment may be performed. In this way, by subjecting at least one of the holding surface 22A of the wafer table 20 and the held surface of the dicing tape T to a rough surface treatment, a minute space is formed between them, which is caused by capillary action. The refractive liquid B spreads efficiently without a gap. As a result, the contact area between the surface subjected to the rough surface treatment and the refractive liquid B is increased, and the wafer W is held in close contact with a larger contact force.

  One method for roughening the holding surface 22A of the wafer table 20 (rough surface treatment) is a texturing (by facing) method. For example, as shown in FIG. 8, a small groove 32 is formed concentrically or spirally on the holding surface 22 </ b> A of the glass wafer table 20 so that the refractive liquid 26 spreads uniformly along the groove 32. become.

  The texturing grooves, depending on the refractive liquid and the material of the wafer table 20 (table plate 22), have a groove width of about 0.2 mm and a pitch of about 0.4 mm, and a groove width of 0.1 mm and a pitch of about 0.2 mm. It may be a groove formed on the entire surface corresponding to the wafer diameter on the holding surface 22A of the wafer table 20.

  In addition to the texturing, there is a method of simply roughening the holding surface 22A of the wafer table 20. For example, GC abrasive grain # 2000 may be used and the holding surface 22A may be lapped for about 20 minutes. Moreover, you may use the abrasive grain about # 500. The holding surface 22A may be roughened by using abrasive grains such as WA other than GC. By doing so, the holding surface 22A becomes on the ground glass and is scattered in the air due to surface roughness and becomes cloudy. The surface roughness may be 0.1 mm or less in terms of Ra, but is not limited to this, and it is better that the liquid is buried in the roughness gap so that the interfacial tension is increased and the surface area is increased.

  In this way, by roughening the holding surface 22A of the wafer table 20 by a method such as texturing, the refractive liquid B for compensating the refractive index is dropped on the holding surface 22A, and the back surface of the wafer W (the dicing tape T is applied). When the holding surface 22A is ringed, the refraction liquid 26 spreads uniformly over the holding surface 22A, and a uniform refractive index distribution can be obtained in the plane of the wafer W.

  As described above, according to the laser dicing apparatus 10 of the present embodiment, even when laser dicing is performed on a wafer, particularly a thin wafer having a thickness of 50 μm or less, ablation is prevented and irradiation efficiency is improved. Since the work distance can be made constant, the laser dicing process with high accuracy can be performed.

  In addition, since the work distance can be made constant by correcting the warpage and deflection of the wafer without destroying the fine-structured device formed on the wafer surface, even a wafer, particularly a thin wafer of 50 μm or less, can be used. The modified layer can be formed accurately inside.

[Wafer processing method]
The wafer W diced by the laser dicing method is then divided into chips by applying external stress. This process is performed by, for example, an expanding apparatus.

  FIG. 9 is a process diagram showing an outline of the expanding process by the expanding apparatus.

  As shown in FIG. 9A, the laser-diced wafer W is placed on the peeling table 110 with the surface facing up. The dicing frame F is fixed at a predetermined position by a frame fixing mechanism (not shown).

  When the wafer W is placed on the peeling table 110 and the dicing frame F is fixed, as shown in FIG. 9B, a ring 112 arranged so as to surround the periphery of the peeling table 110 is illustrated. It is pushed up by the lifting mechanism that does not. As a result, the dicing tape T attached to the back side of the wafer W is expanded (expanded) radially. When the dicing tape T is expanded, external stress is applied to the wafer W, and the wafer W is divided from the modified layer L as a starting point. Since the modified layer L is formed along the street, the wafer W is divided along the street. Since the street is set between individual chips, the wafer W is divided into individual chips.

  Thus, the laser-diced wafer W is divided into individual chips by applying external stress.

[Another aspect of the laser dicing apparatus]
In the laser dicing apparatus described with reference to FIG. 1 and the like, a highly accurate and uniform thickness table plate 22 is mounted on a wafer table 20 attached to the apparatus, and a wafer W is attached by using the table surface of the table plate 22 as a reference. This is a method of making the work distance constant.

  However, as a method of forming the modified layer inside the thin wafer without causing ablation, the thin wafer W is previously attached to a transparent plate (hereinafter referred to as “transparent support substrate 202”) with a high precision and uniform thickness. It may be prepared as a single object. In this case, instead of the wafer table 20 mounted on the apparatus, the wafer W is attached to the transparent support substrate 202 and handled as an integrated wafer unit 200. Therefore, the transparent support substrate 202 serves as the wafer table 20.

  However, the transparent support substrate 202 of the wafer unit 200 is a dummy portion, and the true portion for forming the modified layer is a thin wafer W attached to the transparent support substrate.

  The wafer unit 200 is mounted on a stage device (not shown) that can hold the peripheral edge of the wafer unit 200. Then, by operating and controlling the stage apparatus in the same manner as in the case of the wafer table 20 described above, the modified layer can be formed with high accuracy without causing ablation inside the wafer.

  In the method using this wafer unit 200, bumps 204, which are protrusions such as electrodes, are formed on the surface of the wafer, such as a wafer W for TSV (through silicon via). This is particularly effective when there is no flat surface.

  Next, a process of creating the wafer unit 200 will be described with reference to FIG.

  First, as shown in FIG. 10A, a transparent support substrate 202 for supporting a thin wafer W is prepared.

  Next, as shown in FIG. 10B, the wafer W to be laser-diced is placed on the transparent support substrate 202 with the side on which the bumps 204 are formed facing down.

  Next, as shown in FIG. 10C, the outer periphery of the wafer W and the transparent support substrate 202 is sealed with a sealing material 206, for example, a tape or a wax agent, so that the wafer W and the transparent support substrate 202 are sealed. Surrounding the gap formed on the contact surface.

  As the tape used as the sealing material 206, a polyolefin film or a polyethylene film can be preferably used. Then, this tape is attached so as to straddle the outer peripheral portion of the wafer W and the outer peripheral portion of the transparent support substrate 202 to seal the outer peripheral portion.

  This state is referred to as a unit intermediate 200A. The sealing material 206 is formed with fine holes (not shown) through which injection tubules 222 for injecting the refractive liquid B described later into the gap are inserted.

  Next, as illustrated in FIG. 10D, the unit intermediate 200 </ b> A is mounted on the mounting table 210 in the vacuum chamber 208. A vacuum pump 212 for reducing the pressure in the vacuum chamber is connected to the vacuum chamber 208 via a pipe 214, and an open / close valve 216 is provided on the pipe 214. In the vacuum chamber 208, an opening / closing door for inserting and removing the unit intermediate 200A is omitted.

  In addition, the vacuum chamber 208 is provided with an injection device 218 for injecting the refractive liquid B into a gap formed on the contact surface between the wafer W and the transparent support substrate 202. The injection device 218 includes a storage bottle 220 that stores the refractive liquid B, an injection tube 222 that injects the refractive liquid B from the storage bottle 220 into the gap, and a replenishment tube 224 that replenishes the storage bottle 220 with the refractive liquid B. Composed.

  First, the vacuum chamber 208 is depressurized by the vacuum pump 212. As a result, the gap formed on the contact surface between the wafer W and the transparent support substrate 202 is decompressed.

  Next, as shown in FIG. 10E, the injection thin tube 222 of the injection device 218 is inserted from the micro hole of the sealing material 206, and the refractive liquid B is injected into the gap in the reduced pressure state. This gap is formed by the bumps 204, and the gap is as small as the bump height. Therefore, the refracting liquid B is filled in the gap by capillary action. In this injection step, the vacuum pump 212 is preferably stopped.

  When the gap is filled with the refractive liquid B, the micropores of the tape are closed with wax or the like. Thereby, the wafer unit 200 is formed in which the wafer W and the transparent support substrate 202 are in close contact with each other by the interfacial tension of the liquid film 26B of the refractive liquid B.

  Next, as shown in FIG. 10F, the wafer unit 200 is taken out from the vacuum chamber 208. Even when atmospheric pressure is applied from the outside, the wafer unit 200 is filled with the refractive liquid B between the wafer W and the transparent support substrate 202 and does not bend extremely. .

  The wafer unit 200 thus formed is held by the above-described stage device of the laser dicing apparatus, and laser dicing processing is performed.

  Then, the wafer unit 200 that has completed the laser dicing process peels the wafer W from the transparent support substrate 202 by the method shown in FIG.

  First, as shown in FIG. 11A, the wafer unit 200 from which the sealing material 206 has been removed is placed in a pressure-resistant container 300 for peeling with the wafer side facing down. The pressure vessel 300 includes a container body 302 and a lid 304. A disk-shaped depression 306 that is slightly larger than the wafer W of the wafer unit 200 is formed at the bottom of the container main body 302. Further, a vacuum pump 308 for reducing the pressure inside the container is connected to the container body 302 via a pipe 310. Furthermore, the above-mentioned recess 306 is opened at the tip of a liquid supply pipe 312 that supplies liquid (for example, water) to the recess 306.

  Then, as shown in FIG. 11B, the wafer surface of the wafer unit 200 is placed so as to fit into the recess 306.

  In this state, the vacuum pump 308 is driven to depressurize the pressure vessel 300. By this pressure reduction, the refractive liquid B that forms the liquid film 26B interposed between the wafer W and the transparent support substrate 202 is evaporated and removed. Thereby, the interfacial tension between the wafer W and the transparent support substrate 202 is eliminated.

  Next, as shown in FIG. 11C, a small amount of liquid is supplied from the liquid supply pipe 312 to the recess 306 to form a liquid film 314 between the wafer surface and the recess surface. As a result, the wafer surface and the recessed surface are brought into close contact with each other by the interfacial tension of the liquid film 314.

  In this state, the lid 304 of the pressure vessel 300 is removed, and the upper surface of the transparent support substrate 202 is sucked with the suction pad 316 provided separately, and the suction pad 316 is raised. Thereby, the transparent support substrate 202 is peeled from the wafer W.

  The present invention is not limited to the above-described embodiment, and is included in a range not departing from the concept of the present invention.

  DESCRIPTION OF SYMBOLS 10 ... Laser dicing apparatus, 20 ... Wafer table, 22 ... Table plate, 22A ... Holding surface, 22B ... Table surface (laser beam incident surface), 24 ... Table plate holding frame, 26A ... Liquid film of refraction liquid A, 26B ... Liquid film of refracting liquid B, 50 ... laser antireflection plate, 60 ... laser irradiation device, 62 ... laser oscillation device, 64 ... collimator lens, 66 ... condensing lens, 66A ... lens surface (laser light emitting surface), 68 ... Actuator, 69 ... enclosure member, 110 ... peeling table, 112 ... ring, 200 ... wafer unit, 202 ... transparent support substrate, 204 ... bump, 206 ... sealing material, 208 ... decompression chamber, 210 ... mounting table, 212 ... vacuum Pump, 214 ... piping, 216 ... open / close valve, 218 ... injection device, 220 ... storage bottle, 222 ... injection capillary, 224 ... replenishment piping, 00 ... Pressure-resistant container for peeling, 302 ... Container body, 304 ... Lid, 306 ... Depression, 308 ... Vacuum pump, 310 ... Pipe, 312 ... Liquid supply pipe, 314 ... Liquid film, W ... Wafer, T ... Dicing tape, F: Dicing frame, A: Refraction liquid, B: Refraction liquid, a: Laser light incident surface of wafer, P: Condensing point

Claims (18)

In a laser dicing apparatus that irradiates a wafer having a plurality of devices formed on the surface with laser light and forms a modified layer inside the wafer,
A wafer table that is formed so as to transmit the laser light and has a flat holding surface for tightly holding the back side of the wafer;
Laser irradiation means for irradiating the laser beam from the back side of the wafer through the wafer table,
In the gap between the condensing lens surface of the laser irradiation means and the laser light incident surface of the wafer table, a liquid film of the refractive liquid A having a refractive index larger than that of air is held by interfacial tension. Laser dicing equipment.   The laser dicing apparatus according to claim 1, wherein the wafer has a thickness of 50 μm or less.   3. The laser dicing apparatus according to claim 1, wherein the gap between the condenser lens surface and the laser light incident surface of the wafer table has a shortest distance in a range of 0.01 to 2 mm.   The said wafer table has a table plate which can permeate | transmit a laser beam, and the front and back are flat, This table plate is formed with either quartz glass, germanium, a silicon | silicone, or a plastic resin. The laser dicing apparatus according to any one of the above.   5. The laser dicing apparatus according to claim 1, wherein the rear surface side of the wafer is held in close contact with the holding surface of the wafer table with a refractive liquid B having a refractive index larger than that of air interposed therebetween.   The laser dicing apparatus according to claim 5, wherein the refractive liquid A and the refractive liquid B are the same type of refractive liquid.   The laser dicing apparatus according to claim 5 or 6, wherein the refractive liquid A and the refractive liquid B are liquids containing isopropyl alcohol.   The laser dicing apparatus according to claim 1, wherein the laser light incident surface and the holding surface of the wafer table are subjected to a rough surface treatment.   The wafer table is horizontally installed with the holding surface facing downward, and the laser irradiation unit irradiates the wafer held on the wafer table with laser light from above the wafer table. Item 9. The laser dicing apparatus according to any one of Items 1 to 8. In a laser dicing method in which a wafer having a plurality of devices formed on the surface is irradiated with laser light from a laser irradiation means to form a modified layer inside the wafer,
A wafer holding step of holding the wafer by a wafer table that is formed so as to transmit the laser light and has a flat holding surface that holds the back side of the wafer closely;
A moving step of moving the laser irradiation means and / or the wafer table closer so that a gap of a predetermined distance is formed between the condensing lens surface of the laser irradiation means and the laser light incident surface of the wafer table;
A liquid film forming step of forming a liquid film by supplying a refractive liquid A having a refractive index larger than air to the gap;
An irradiation step of irradiating the laser beam from the back surface side of the wafer through the wafer table.   The laser dicing method according to claim 10, wherein the wafer has a thickness of 50 μm or less.   The laser dicing method according to claim 10 or 11, wherein the gap between the condensing lens surface and the laser light incident surface of the wafer table has a shortest distance of 0.01 to 2 mm.   The laser dicing method according to any one of claims 10 to 12, wherein the back surface side of the wafer is held in close contact with the holding surface of the wafer table with a refractive liquid B having a refractive index larger than that of air interposed therebetween.   The laser dicing method according to claim 13, wherein the refractive liquid A and the refractive liquid B are the same type of refractive liquid.   The laser dicing method according to claim 13 or 14, wherein the refractive liquid A and the refractive liquid B are liquids containing isopropyl alcohol.   16. The laser dicing method according to claim 10, wherein the laser light incident surface and the holding surface of the wafer table are roughened.   The wafer table is horizontally installed with the holding surface facing downward, and the laser irradiation unit irradiates the wafer held on the wafer table with laser light from above the wafer table. Item 17. The laser dicing method according to any one of Items 10 to 16. Laser dicing, in which a plurality of devices are formed on the front surface, and a laser beam is irradiated to a wafer mounted on a frame via a tape attached to the back surface to form a modified layer inside the wafer. Process,
Expanding the tape of the wafer that has been subjected to laser dicing processing to expand the wafer into individual chips on the tape, and a wafer processing method comprising:
A wafer processing method using the laser dicing method according to claim 10 in the laser dicing step.

Images