JP2006187827A - Polishing pad - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a polishing pad which minimizes edge sagging of a wafer, reduces generation of unusual sound, vibration and damping, by suppressing the slidability of a polishing head, and is excellent in pad service life and polishing rate. <P>SOLUTION: The polishing pad 5 has a lamination structure consisting of a polishing layer 1 formed of a foamed polyurethane resin, an adhesive layer 3, a cushioning layer 2, and a tackifier layer 4 to be stuck to a polishing device surface plate, which are laminated on each other in this order. Herein, when the tackifier layer of the polishing pad is stuck to a non-repulsive board, a rebound resilience value of a front surface of the polishing layer measured according to JIS K6400, is set to 35 to 45%. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a polishing pad used when planarizing irregularities on a wafer surface by chemical mechanical polishing (CMP), a method for polishing a semiconductor wafer using the polishing pad, and a surface of a semiconductor wafer using the polishing pad. The present invention relates to a semiconductor device manufacturing method including a polishing step.

  When manufacturing a semiconductor device, a step of forming a conductive layer on the wafer surface and forming a wiring layer by photolithography, etching, or the like, or a step of forming an interlayer insulating film on the wiring layer These steps cause irregularities made of a conductor such as metal or an insulator on the wafer surface. In recent years, miniaturization of wiring and multilayer wiring have been advanced for the purpose of increasing the density of semiconductor integrated circuits, and along with this, technology for flattening the irregularities on the wafer surface has become important.

  As a method for flattening the irregularities on the wafer surface, a CMP method is generally employed. In CMP, polishing flattening is performed by adding a polishing slurry in which fine particles (abrasive grains) are dispersed in a gap between a polishing pad and a semiconductor wafer. In the CMP processing step, for example, a wafer is mounted on a rotatable polishing head and pressed by a mechanical force against a polishing pad affixed on a rotating platen (platen), and the chemical action and mechanical properties of the polishing slurry. It is the mainstream that polishing is performed by action.

  As described above, when polishing is performed by a relative movement in pressure contact like a polishing head (wafer) and a platen (pad), the slidability is important. Inferior slidability causes damping of the polishing head, vibration of the polishing apparatus, abnormal noise, and the like.

  In addition, it is known that a phenomenon called edge sagging occurs in which the polishing rate increases at the outer peripheral portion of the wafer due to elastic restoration of the pressurizing portion of the pad. Wafer mounting guides and retaining rings (retainer rings) are recognized to have some effect on this edge sag, but are not perfect.

  Regarding the relationship between rebound resilience and flattening characteristics, a polishing pad system has been proposed in which the resilience etc. are changed by changing the configuration of the polishing pad in accordance with the change in the object to be polished during the polishing process (Patent Document). 1). Further, a polishing pad provided with a rigid element and an elastic element adjacent to the polishing surface is disclosed (Patent Document 2).

However, the technique described in Patent Document 1 cannot be a general technique because the system control and the configuration of the polishing pad are complicated and technically difficult. Moreover, in the technique described in Patent Document 2, the role of the rigid element is to eliminate the stress bias caused by the integration density on the wafer surface. In other words, it is intended to improve the planarization by giving the polishing surface rigidity while adapting to the topography of the wafer surface, but by controlling the resilience, which is the physical characteristic of the polishing pad itself, the slidability and wafer It is not a technique to improve the edge of the edge.
Japanese translation of PCT publication No. 2001-513352 JP-A-6-77185

  It is an object of the present invention to provide a polishing pad that can reduce edge noise of a wafer, suppress sliding properties of a polishing head, and reduce noise, vibration, and damping during polishing. Furthermore, it is providing the polishing pad which is excellent also in a pad lifetime and a polishing rate. Another object of the present invention is to provide a method for polishing a semiconductor wafer using the polishing pad, and a method for manufacturing a semiconductor device including a step of polishing the surface of the semiconductor wafer using the polishing pad.

  As a result of intensive research to achieve the above object, the present inventors have found the relationship between the rebound resilience value of the polishing layer surface, sliding characteristics, and edge sag of the wafer, and the rebound resilience value of the polishing layer surface is controlled within a specific range. As a result, the inventors have found that the above problems can be solved, and have completed the present invention.

  That is, the present invention is a polishing pad in which a polishing layer made of polyurethane foam resin, an adhesive layer, a cushion layer, and a pressure-sensitive adhesive layer for bonding to a polishing apparatus surface plate are laminated in this order, The present invention relates to a polishing pad, wherein the pad adhesive layer is affixed to a non-repulsive plate, and the rebound resilience value of the polishing layer surface is 35 to 45% when measured according to JIS K6400.

  As a result of investigating the relationship between the thickness of the polishing layer and the cushion layer and the rebound resilience value, as shown in FIGS. 1 and 2, the rebound resilience value of the polishing layer and the cushion layer is proportional to the inverse of the logarithm of the thickness as the thickness increases. It has been found that when the thickness exceeds a certain thickness, it converges to a value inherent to the material (defined as the intrinsic rebound resilience value in the present invention) and becomes almost independent of the thickness. In the present invention, the lower limit thickness indicating the intrinsic rebound resilience value is defined as the critical thickness. That is, it was found that the rebound resilience value varies depending on the thickness when the thickness is less than the critical thickness, but the constant rebound resilience value described above is exhibited regardless of the thickness when the thickness exceeds the critical thickness.

  Further, the inventors of the present invention are that the polishing layer having higher rigidity (elastic modulus) is more susceptible to the impact resilience of the cushion layer (lower layer). It was also found that the influence of the cushion layer varies depending on the difference in thickness and rigidity, and that the greater the rigidity, the more susceptible to the impact resilience of the cushion layer.

  By controlling the rebound resilience value of the polishing layer surface of the polishing pad to 35 to 45%, the edge sag of the wafer is suppressed, and the slidability of the polishing head is suppressed to suppress abnormal noise, vibration, and Damping can be reduced. The rebound resilience value of the polishing layer surface is preferably 38 to 43%. When the rebound resilience value is less than 35%, the sliding resistance of the polishing head increases and abnormal noise and vibration are generated. Further, the edge sag of the wafer is increased and the flattening characteristics are also deteriorated. On the other hand, when the rebound resilience value exceeds 45%, the polishing head rebounds (damping, the polishing head moves up and down), and the polishing rate is reduced.

  In the present invention, it is preferable that a rebound resilience adjusting layer is further laminated between the polishing layer and the cushion layer via an adhesive layer in the polishing pad.

  As described above, the rebound resilience value can be relatively easily obtained by providing a rebound resilience adjustment layer between the polishing layer and the cushion layer by utilizing the fact that the polishing layer having high rigidity is cheaply affected by the impact resilience of the immediate lower layer. Can be controlled within the range. In other words, the amount of displacement of the polishing pad due to the polishing pressure (the displacement amount of the cushion layer occupies most) is preferably within a certain range, but if a rigid polishing layer is used or the thickness of the polishing layer is increased, the resilience value Therefore, it is necessary to reduce the thickness of the cushion layer. Then, the amount of displacement of the polishing pad becomes small and deviates from the optimum range. Conversely, if a polishing layer with low rigidity is used or the thickness of the polishing layer is reduced, it is necessary to increase the thickness of the cushion layer in order to obtain a predetermined rebound resilience value. The amount increases and deviates from the optimum range. The amount of displacement refers to the amount of displacement by which the polishing pad surface moves up and down when the polishing pad surface traces the waviness of the wafer by the polishing pressure.

  By providing a rebound resilience adjustment layer between the polishing layer and the cushion layer, the thickness of the cushion layer can be maintained so that the amount of displacement with respect to the polishing pressure is in the optimum range, and the rebound resilience value of the polishing pad is optimized. Since the range can be adjusted, the polishing pad is excellent in polishing characteristics (flattening characteristics, polishing rate, and in-plane uniformity).

  In the present invention, it is preferable that the intrinsic rebound resilience value of the rebound resilience adjustment layer is smaller than the intrinsic rebound resilience value of the cushion layer.

  In the present invention, it is preferable that the intrinsic rebound resilience value of the rebound resilience adjustment layer is larger than the intrinsic rebound resilience value of the cushion layer.

  Furthermore, in the present invention, it is preferable that each thickness of the polishing layer, the rebound resilience adjusting layer, and the cushion layer is smaller than the critical thickness of each layer. This is because the relationship between the thickness and the rebound resilience value disappears above the critical thickness, and the rebound resilience value cannot be adjusted by the thickness.

  The present invention relates to a method for polishing a semiconductor wafer using the polishing pad.

  The present invention also relates to a method for manufacturing a semiconductor device including a step of polishing a surface of a semiconductor wafer using the polishing pad.

  The amount of displacement in the thickness direction due to the polishing pressure of the polishing pad during polishing is maintained within a range in which polishing characteristics such as flattening characteristics, polishing rate, and in-plane uniformity can be satisfactorily maintained, and the thickness of the polishing layer and cushion layer By controlling the rebound resilience value and thereby optimizing the rebound resilience value of the polishing layer, the slidability during polishing can be improved and the edge sag phenomenon can be prevented.

  Also, by providing a rebound resilience adjusting layer between the cushion layer and the polishing layer, while maintaining good slidability and edge sag prevention, a polishing layer with high hardness and elastic modulus, conversely hardness and elastic modulus etc. Therefore, it is possible to use a polishing layer having a small thickness and a polishing layer having a large thickness, and a polishing pad having both a high polishing rate, prevention of edge sagging, long pad life, and improvement in slidability can be obtained.

  The configuration of the polishing pad of the present invention is shown in FIGS. FIG. 3 shows the configuration of the polishing pad 5 in which the polishing layer 1, the adhesive layer 3, the cushion layer 2, and the pressure-sensitive adhesive layer 4 to be attached to the polishing apparatus surface plate are laminated in this order. FIG. 4 shows a polishing in which a polishing layer 1, an adhesive layer 3, a rebound resilience adjusting layer 6, an adhesive layer 3, a cushion layer 2, and an adhesive layer 4 to be attached to a polishing apparatus surface plate are laminated in this order. The structure of the pad 5 is represented.

  The polishing layer 1 is made of a foamed polyurethane resin. Polyurethane resins are preferred because they are excellent in abrasion resistance and can be obtained from polymers having desired physical properties by variously changing the raw material composition.

  The polyurethane foam resin has as a main component a first component containing an isocyanate group-containing compound and a second component containing an active hydrogen group-containing compound.

  As the isocyanate group-containing compound, known isocyanate compounds in the field of polyurethane can be used without particular limitation. In particular, the use of diisocyanate compounds and derivatives thereof, particularly isocyanate prepolymers, is preferred because the resulting polyurethane foam has excellent physical properties. Incidentally, as a method for producing polyurethane, a prepolymer method and a one-shot method are known, but any method can be used in the present invention.

  Examples of the isocyanate compound include 2,4-toluene diisocyanate, 2,6-toluene diisocyanate, 2,2′-diphenylmethane diisocyanate, 2,4′-diphenylmethane diisocyanate, 4,4′-diphenylmethane diisocyanate, and 1,5-naphthalene diisocyanate. , P-phenylene diisocyanate, m-phenylene diisocyanate, p-xylylene diisocyanate, m-xylylene diisocyanate and other aromatic diisocyanates, ethylene diisocyanate, 2,2,4-trimethylhexamethylene diisocyanate, 1,6-hexamethylene diisocyanate Aliphatic diisocyanates such as 1,4-cyclohexane diisocyanate, 4,4′-dicyclohexylmethane diisocyanate, Isophorone diisocyanate, cycloaliphatic diisocyanates such as norbornane diisocyanate and the like. These may be used alone or in combination of two or more.

  As the isocyanate compound, in addition to the diisocyanate compound, a polyfunctional polyisocyanate compound having three or more functions can be used. As a polyfunctional isocyanate compound, a series of diisocyanate adduct compounds are commercially available as Desmodur-N (manufactured by Bayer) or trade name Duranate (manufactured by Asahi Kasei Kogyo Co., Ltd.). These trifunctional or higher functional polyisocyanate compounds are preferably used by adding to the diisocyanate compound because they are easily gelled during prepolymer synthesis when used alone.

In the present invention, the isocyanate group-containing compound suitable for use as the first component is an isocyanate prepolymer which is a reaction product of the above-described isocyanate compound and an active hydrogen group-containing compound. As such an active hydrogen group-containing compound, a polyol compound or a chain extender described later is used, and an equivalent ratio NCO / H ^{* of} the isocyanate group (NCO) to the active hydrogen (H ^* ) is 1.2 to 5.0. The isocyanate prepolymer which is an oligomer having an isocyanate group terminal is produced by a heat reaction in the range of preferably 1.6 to 2.6. The use of commercially available isocyanate prepolymers is also suitable.

  The active hydrogen group-containing compound used in the present invention is an organic compound having at least two or more active hydrogen atoms, and is a compound usually referred to as a polyol compound or a chain extender in the technical field of polyurethane.

  The active hydrogen group is a functional group containing hydrogen that reacts with an isocyanate group, and examples thereof include a hydroxyl group, a primary or secondary amino group, and a thiol group (SH).

  Examples of the polyol compound include those usually used in the technical field of polyurethane. For example, a polyether polyol typified by polytetramethylene ether glycol and polyethylene glycol, a polyester polyol typified by polybutylene adipate, a polycaprolactone polyol, a reaction product of a polyester glycol such as polycaprolactone and an alkylene carbonate, etc. The polyester polycarbonate polyol exemplified in the above, obtained by reacting ethylene carbonate with a polyhydric alcohol, and then reacting the obtained reaction mixture with an organic dicarboxylic acid, transesterification of a polyhydroxyl compound and an aryl carbonate. Examples thereof include polycarbonate polyol. These may be used alone or in combination of two or more.

  In addition, the number average molecular weight of these polyol compounds is not particularly limited, but is preferably about 500 to 2000 from the viewpoint of the elastic characteristics of the obtained polyurethane foam. When the number average molecular weight of the polyol compound is less than 500, the foamed polyurethane obtained using the polyol compound does not have sufficient elastic properties and tends to be a brittle polymer, and the abrasive sheet made of this foamed polyurethane becomes too hard and is subject to polishing. It may cause scratches on the polished surface of the workpiece. Moreover, since it becomes easy to wear, it is not preferable from the viewpoint of the life of the polishing sheet. On the other hand, when the number average molecular weight exceeds 2000, an abrasive sheet made of a polyurethane foam obtained by using this becomes soft and it is difficult to obtain sufficiently satisfactory planarity, which is not preferable.

  Examples of the polyol compound include ethylene glycol, 1,2-propylene glycol, 1,3-propylene glycol, 1,4-butanediol, 1,6-hexanediol, neopentyl glycol, in addition to the above-described high molecular weight polyol. 1,4-cyclohexanedimethanol, 3-methyl-1,5-pentanediol, diethylene glycol, triethylene glycol, 1,4-bis (2-hydroxyethoxy) benzene and other low molecular weight polyols may be used in combination. . These may be used alone or in combination of two or more.

  Among the active hydrogen group-containing compounds, what is called a chain extender is a compound having a molecular weight of about 500 or less. Specifically, ethylene glycol, 1,2-propylene glycol, 1,3-propylene glycol, 1,4-butanediol, 1,6-hexanediol, neopentyl glycol, 1,4-cyclohexanedimethanol, 3- Aliphatic low molecular weight glycols and triols typified by methyl-1,5-pentanediol, diethylene glycol, triethylene glycol, trimethylolpropane, etc., 1,4-bis (2-hydroxyethoxy) benzene, m-xylylene diene Aromatic diols represented by ol, etc., 4,4′-methylenebis (o-chloroaniline), 2,6-dichloro-p-phenylenediamine, 4,4′-methylenebis (2,3-dichloroaniline) 3,5-bis (methylthio) -2,4-toluenediamine, 3,5 Bis (methylthio) -2,6-toluenediamine, 3,5-diethyltoluene-2,4-diamine, 3,5-diethyltoluene-2,6-diamine, trimethylene glycol-di-p-aminobenzoate, 1 , 2-bis (2-aminophenylthio) ethane, 4,4′-diamino-3,3′-diethyl-5,5′-dimethyldiphenylmethane, and other polyamines. These may be used alone or in combination of two or more.

  The ratio of the isocyanate compound, the polyol compound, and the chain extender in the present invention can be variously changed depending on the molecular weight of each, the desired physical properties of the polishing layer produced therefrom. In order to obtain a polishing layer having desired polishing characteristics, the number of isocyanate groups of the isocyanate compound relative to the total number of functional groups of the polyol compound and the chain extender is preferably in the range of 0.95 to 1.15, more preferably 0.99. ~ 1.10.

  The foamed polyurethane resin can be produced by applying known urethanization techniques such as a melting method and a solution method, but for the foamed polyurethane resin of the present invention, it is necessary to take pores (bubbles) into the polyurethane, Further, it is preferable to manufacture by a melting method in consideration of cost, working environment and the like.

  The foamed polyurethane resin of the present invention is obtained by mixing and curing a first component containing an isocyanate group-containing compound and a second component containing an active hydrogen group-containing compound. In the prepolymer method, the isocyanate-terminated prepolymer becomes an isocyanate group-containing compound, and the chain extender becomes an active hydrogen group-containing compound. In the one-shot method, the organic polyisocyanate becomes an isocyanate group-containing compound, and the chain extender and the polyol compound become active hydrogen group-containing compounds. If necessary, stabilizers such as antioxidants, lubricants, pigments, fillers, antistatic agents, and other additives may be added.

  Examples of the method for producing a foamed polyurethane resin include, but are not limited to, a method of adding hollow beads, a method of forming a foam by a mechanical foaming method, a chemical foaming method, and the like. Each method may be used in combination, but a mechanical foaming method using a silicone surfactant which is a copolymer of polyalkylsiloxane and polyether and does not have an active hydrogen group is more preferable. As such a silicone-based surfactant, SH-192 (manufactured by Toray Dow Corning Silicon) is exemplified as a suitable compound.

An example of a method for producing a foamed polyurethane resin constituting the polishing layer will be described below. The manufacturing method of this foaming polyurethane resin has the following processes.
1) Stirring step for producing a cell dispersion of isocyanate-terminated prepolymer Add a silicone surfactant to the isocyanate-terminated prepolymer, stir with a non-reactive gas, and disperse the non-reactive gas as fine bubbles to disperse the bubble. Use liquid. When the prepolymer is solid at room temperature, it is preheated to an appropriate temperature and melted before use. 2) Curing Agent (Chain Extender) Mixing Step A chain extender is added to the above cell dispersion and mixed and stirred.
3) Curing step An isocyanate-terminated prepolymer mixed with a chain extender is cast and cured by heating.

  As the non-reactive gas used to form the fine bubbles, non-flammable gases are preferable, and specific examples include nitrogen, oxygen, carbon dioxide, rare gases such as helium and argon, and mixed gases thereof. The use of air that has been dried to remove moisture is most preferable in terms of cost.

  In the method for producing the foamed polyurethane resin, heating and post-curing the foam that has been reacted until the cell dispersion is poured into the mold and no longer flows has the effect of improving the physical properties of the foam. Is preferred. The bubble dispersion may be poured into the mold and immediately placed in a heating oven for post-cure. Under such conditions, heat is not immediately transferred to the reaction components, so the bubble diameter does not increase. . The curing reaction is preferably performed at normal pressure because the bubble shape is stable.

  In the foamed polyurethane resin, a catalyst that promotes a known polyurethane reaction such as tertiary amine or organotin may be used. The type and addition amount of the catalyst are selected in consideration of the flow time for pouring into a mold having a predetermined shape after the mixing step.

  In the present invention, the method for producing the polyurethane resin foam is not particularly limited, and a general method can be used. For example, a prepolymer as a raw material for the polishing layer is placed in a reaction vessel, a curing agent is added, stirred, and then poured into a casting mold of a predetermined size to produce a block, and the block is formed into a bowl-shaped or band saw-shaped slicer. A thin sheet may be used in the method of slicing using the above-mentioned method or the above-described casting step. Alternatively, the raw material resin may be dissolved and extruded from a T-die to directly obtain a sheet-like foam.

  In this invention, it is preferable that the average cell diameter of the said foam is 70 micrometers or less. When deviating from this range, the planarity (flatness) of the polished material after polishing tends to decrease.

  In the present invention, the foamed polyurethane resin preferably has a specific gravity of 0.5 to 1.0. When the specific gravity is less than 0.5, the surface strength of the polishing layer decreases, and the planarity of the material to be polished tends to decrease. On the other hand, when the specific gravity is larger than 1.0, the number of bubbles on the surface of the polishing layer is reduced and the planarity is good, but the polishing rate tends to decrease.

  In the present invention, the foamed polyurethane resin preferably has a hardness of 45 to 65 degrees as measured by an Asker D hardness meter. When the Asker D hardness is less than 45 degrees, the planarity of the material to be polished is lowered. When the Asker D hardness is more than 65 degrees, the planarity is good but the uniformity of the material to be polished is lowered. There is a tendency.

  The thickness variation of the polishing layer is preferably 100 μm or less. When the thickness variation exceeds 100 μm, the polishing layer has a large waviness, and there are portions where the contact state with the material to be polished is different, which adversely affects the polishing characteristics. In order to eliminate the thickness variation of the polishing layer, in general, the surface of the polishing layer is dressed with a dresser in which diamond abrasive grains are electrodeposited and fused in the initial stage of polishing. As a result, the dressing time becomes longer and the production efficiency is lowered.

  Examples of a method for suppressing the variation in the thickness of the polishing layer include a method of buffing the surface of the polishing layer sliced to a predetermined thickness. Moreover, when buffing, it is preferable to carry out stepwise with abrasives having different particle sizes.

  The polishing surface of the polishing layer of the present invention that comes into contact with the material to be polished preferably has a surface shape that holds and renews the slurry. The polishing layer made of foam has many openings on the polishing surface and has the function of holding and renewing the slurry. However, in order to further maintain the slurry and renew the slurry efficiently, the polishing layer is also polished. In order to prevent destruction of the material to be polished due to adsorption with the material, it is preferable that the polished surface has an uneven structure. The concavo-convex structure is not particularly limited as long as the shape holds and renews the slurry. For example, an XY lattice groove, a concentric circular groove, a through hole, a non-penetrating hole, a polygonal column, a cylinder, a spiral groove, Examples include eccentric circular grooves, radial grooves, and combinations of these grooves. In addition, these uneven structures are generally regular, but in order to make the slurry retention and renewability desirable, the groove pitch, groove width, groove depth, etc. should be changed for each range. Is also possible.

  The method for producing the concavo-convex structure is not particularly limited. For example, a method of machine cutting using a jig such as a tool of a predetermined size, pouring a resin into a mold having a predetermined surface shape, and curing. Using a press plate having a predetermined surface shape, a method of producing a resin by pressing, a method of producing using photolithography, a method of producing using a printing technique, a carbon dioxide laser, etc. Examples include a manufacturing method using laser light.

  When the thickness of the polishing layer is of primary importance, it can be realized by selecting or adjusting the inherent rebound resilience of the polishing layer so that the rebound resilience of the polishing pad at that thickness can achieve the optimum rebound resilience . As described above, for example, in the case of mechanically foamed urethane, the intrinsic rebound resilience of the material used for the polishing layer can be adjusted by the amount of mixed bubbles, the dispersed state, and the average bubble diameter. Further, in the case of including one or more fillers for other foamed polyurethane resins, it can be adjusted by adjusting the type, size, and dispersion density of each filler related to rigidity. Further, it is common to pattern the surface of the polishing layer. In this case, the resilience can be slightly adjusted by the surface pattern.

  The cushion layer 2 supplements the characteristics of the polishing layer. The cushion layer is necessary in order to achieve both planarity and uniformity in a trade-off relationship in CMP. Planarity refers to the flatness of a pattern portion when a material having fine irregularities generated during pattern formation is polished, and uniformity refers to the uniformity of the entire material to be polished. The planarity is improved by the characteristics of the polishing layer, and the uniformity is improved by the characteristics of the cushion layer. In the polishing pad of the present invention, the cushion layer is softer than the polishing layer.

  The material used for the cushion layer is not particularly limited as long as it is softer than the polishing layer. For example, fiber nonwoven fabrics such as polyester nonwoven fabric, nylon nonwoven fabric and acrylic nonwoven fabric, resin impregnated nonwoven fabrics such as polyester nonwoven fabric impregnated with polyurethane, polymer resin foams such as polyurethane foam and polyethylene foam, rubber properties such as butadiene rubber and isoprene rubber Examples thereof include resins and photosensitive resins.

  When the polishing pad is composed of a polishing layer, an adhesive layer, a cushion layer, and an adhesive layer to be attached to the polishing apparatus surface plate, the thickness of the selected cushion layer and polishing layer in the polishing pad design. The amount of displacement of the pad due to the polishing pressure (mostly the amount of displacement of the cushion layer) is within the optimum range from the thickness of the cushion layer and the polishing layer, and the relationship between the resilience and the resilience is obtained and charted. It is desirable to select a combination of thicknesses that achieves the pad's impact resilience within the optimum range.

  In the present invention, by providing the impact resilience adjusting layer 6 between the polishing layer 1 and the cushion layer 2, it is possible to use a polishing layer having a considerably high rigidity as described above, and for example, by increasing the thickness of the polishing layer. A long-life pad can be provided. In this case, if a rigid polishing layer or a thick polishing layer is used, the rebound resilience increases. However, taking advantage of the fact that the polishing layer is easily affected by the immediate lower layer, an ultra-low rebound resilience sheet material is optimally thick. If the pad is ground and provided between the cushion layer and the polishing layer, the thickness of the cushion layer can be maintained so that the displacement of the pad is within the optimum range as described above. If a polishing layer having high rigidity is used, a pad having both flattening characteristics and sliding characteristics can be provided. If a thick polishing layer is used, a pad having both a long polishing life and suppression of sliding characteristics can be provided.

  Further, if the hardness and elastic modulus of the polishing layer are reduced, the polishing rate can be increased, but the rebound resilience is reduced and the edge sag and flattening characteristics are deteriorated. However, if the cushion layer is thickened to increase the resilience, the planarization characteristics and the in-plane uniformity are deteriorated. However, if a high rebound resilience material is provided between the cushion layer and the polishing layer as a rebound resilience adjusting layer as in the present invention, the thickness of the cushion layer is maintained within the optimum range with respect to the polishing characteristics, and polishing with less edge sag A polishing pad having a large rate can be obtained.

  When a low intrinsic rebound resilience material is used as the rebound resilience adjustment layer, it is necessary that the inherent rebound resilience value of the material used is smaller than the intrinsic rebound resilience value of the cushion layer. And it is preferable that it is sufficiently smaller than the cushion layer. This is because the smaller the intrinsic rebound resilience value of the rebound resilience adjustment layer, the greater the effect as the adjustment layer, and thus the thinner the material.

  This is important considering that the amount of displacement of the low-rebound elastic material is generally large with respect to the pressure. In this case, the displacement of the pad due to polishing is mostly the displacement of the impact resilience adjusting layer and the cushion layer. However, if the resilience resilience adjusting layer is thick, the displacement becomes large. It is necessary to reduce the thickness. This is because the original role of the cushion layer, that is, the effect of imparting appropriate elasticity to the polishing surface of the polishing pad is reduced. Therefore, it is important to adjust the thicknesses of the cushion layer and the rebound resilience adjusting layer so that the displacement amount of the pad due to polishing is optimally maintained and the rebound resilience value of the polishing pad is optimized. As is clear from this, the thickness of the rebound resilience adjusting layer needs to be smaller than the thickness of the cushion layer. The thickness is preferably such that the thickness of the cushion layer can be maintained within a range where the original effect can be exerted, and the rebound resilience of the polishing layer can be adjusted so that the rebound resilience of the entire pad is within the optimum range. Thin is preferred.

  Examples of the rebound resilience adjusting layer of the low rebound resilience material include, but are not limited to, low rebound resilience independent foamed urethane, sponge-like foamed urethane, sponge-like rubber sheet, and the like.

  When using a high rebound resilience material as the rebound resilience adjustment layer, the inherent rebound resilience value needs to be greater than the intrinsic rebound resilience value of the cushion layer. This is because the rebound resilience decreases when the rebound resilience value of the resilience adjustment layer is smaller than the rebound resilience value of the cushion layer. And it is preferable that it is sufficiently larger than the resilience of the polishing layer.

  The higher the intrinsic resilience value, the thinner the material, but this is important because it does not reduce the role of the cushion layer. The thickness of the rebound resilience adjusting layer is preferably thin enough not to reduce the role of the cushion layer, and is a thickness capable of adjusting the rebound resilience of the polishing layer to the optimum range.

  Examples of the high rebound resilience adjusting layer include, but are not limited to, steel plates, glass fiber reinforced plastics, carbon fiber reinforced plastics, non-foamed hard urethane, epoxy resins, polycarbonate resins, and high hardness and high resilience rubbers. is not.

  The rebound resilience adjusting layer needs to be provided between the polishing layer and the cushion layer. This is because when the impact resilience adjusting layer is provided between the cushion layer and the polishing surface plate, the polishing layer is greatly affected by the cushion layer immediately below, and thus the effect cannot be expected.

  Moreover, it is preferable that the thickness of the polishing layer, the cushion layer, and the rebound resilience adjusting layer is not more than the critical thickness. This is because when the thickness exceeds the critical thickness, the relationship between the thickness and the rebound resilience value disappears, and the rebound resilience value cannot be adjusted by the thickness.

  Examples of the adhesive layer that is a means for attaching the polishing layer, the cushion layer, and the impact resilience adjusting layer include a double-sided tape.

  The double-sided tape has a general configuration in which adhesive layers are provided on both sides of a substrate such as a nonwoven fabric or a film. In consideration of preventing the penetration of the slurry into the cushion layer and the rebound resilience adjusting layer, it is preferable to use a film for the substrate. Examples of the composition of the adhesive layer include rubber adhesives and acrylic adhesives. Considering the content of metal ions, an acrylic adhesive is preferable because the metal ion content is low.

  As an adhesive layer for affixing on a polishing apparatus surface plate, a double-sided tape is mentioned, for example. As the double-sided tape, one having a general configuration in which adhesive layers are provided on both sides of a base material can be used as described above. As a base material, a nonwoven fabric, a film, etc. are mentioned, for example. In consideration of peeling from the platen after use of the polishing pad, it is preferable to use a film for the substrate. Moreover, as a composition of an adhesion layer, a rubber adhesive, an acrylic adhesive, etc. are mentioned, for example. Considering the content of metal ions, an acrylic adhesive is preferable because the metal ion content is low.

  The semiconductor device is manufactured through a step of polishing the surface of the semiconductor wafer using the polishing pad. A semiconductor wafer is generally a laminate of a wiring metal and an oxide film on a silicon wafer. The method and apparatus for polishing the semiconductor wafer are not particularly limited. For example, as shown in FIG. 5, a polishing platen 7 that supports the polishing pad 5, a support table (polishing head) 9 that supports the semiconductor wafer 8, and the wafer. This is performed using a backing material for performing uniform pressurization and a polishing apparatus equipped with a supply mechanism for the abrasive 10. The polishing pad 5 is attached to the polishing surface plate 7 by sticking with the adhesive layer 4. The polishing surface plate 7 and the support base 9 are disposed so that the polishing pad 5 and the semiconductor wafer 8 supported by each of the polishing surface plate 7 and the support table 9 are opposed to each other, and are provided with rotating shafts 11 and 12 respectively. A pressure mechanism for pressing the semiconductor wafer 8 against the polishing pad 5 is provided on the support base 9 side. In polishing, the semiconductor wafer 8 is pressed against the polishing pad 5 while rotating the polishing surface plate 7 and the support base 9, and polishing is performed while supplying an abrasive (slurry) 10. The flow rate of the slurry, the polishing load, the polishing platen rotation speed, and the wafer rotation speed are not particularly limited and are appropriately adjusted.

  As a result, the protruding portion of the surface of the semiconductor wafer 8 is removed and polished flat. Thereafter, a semiconductor device is manufactured by dicing, bonding, packaging, or the like. The semiconductor device is used for an arithmetic processing device, a memory, and the like.

  Examples and the like specifically showing the configuration and effects of the present invention will be described below. The evaluation items in Examples and the like were measured as follows.

<Measurement method of rebound resilience value and intrinsic rebound resilience value>
A sample (polishing pad) prepared on a non-repulsive plate (Arai Kasei Co., Ltd., EGR-6) is attached by means used when affixing to the platen, and the rebound resilience value of the polishing layer surface is measured according to JIS K6400 (foam test). It measured based on the impact resilience measuring method-A) of the method. Note that the rebound resilience value of the polishing pad having the polishing layer surface patterned is measured entirely on the pattern surface of the sample, and the maximum value is defined as the rebound resilience value of the polishing layer surface. In order to obtain the maximum value, it is necessary to drop the measurement hard sphere so as not to hit the edge of the convex portion or the concave portion, and to drop the ball at the center of the flat portion of the pattern convex portion.

  The intrinsic rebound resilience value of the polishing layer, cushion layer, and rebound resilience adjustment layer is measured by measuring the rebound resilience value while increasing the thickness, and the value obtained when the rebound resilience value hardly changes depending on the thickness. The elastic value was used. Further, the thickness when the intrinsic rebound resilience value was shown was the critical thickness.

<Evaluation of polishing characteristics>
1. Polishing equipment Okamoto Machine Tool Co., Ltd. SSP600S
2. Polishing conditions (1) Polishing load: 350 g / cm ²
(2) Polishing speed: polishing plate rotation speed = 30 rpm, polishing head rotation speed = 35 rpm
(3) Polishing slurry: SS21 made by Cabot, dripping at 150 ml / min Dress conditions (1) Dresser: Diamond dresser (Abrasive grain number # 100)
(2) Dress load: 450 g / cm ²
(3) Dressing speed: surface plate rotation speed = 30 rpm, dress head rotation speed = 15 rpm
5. Evaluation wafer (1) Pattern film wafer: A thermal silicon oxide film is deposited on an 8-inch silicon wafer to a thickness of 0.5 μm, and a pattern is prepared. Semi-standard wafer including a test pattern with an initial step of 0.5 μm produced by deposition (specimen pattern: 4000 μm × 4000 μm area with 5 lines of width = 270 μm arranged for every space of width = 30 μm)
(2) Solid film wafer: Wafer with notch obtained by depositing a thermal silicon oxide film of 10,000 mm on an 8-inch silicon wafer. Wafer oxide film thickness measurement (1) Device: Otsuka Electronics Co., Ltd., interference type film thickness measuring device Photo- (MCPD-3000 system)
7). Evaluation method (1) Polishing rate: Using a solid film wafer, the operation of polishing for 2 minutes after dressing for 0.5 minutes is repeated while replacing with a new wafer, and the polishing rate at a predetermined measurement position of the wafer is measured. The average value was taken as the polishing rate.

  (2) Wafer slidability: Dressing with a solid film wafer for 25 minutes, then polishing for 2 minutes is repeated while exchanging with a new wafer until the pattern of the polishing layer is consumed. Abnormal noise, vibration, and presence / absence of shaking of the polishing head and the degree thereof were evaluated by visual observation and sound.

  (3) Wafer edge sagging: The wafer and polishing method are the same as those in (2) above, one on the wafer passes through the notch and the center line passing through the center of the wafer, and the other passes through the center point of the wafer. Measure the polishing rate at a point 0.4 inches from the center of the wafer on the two centerlines that intersect the line at right angles to the centerline of the wafer, and from the polishing rate at the measurement point from the center of the wafer to the edge Wafer edge sagging was evaluated.

  (4) Life of the polishing pad: The life when the polishing rate is reduced to 80% of the initial rate by repeating the operation of dressing with a solid film wafer for 25 minutes and then polishing for 2 minutes while replacing with a new wafer. And the total polishing time at that time was measured.

Example 1
(Preparation of polishing layer)
In a reaction vessel, 100 parts by weight of a polyether-based prepolymer (Uniroy, Amplen L-325) and 3 parts by weight of a silicone-based nonionic surfactant (Toray Dow Corning Silicone, SH192) were added, and the temperature was 80 ° C. Adjusted. Using a stirrer, the mixture was vigorously stirred for about 4 minutes so that bubbles were taken into the reaction system at 900 rpm. 26 parts by weight of 4,4-methylenebis (o-chloroaniline) (Ihara Chemical Co., Ltd., Iharacamine MT) preheated to 120 ° C. was added thereto and stirred vigorously for 1 minute, and then reacted to a pan-type open mold. The solution was poured. And when the fluidity | liquidity of the reaction solution poured into the mold was lose | eliminated, it put in an open and post-cured at 110 degreeC for 6 hours, and obtained the foaming polyurethane resin block. This block was sliced using a pando saw type slicer (manufactured by Fecken) to obtain a foamed polyurethane sheet. A 400 mesh belt sander (manufactured by Riken Corundum Co., Ltd., belt paper C54P) was attached to a buffing machine (Amitech Co., Ltd.), and the foamed urethane sheet was ground to obtain an abrasive sheet having a thickness of 1.27 mm. After the abrasive sheet was punched to a diameter of 61 cm, a surface pattern was processed (groove shape: concentric circle, groove width: 0.25 mm, groove depth: 0.4 mm, groove pitch: 1.5 mm) to produce a polishing layer. . The inherent rebound resilience value of the polishing sheet was 31%, and the critical thickness was 6 mm.
(Production of cushion layer)
A 400 mesh belt sander (manufactured by Riken Corundum Co., Ltd., belt paper C54P) is attached to a buffing machine (manufactured by Amitech Co., Ltd.), and a PEF sheet (manufactured by Toray Industries, Inc., polyethylene foam, Torepefu) is ground with a roll to roll to a thickness of 0. An 8 mm cushion layer was obtained. The intrinsic resilience value of the cushion layer was 26%, and the critical thickness was 6 mm.
(Preparation of polishing pad)
Using a laminator on the opposite side of the polishing layer surface pattern processing surface, a double-sided tape (manufactured by Sekisui Chemical Co., Ltd., double tack tape) is applied, and the cushion layer is attached to this double-sided tape using a laminator. It was. The double-sided tape was affixed to the surface corresponding to the polishing surface plate side of the cushion layer using a laminating machine to prepare a polishing pad.

Comparative Example 1
In Example 1, a polishing pad was prepared in the same manner as in Example 1 except that the thickness of the cushion layer was 0.5 mm.

Comparative Example 2
A polishing pad was prepared in the same manner as in Example 1 except that the thickness of the polishing layer was 3 mm and the thickness of the cushion layer was 0.5 mm in Example 1.

Comparative Example 3
In Example 1, a polishing pad was produced in the same manner as in Example 1 except that the thickness of the cushion layer was 5 mm.

Comparative Example 4
In Example 1, a polishing pad was produced in the same manner as in Example 1 except that the thickness of the polishing layer was 0.8 mm.

Comparative Example 5
In Example 1, a polishing pad was produced in the same manner as in Example 1 except that the thickness of the polishing layer was 3 mm.

Example 2
(Preparation of polishing layer)
In Example 1, a polishing layer was produced in the same manner as in Example 1 except that a polishing sheet having a thickness of 2.5 mm was used and the groove depth was increased 1.75 times.
(Production of cushion layer)
In Example 1, a cushion layer was produced in the same manner as in Example 1 except that the thickness was 1 mm.
(Preparation of rebound resilience adjustment layer)
A 400 mesh belt sander (Riken Corundum, belt paper C54P) is attached to a buffing machine (Amitech), and a sponge sheet of ultra-low resilience urethane (Arai Kasei, EGR-5) is ground and thickened. A rebound resilience adjusting layer having a thickness of 0.2 mm was obtained. The intrinsic resilience value of the resilience adjusting layer was 15%, and the critical thickness was 2 mm.
(Preparation of polishing pad)
Using a laminator on the opposite side of the polishing layer surface pattern processing surface, a double-sided tape (Sekisui Chemical Co., Ltd., double tack tape) is applied, and the rebound resilience adjustment layer using a laminator on this double-sided tape Was pasted. The laminating machine was used to attach the double-sided tape to the other side of the impact resilience adjusting layer, and the cushion layer was applied to the double-sided tape using a laminating machine. A polishing pad was prepared by applying the double-sided tape to a surface corresponding to the surface of the cushion layer using a laminator.

Comparative Example 6
(Preparation of polishing layer)
In Example 1, a polishing layer was produced in the same manner as in Example 1 except that a polishing sheet having a thickness of 2.5 mm was used and the groove depth was increased 1.75 times.
(Cushion layer)
In Example 1, a cushion layer was produced in the same manner as in Example 1 except that the thickness was 1 mm.
(Preparation of polishing pad)
A polishing pad was prepared in the same manner as in Example 1 except that the polishing layer and the cushion layer were used.

Example 3
(Preparation of polishing layer)
In Example 1, a polishing layer was produced in the same manner as in Example 1 except that the silicone-based nonionic surfactant was changed to 15 parts by weight. The inherent rebound resilience value of the polishing sheet was 25%, and the critical thickness was 6 mm.
(Production of cushion layer)
A 400 mesh belt sander (manufactured by Riken Corundum Co., Ltd., belt paper C54P) is attached to a buffing machine (Amitech Co., Ltd.), and a urethane sheet (manufactured by Sekisui Chemical Co., Ltd., THS) is ground with a roll-to-roll to a thickness of 1.32 mm. The cushion layer was obtained. The intrinsic resilience value of the cushion layer was 18%, and the critical thickness was 6.6 mm.
(Preparation of rebound resilience adjustment layer)
A 400 mesh belt sander (manufactured by Riken Corundum Co., Ltd., belt paper C54P) is attached to a buffing machine (Amitech Co., Ltd.), and a high resilience sheet (Toyobo, Perprene P90BD) is ground to give a rebound resilience of 0.5 mm in thickness. An adjustment layer was obtained. The rebound resilience adjusting layer had an intrinsic rebound resilience value of 65% and a critical thickness of 2 mm.
(Preparation of polishing pad)
A polishing pad was produced in the same manner as in Example 2.

Comparative Example 7
(Preparation of polishing layer)
A polishing layer was produced in the same manner as in Example 3.
(Cushion layer)
A cushion layer was produced in the same manner as in Example 3.
(Preparation of polishing pad)
A polishing pad was prepared in the same manner as in Example 1 except that the polishing layer and the cushion layer were used.

Example 4
(Preparation of polishing layer)
In Example 1, a polishing layer was produced in the same manner as in Example 1 except that the surface pattern processing was concentric, groove width 0.25 mm, groove depth 0.5 mm, and groove pitch 1.5 mm.
(Production of cushion layer)
A cushion layer was produced in the same manner as in Example 1.
(Preparation of polishing pad)
A polishing pad was produced in the same manner as in Example 1 except that the polishing layer was used.

Example 5
(Preparation of polishing layer)
In Example 1, a polishing layer was produced in the same manner as in Example 1 except that the surface pattern processing was not performed.
(Production of cushion layer)
A cushion layer was produced in the same manner as in Example 1.
(Preparation of polishing pad)
A polishing pad was produced in the same manner as in Example 1 except that the polishing layer was used.

表１から明らかなように、研磨層表面の反発弾性値が３５〜４５％である研磨パッドを用いることにより、ウエハの縁ダレを抑制し、かつ研磨ヘッドの摺動性を抑制して研磨時の異音、振動、及びダンピングを軽減することができる。

As is clear from Table 1, by using a polishing pad having a rebound resilience value of 35 to 45% on the surface of the polishing layer, the edge sag of the wafer is suppressed, and the slidability of the polishing head is suppressed during polishing. Noise, vibration, and damping can be reduced.

Graph showing the relationship between the thickness of the polishing layer and the impact resilience value Graph showing the relationship between cushion layer thickness and impact resilience value Schematic sectional view showing an example of the polishing pad of the present invention Schematic sectional view showing another example of the polishing pad of the present invention Schematic configuration diagram showing an example of a polishing apparatus used in CMP polishing

Explanation of symbols

1: Polishing layer 2: Cushion layer 3: Adhesive layer 4: Adhesive layer 5: Polishing pad 6: Rebound resilience adjusting layer 7: Polishing surface plate 8: Semiconductor wafer 9: Support base (polishing head)
10: Abrasive (slurry)
11, 12: Rotating shaft

Claims (7)

A polishing pad in which a polishing layer made of a polyurethane foam resin, an adhesive layer, a cushion layer, and an adhesive layer for adhering to a polishing apparatus surface plate are laminated in this order, and the adhesive layer of the polishing pad is A polishing pad characterized by having a rebound resilience value on a polishing layer surface of 35 to 45% when adhered to a non-repulsive plate and measured according to JIS K6400. The polishing pad according to claim 1, further comprising a rebound resilience adjusting layer laminated between the polishing layer and the cushion layer via an adhesive layer. The polishing pad according to claim 2, wherein a specific rebound resilience value of the rebound resilience adjustment layer is smaller than a specific rebound resilience value of the cushion layer. The polishing pad according to claim 2, wherein a specific rebound resilience value of the rebound resilience adjustment layer is larger than a specific rebound resilience value of the cushion layer. The polishing pad according to any one of claims 2 to 4, wherein each thickness of the polishing layer, the rebound resilience adjusting layer, and the cushion layer is smaller than the critical thickness of each layer. A method for polishing a semiconductor wafer using the polishing pad according to claim 1. A method for manufacturing a semiconductor device, comprising a step of polishing a surface of a semiconductor wafer using the polishing pad according to claim 1.

Images