JP2007505749A - Polishing pad for chemical mechanical polishing - Google Patents

Abstract

The present invention, in one embodiment, provides a polishing pad (100) for chemical mechanical polishing. The polishing pad includes a polishing body (110). The polishing body comprises a thermoplastic foam substrate (115) having a surface (120) with concave cells (125). The abrasive (130) coats the inner surface (135) of the concave cell. The abrasive includes an inorganic metal oxide containing carbide or nitride. Another embodiment of the invention is a method for manufacturing a polishing pad (200).

Description

  The present invention is directed to chemical mechanical polishing to create smooth, ultra-flat surfaces on articles such as glass, semiconductors, dielectrics, metals and their composites, magnetic mass storage media and integrated circuits. To do.

  Chemical mechanical polishing (CMP) has been successfully used to planarize both metal and dielectric thin films. In one reasonable planarization mechanism, it is believed that the polishing process requires intimate contact between the wafer surface and high portions of pad material in the presence of the slurry. In this scenario, the corroded material resulting from the reaction between the slurry and the wafer surface being polished is removed by shear at the pad-wafer interface. The elastic properties of the pad material greatly affect the final flatness and polishing rate. As a result, the elastic properties are a function of both the intrinsic polymer and its foam structure.

  Historically, polyurethane-based pads have been used for CMP due to their high strength, high hardness, high modulus, and high elongation at break. Such pads have good uniformity and can effectively shrink in shape, but the ability of the pad to remove surface material quickly and uniformly decreases rapidly with use. The time-dependent decrease in material removal rate observed for polyurethane-based pads has been attributed to changes in the mechanical response of such polishing pads under critical shear conditions. It is generally believed that the degradation of the functionality of polyurethane-based CMP pads is due to the degradation of the pads due to the interaction between the pad used during polishing and the slurry.

In addition, in the case of polyurethane pads, decomposition results in surface modifications within the pad and the pad itself that can be detrimental to uniform polishing. Alternatively, in some cases, surface modification of the material used for CMP polishing may improve performance in use. However, such modifications are temporary and therefore frequently require CMP pad replacement or reprocessing. Also, polyurethane pads usually require a trial run period before polishing and further maintenance and reprocessing after use. It is also often necessary to keep the traditional pad moist while the polishing apparatus is inactive. These characteristics undesirably reduce the overall efficiency of CMP when using polyurethane or similar conventional pads.
US Pat. No. 6,579,604 US patent application Ser. No. 10 / 24,104

  Therefore, what is needed is an improved CMP pad that can create a highly flat surface during CMP and improve lifetime, while avoiding experiencing the problems described above.

  To address the prior art deficiencies discussed above, the present invention, in one embodiment, provides a polishing pad for chemical mechanical polishing. The polishing pad includes a polishing body with a thermoplastic foam substrate. The thermoplastic foam substrate includes a surface with concave cells. The abrasive | polishing agent which coat | covers the internal surface of a concave cell contains the inorganic metal oxide containing a carbide | carbonized_material or nitride.

  Another embodiment of the invention is directed to a method for manufacturing a polishing pad. The method includes exposing independent cells within the thermoplastic foam substrate to provide a substrate surface that includes concave cells. The method further includes coating the inner surface of the concave cell with an abrasive containing an inorganic metal oxide, during which the carbides and nitrides are incorporated into the inorganic metal oxide.

  For a more complete understanding of the present invention, reference should be made to the following description taken in conjunction with the accompanying drawings.

  The present invention benefits from the previously unnoticed advantage of using a thermoplastic polymer as a substrate for depositing a uniform coating of abrasive on the concave cells. It has been found that the inner surface of the concave cell forms an excellent container for accommodating a uniform coating of abrasive. Since the center cell has the lowest surface energy, it is assumed that the center of the concave cell serves as an excellent nucleation point for the coating. When coating starts at this point, it is likely that the inner surface of the concave cell will be more easily covered with the abrasive, and as a result, the polishing performance of the pad having such a surface will be promoted.

  The abrasive | polishing agent of this invention contains the inorganic metal oxide containing a nitride or a carbide | carbonized_material. By using such metal oxides to coat the pad surface, the polishing pad surface can advantageously be made permanently hydrophilic. Preferably, the inorganic metal oxide has a lattice of atoms incorporating nitrides or carbides in the lattice. By using such a surface coating of an abrasive, the mechanical properties of the surface of the polishing pad are modified and polishing is improved.

  For example, by changing the content of nitride or carbide in a specific inorganic metal oxide, the mechanical properties of the polishing pad surface can be fine-tuned to make it suitable for the surface being polished. Can do. As a result, by making the mechanical properties of the polishing pad surface suitable for the surface being polished, selectivity of the polishing rate can be improved and defects caused by the process, such as scratches, can be reduced. . Fine-tuning the mechanical properties of the surface should be achieved using plasma-enhanced chemical vapor deposition (PECVD) surface coating and modification of the thermoplastic polymer substrate caused by secondary thermally induced reactions in most thermoplastic substrates Can do.

  One embodiment of the present invention is a polishing pad for chemically and mechanically polishing a semiconductor device. FIG. 1 illustrates an exemplary polishing pad 100 of the present invention. The polishing pad 100 includes a polishing body 110. The polishing body 110 includes a thermoplastic foam substrate 115 having a surface 120 that includes concave cells 125. The abrasive | polishing agent 130 which coat | covers the inner surface 135 of the concave cell 125 contains the inorganic metal oxide containing a carbide | carbonized_material or nitride.

  The abrasive 130 is one or more inorganic metal oxides formed by grafting secondary reactants on the surface 120 of the thermoplastic foam substrate 115 in a plasma assisted chemical vapor deposition (PECVD) process. A ceramic compound composed of As described further below, in the present invention, the PECVD method can be modified to promote the inclusion of carbides or nitrides in the inorganic metal oxide.

  It is preferable that one or both of carbide and nitride are incorporated in the lattice of the inorganic metal oxide. For example, when the inorganic metal oxide includes silicon oxide, the lattice can include silicate within a polymerized Si-O-Si structure having a tetrahedral and distorted tetrahedral arrangement. The nitride can include silicon nitride incorporated into these lattices. Alternatively, if the inorganic metal oxide includes titanium oxide, the nitride can include titanium nitride incorporated into the lattice of titanium oxide. Similarly, silicon carbide and titanium carbide can have inorganic metal oxides incorporated into abrasives 130 containing silicon oxide and titanium oxide, respectively. In some preferred embodiments, a nitride such as silicon nitride includes about 10 mole percent abrasive 130, while in other embodiments, a carbide such as silicon carbide includes about 10 mole percent abrasive. In some preferred embodiments, the abrasive 130 includes both these concentrations of nitride and carbide.

  As the PECVD process takes a long time, the silanol concentration in the abrasive 130 decreases, resulting in a decrease in the oxygen to silicon ratio. In some advantageous embodiments of the polishing pad 100 where the abrasive 130 comprises silicon oxide, the O: Si ratio is at least about 8: 1, and in some cases at least about 9.9: 1. .

  Inclusion of nitrides and carbides in the abrasive 130 provides an additional means not previously noticed that changes the mechanical properties of the polishing pad 100 and thereby changes the polishing properties of the pad. In some preferred embodiments, the polishing pad 100 has a hardness greater than 60 KPa, more preferably greater than 70 KPa. In other preferred embodiments, the polishing pad 100 has a hardness between about 62 KPa and 70 KPa. In certain preferred embodiments, the polishing pad 100 has a modulus of elasticity greater than about 3 MPa, more preferably 4 MPa or greater. In other preferred embodiments, the polishing pad has an elastic modulus between about 4 MPa and 8.2 MPa, respectively. In other embodiments, the polishing pad 100 has a loss modulus of about 0.4 MPa or more, and more preferably at least about 0.6 MPa.

  As disclosed in U.S. Pat. No. 6,579,604 and U.S. Patent Application No. 10/241074, which are incorporated herein by reference, the inorganic metal oxide of abrasive 130 is used as a secondary reactant in the PECVD process. Produced from various oxygen-containing organometallic compounds. For example, the secondary plasma mixture can include a transition metal such as titanium, manganese, or tantalum. However, any metal element that forms a volatile organometallic compound such as a metal ester containing one or more oxygen atoms and can be grafted onto the polymer surface is suitable. Silicon can also be used as the metal portion of the organometallic secondary plasma mixture. In these embodiments, the organic portion of the organometallic reagent can be an ester, acetic acid, or an alkoxy fragment. In a preferred embodiment, the inorganic metal oxide of abrasive 130 includes silicon oxide or titanium oxide, such as silicon dioxide or titanium dioxide; a tetraethoxysilane polymer; or a titanium alkoxide polymer, respectively.

  Other secondary plasma reactants include ozone, alkoxysilane, water, ammonia, alcohol, mineral spirit or hydrogen peroxide. In some preferred embodiments, the secondary plasma reactant includes a titanium ester; a tantalum alkoxide in which the alkoxide moiety includes a tantalum alkoxide having 1 to 5 carbon atoms; a manganese acetate solution dissolved in water; a mineral spirit. Manganese alkoxide, manganese acetate, manganese acetylacetonate, aluminum alkoxide, alkoxy aluminate, aluminum oxide, zirconium alkoxide in which the alkoxide has 1 to 5 carbon atoms, alkoxy zirconate; Magnesium acetate; and magnesium acetylacetonato. Other examples are also contemplated for secondary plasma reactants such as alkoxysilanes and ozone, alkoxysilanes and ammonia, titanium esters and water, titanium esters and alcohols, or titanium esters and ozone.

  Some preferred embodiments of the thermoplastic foam substrate 115 include cross-linked polyolefins such as polyethylene, polypropylene, and combinations thereof. In certain preferred embodiments, the thermoplastic foam substrate 115 comprises a closed cell foam of cross-linked homopolymer or copolymer. Examples of independent cell foam cross-linked homopolymers including polyethylene (PE) include Volara ™ and Volextra ™ from Voltek (Lawrence, Mass.); JMS Plastics Supply, Inc. Aliplast (TM) from (Neptune, NJ); or Senflex T-cell (TM) (Rogers Corp., Rogers, CT). Examples of independent cell foams of crosslinked copolymers including polyethylene and ethylene vinyl acetate copolymer (EVA) include: Volara ™ and Volextra ™ (from Voltek Corp.); Senflex EVA ™ (from Rogers Corp.). ); And J-foam ™ (from JMS Plastics JMS Plastics Supply, Inc.).

  In other preferred embodiments, the independent cell thermoplastic foam substrate 115 is a crosslinked ethylene vinyl acetate copolymer and a low density polyethylene copolymer (ie, preferably between about 0.1 and about 0.3 gm / cc). ). In other advantageous embodiments, the blend has an ethylene vinyl acetate: polyethylene weight ratio of between about 1: 9 and about 9: 1. In certain preferred embodiments, the blend comprises about 5 to about 45% by weight of EVA, preferably about 6 to about 25% by weight, and more preferably about 12 to about 24% by weight. Such a blend desirably produces a stand-alone cell 140 having a small dimension (eg, a diameter between about 10 and about 500 microns, more preferably between 50 and 150 microns) of the thermoplastic foam substrate 115. It is thought to contribute. Further, in a more preferred embodiment, the blend has an ethylene vinyl acetate: polyethylene weight ratio of between about 0.6: 9.4 and about 1.8: 8.2. Further, in a more preferred embodiment, the blend has an ethylene vinyl acetate: polyethylene weight ratio of between about 0.6: 9.4 and about 1.2: 8.8.

  As further illustrated in FIG. 1, the thermoplastic foam substrate 115 includes independent cells 140. As used herein, the term independent cell 140 refers to any volume defined by a film inside the substrate 115 and occupied by air or other gas used as a blowing agent such as nitrogen or helium. . The independent cell 140 forms a substantially concave cell 135 that is formed when the substrate 115 is peeled. However, the concave cell 135 need not have a smooth or curved wall. Rather, the concave cell 135 can have an irregular shape and dimensions. Several factors, such as the composition of the thermoplastic foam substrate 115 and the procedure used to manufacture the thermoplastic foam substrate 115, can affect the shape and dimensions of the independent cells 140 and the concave cells 135.

  As further illustrated in FIG. 1, the thermoplastic foam substrate 115 can optionally be bonded to a substrate 145. In some preferred embodiments, the substrate 145 is rigid. The rigid backing limits the compressibility and elongation of the foam during polishing, thereby advantageously reducing the corrosion and dishing effects that occur during CMP metal polishing. In some cases, substrate 145 comprises high density polyethylene (ie, greater than about 0.98 gm / cc), and more preferably concentrated high density polyethylene. In some cases, bonding to the thermoplastic foam substrate 115 is accomplished by chemical bonding using a conventional adhesive 150, such as an epoxy or other material known to those skilled in the art. In some preferred embodiments, bonding is accomplished by extrusion coating a molten substrate 145 onto a thermoplastic foam substrate 115. In still other preferred embodiments, the substrate 145 is thermally welded to the thermoplastic foam substrate 115.

  Another aspect of the present invention is a method for producing a polishing pad for chemical mechanical polishing. 2-4 illustrate selected steps in an exemplary method of manufacturing the polishing pad 200. FIG. Any of the embodiments of the polishing pad and its components, including the primary and secondary plasma reactants described above, can be incorporated into the method of manufacturing the polishing pad 200.

  Referring now to FIG. 2, a portion of the polishing pad 200 after exposing the independent cells 210 inside the thermoplastic foam substrate 220 to provide a substrate surface 230 with concave cells 240 is shown. An assembled polishing pad 200 is shown. Concave cell 240 is formed on substrate surface 230 by stripping. As used herein, the term stripping means any method of cutting off a thin layer of the surface of the substrate 220 to expose the concave cells 240 inside the thermoplastic foam substrate 220. Stripping can be accomplished by using conventional techniques well known to those skilled in the art.

  3 and 4 illustrate selected stages in the surface coating process. Referring initially to FIG. 3, in PECVD, a partially completed polishing pad 200 is illustrated after exposing the substrate surface 230 to a first plasma reactant and then to a secondary plasma reactant. Yes.

  Upon exposure to the initial plasma reactant, a modified surface 310 of the thermoplastic foam substrate 220 is formed. It is important to carefully control the plasma treatment conditions and time to avoid undue damage to the thermoplastic foam substrate 220. For example, if the temperature of the radio flow discharge electrode is too high or uncontrolled, the thermoplastic foam substrate 220 is melted, twisted, or cracked in the substrate. In some preferred embodiments of this method, the temperature of the radio flow discharge electrode is maintained between about 20 ° C. and 100 ° C., and more preferably between about 30 ° C. and 50 ° C. In some cases, high frequency operating power between about 250 to about 1000 watts, and more preferably between about 300 watts to 400 watts is used. In certain preferred embodiments, the initial plasma reactant comprises neon and, more preferably, an inert gas such as argon or helium. In some cases, the initial plasma exposure proceeds for between about 1 second and 60 seconds, more preferably about 30 seconds. In some embodiments, the PECVD reaction chamber is maintained between about 300 mTorr and about 400 mTorr, and more preferably about 350 mTorr.

  Also shown in FIG. 3 is a partially completed polishing pad 200 after exposing the modified surface 310 to a secondary plasma reactant. In some preferred embodiments, the secondary plasma reactant comprises tetraethoxysilane (TEOS) or titanium alkoxide (TYZOR). In some cases, the secondary plasma reactant also includes a primary plasma reactant, such as TEOS or TYZOR vapor mixed with helium or argon gas. Exposure to the secondary plasma reactant causes grafting of the secondary plasma reactant to the modified surface 310 to form an abrasive 320 that includes an inorganic metal oxide. The abrasive 320 covers the inner surface 330 of the concave cell 240.

  Again, the conditions and time of exposure to the secondary plasma reactant are carefully controlled to avoid damaging the thermoplastic foam substrate 220 or the abrasive 320 and to achieve a lasting coating of the abrasive 320. In some embodiments, the PECVD reaction chamber is maintained between about 300 millitorr to about 400 millimeters, and more preferably about 350 millimeters. In some examples, the temperature of the radio flow discharge electrode is maintained between about 20 ° C. and 100 ° C., and more preferably between about 30 ° C. and 50 ° C. In some cases, high frequency operating power between about 50 to about 500 watts, and more preferably between about 250 to about 350 watts is used.

  Referring now to FIG. 4, a partially completed polishing pad 200 is illustrated after exposure to a secondary plasma reactant for at least about 30 minutes. In some preferred embodiments, the exposure to the secondary plasma reactant is performed for about 30 minutes to about 60 minutes. In another preferred embodiment, the exposure to the secondary plasma reactant is performed for about 30 to about 45 minutes. Such a length of time exposure advantageously enhances the incorporation of nitrides or carbides, or both, into the inorganic metal oxide of the abrasive 320. In some embodiments of the method, nitrogen gas is contained within the independent cells 410 of the thermoplastic foam substrate 220. Nitrogen gas reacts with the secondary plasma reactant to form nitride. In another embodiment of this method, at least a portion of the thermoplastic foam substrate 220 reacts with the secondary plasma reactant to form carbide. For example, in some embodiments, carbon radical species within a depth of about 1 micron of thermoplastic foam substrate 220 from the modified surface 310 can react with the secondary plasma reactant.

  As discussed above and illustrated in the examples below, depositing abrasive 320 by PECVD modifies the surface properties of certain thermoplastic foams 230, such as polyolefin foams. Surface coating of the thermoplastic foam surface 310 with abrasive 320 for up to about 30 minutes occurs by one mechanism. However, after this time, the surface coating occurs by different mechanisms. As a result, this produces a polishing pad surface 410 that clearly differs in surface micromechanics and chemical properties as a function of coating time.

In some cases, as the coating time increases, the temperature of the thermoplastic foam substrate 220 increases. As a result, this results in outgassing of nitrogen gas that is used in foaming the substrate 220 and is present in the independent cells 410 of the substrate 240. For example, in some embodiments where the abrasive 320 includes silicon oxide, the released nitrogen gas reacts with silicon species on the pad surface and in stoichiometric conversion from SiO ₄ to Si ₃ N ₄ . This is the basis for forming the Si ₃ N ₄ species. Of course, a similar reaction can occur in embodiments where the abrasive comprises other metal oxides such as titanium oxide.

  Similarly, if the coating time is long, a significant amount of carbon radicals are generated on the surface 310 of the pad 200 by ion bombardment of the surface 310 of the thermoplastic foam substrate 240. For example, in some embodiments where the abrasive 320 includes silicon oxide, these radicals react with the silicon species to form silicon carbide (SiC), after which the silicon carbide coats the pad 200. Captured inside.

When species such as Si ₃ N ₄ and SiC are incorporated into the abrasive 320, it modifies the properties of the polishing pad 200, for example compared to a starting thermoplastic foam substrate or a substrate that has received a brief coating, Increases rigidity and hardness, and changes elastic modulus.

  Having described the invention, the same should be made clearer by referring to the following experiments. It should be understood that the experiments are presented for illustrative purposes only and should not be construed as limiting the invention. For example, the experiments described below may be performed in a laboratory environment, but those skilled in the art will adjust specific numbers, sizes, and quantities to appropriate values for a full-scale device environment. Should be able to.

Experiment The experiment was to 1) characterize the chemical composition of the thermoplastic foam substrate coated with the abrasive as a function of coating time, and 2) the mechanical properties of the thermoplastic foam substrate coated with the abrasive. And 3) to characterize the polishing properties of the polishing pad coated with the abrasive as a function of coating time.

  The thermoplastic foam substrate was molded into a circular polishing pad having a diameter of approximately 120 cm and a thickness of approximately 0.3 cm. A commercially available thermoplastic foam substrate called "J-60SE" (JMS Plastics, Neptune, J-foam from NJ) has about 18% EVA, about 16 to about 20% talc, and the remaining polyethylene and Includes other additives present in commercially available substrates, such as silicates. The J-60 sheet was peeled off with a commercially available cutting edge (model number D5100 K1, from Fecken-Kirfel, Aachen, Germany). The sheet was then washed manually with a water / isopropyl alcohol solution.

  Then, by placing the stripped substrate in a conventional commercially available radio frequency glow discharge (RFGD) plasma reaction chamber (model PE-2, Advanced Energy Systems, Medford, NY) with a temperature controlled electrode arrangement, The J-60SE substrate was coated with an abrasive containing tetraethoxysilane (TEOS). The plasma treatment of the substrate was started by introducing argon as a primary plasma reactant for 30 seconds in a reaction chamber maintained at 350 mm. The electrode temperature was maintained at 30 ° C. and 300 watts of high frequency operating power was used. A secondary reactant comprising TEOS mixed with He or Ar gas was then introduced at a flow rate of 0.10 SLM for about 0 to about 45 minutes. The amount of secondary reactant in the gas stream was adjusted by the back pressure (BP) of the secondary reactant monomer vapor at the temperature of the monomer reservoir (MRT; 50 ± 10 ° C.).

  The polishing characteristics of the J60SE polishing pad were tested by polishing a wafer with a tungsten surface approximately 4000 angstroms thick and a tantalum barrier layer 250 angstroms below. The polishing characteristics of tungsten were evaluated using a commercially available polishing machine (manufacture number EP0222; Ebara Technologies, Sacramento, Calif.). Unless otherwise stated, the removal rate of tungsten polishing was evaluated using a down force of about 25 kPa (downward force) to suppress the substrate and a table speed of about 100 to about 250 rpm (Production Number MSW2000; Rodel, Newark). , From Delaware). A conventional slurry with a pH adjusted to about 2 (manufacture number MSW2000; from Rodel, Newark, Del.) was used.

FIG. 5 shows the FTIR spectrum of the substrate surface after coating TEOS for various lengths of time. Spectra were obtained using an FTIR spectrometer (FTIR 1727, Perkin-Elmer System detector: equipped with a series-I FTIR microscope (MCT detector) and having a spectral range of 10,000 to 370 cm ⁻¹ ). Signals of about 1010 and about 950 cm ⁻¹ are applied to the asymmetric Si—O—Si stretch of silica and the Si—O—X stretch of silicate, respectively, where X is a polymer that is not in a tetrahedral configuration. -(Si-O-Si) _n -indicates a structure). The signal at 850 cm ⁻¹ is due to free and associated silanol (Si—O—H). Silanol associates through hydrogen bonding and the degree of association increases as the surface concentration of silanol increases.

  As shown in FIG. 6, both of these signals are monotonic until the coating time reaches about 30 minutes, as the surface concentration of Si—O decreases net as the coating time increases. Diminished. Thereafter, the deposition kinetics and mechanism changed, indicating that the Si-O concentration on the surface increased. The latter observation contradicts the generally accepted concept of TEOS deposition mechanism and kinetics and facilitates further study of the coating process, particularly at coating times later than 30 minutes.

  The nano-indentation test was used to evaluate the mechanical properties of the surface-coated coating, more specifically, the elastic modulus and hardness. Indentation was performed on thermoplastic foam substrates coated with abrasive at various times. A Nanoindenter installed at NANOTEST 600®, Advanced Material and Charactarization Facility (AMPAC, Orlando, FL) was used for all measurements. The machine is placed on a vibration isolation table and housed in a temperature controlled box. Two separate heaters installed on both sides of the front of the box form a thermal barrier. A temperature controller with an expected stability of ± 0.1 ° C. was set to a value about 2 or 3 ° C. above room temperature. The indenter could be stabilized for at least 30 minutes to obtain thermal stability before starting the experiment.

  Indentation parameters such as indenter indenter type, maximum depth and loading / unloading rate were determined by conducting a preliminary test of the coated polishing pad. The polishing pad surface was found to contain irregularities and pores of various sizes on the order of a few microns. Based on these observations, a spherical indenter with a tip diameter of about 1 mm was selected so that the indenter could fully test the pad material. For the same reason, a spatial resolution of over 500 microns was selected. The indentation was carried out in an ultra-low load range where the initial load, which is a machine parameter, was 0.1 mN. The control parameter was set to a controlled depth and each pad was pushed in at various depths with a maximum depth of 10,000 nanometers. Usually, the result of averaging 10 indentations was examined. The results were evaluated using the Oliver-Pharr method and the Hertz method to prove its validity. The load-depth (Ph) curve obtained from the nanoindentation indenter was analyzed using the Oliver-Pharr method.

  The polishing pad coated with abrasive showed non-uniform penetration during the indentation experiment. Several characteristics can be seen from visual inspection of the indentation (Ph) curve. As shown in FIG. 7, separate events labeled “pop-in” (Xb), “pop-out” or “kink-back” (Xc) can be identified from the curve. For example, “pop-in” occurs during the compression cycle when the tip of the indenter suddenly penetrates into the sample. These events correlate with several experimental parameters such as coating time, loading / unloading speed and indenter depth. Due to the mixed response exhibited by the coated polishing pad, the “pop-in” event occurred more often in the case of indentation indentation depths near 1000 nanometers, “pop-out” It became clear that the event appeared to be unaffected by the indentation depth, and that the loading / unloading rate affected both of these events.

  By further analyzing the Ph curve for various depths and various coating times, it was found that the loading curve rapidly increased and decreased. This X coordinate or the depth of this transition point is called Xa. Similarly, Xb and Xc are the corresponding X coordinates or depth in nanometers. Such non-uniform penetration of the indenter tip into the coating probably results from the beginning of plastic deformation. Plastic deformation is an important attribute of the CMP pad and affects the efficiency of the CMP process. Therefore, it was assumed that the above event in the load-depth curve is an indication of pad performance. The initial surface penetration event (Xa) was found to be a function of PECVD coating time, maximum penetration depth and loading rate. For example, FIG. 8 shows a representative correlation between Xa and TEOS coverage time. This data suggests that Xa is related to the thickness of the surface foam modified by the dielectric coating.

  9 and 10 illustrate the hardness and elastic modulus, respectively, of a polishing pad having various TEOS coating times, calculated using the Oliver-Pharr method. It has been found that the effective surface modulus and hardness increase with increasing coating time. At coating times of 30, 40 and 45 minutes, the polishing pads had hardness values of about 65 KPa, 62 KPa and 70 Kpa, respectively. When the polishing time was shorter, the hardness was 60 KPa or less. At coating times of 30, 40 and 45 minutes, the polishing pad had modulus values of about 4 MPa, 5.5 MPa and 8.2 MPa, respectively. When the polishing time was shorter, the value of the elastic modulus was 3 MPa or less.

  The change in mechanical properties of the pad surface is due to the effect of the coating deposited on the foam substrate. These data indicate that the previously described discontinuities in the FTIR data suggest a change in pad surface chemistry rather than a net removal of the coating derived from TEOS.

Dynamic machine of coated polishing pad sample using a commercial device operating from -125 to 200 ° C at a programmed heating rate of 5 ° C / min, in tension mode, frequency 1 Hz, amplitude 10 microns Analysis (DMA) was performed. Liquid nitrogen was used to achieve a temperature below ambient temperature. Prior to the measurement, the sample was equilibrated for 10 minutes at a predetermined initial temperature. All of the polishing pad samples are manufactured to have the same dimensions of 15 cm × 5 cm and are vacuum dried (30 ° C. at about 1 × 10 ⁻² ) for 24 hours prior to DMA measurement to account for moisture effects I didn't need it.

  As shown in FIG. 11, a study by DMA shows that the loss modulus changes abruptly for a long period of PECVD coating. Compared to the value of loss modulus changing from 0.37 to 0.23 during the coating time of 10 minutes to 40 minutes, the loss modulus is about 0.6 MPa when the coating time is 45 minutes. Soars.

  This is in contrast to the generally accepted view that PECVD coating only modifies the surface of the substrate. This surprising result suggests that another process occurs during surface coating that alters the bulk mechanical properties of the foam substrate. It was assumed that the residual reactants in the thermoplastic foam substrate continued to react in a time dependent manner during the PECVD coating.

X-ray photoelectron spectroscopy (XPS) was used to further characterize the surface modification of the polishing pad coated at various times. A commercial X-ray photoelectron spectrometer was operated at a reference pressure of 10 ⁻¹⁰ and the spectrometer was calibrated using a metal gold standard (Au (4 _{f7 / 2} ): 84.0 ± 0.1 eV). A non-monochromatic MgK X-ray source with an energy of 1253 eV at a power of 250 W was used for the analysis. The charging shift caused by the polishing pad sample was removed by using a binding energy scale based on the binding energy of the hydrogen portion of the adventitious carbon line at 285.0 eV. Peak deconvolution was performed using commercially available software.

XPS analysis provided some insight into the topographic chemistry resulting from the adsorption and dissociation of TEOS. The carbon (1s) signal was resolved into three main peaks: two peaks at about 285.0 eV corresponding to C—C and C—H bonds, and about 286.5 eV corresponding to C—O bonds. It is a peak observed in. The peak centered at about 289 to about 289.3 eV is attributed to the carbamide [—O—C (NH ₂ ) ═O] functionality derived from the residue of the blowing agent used in the process of manufacturing the thermoplastic foam substrate. Assigned. For the samples coated for 40 and 45 minutes, another peak was observed around 283.6 eV, which was temporarily assigned to the C—Si bond.

  FIG. 12 shows Si (2p) envelopes obtained from pads after coating with TEOS for (a) 10 minutes, (b) 20 minutes, (c) 30 minutes, (d) 40 minutes, and (e) 45 minutes. Fig. 6 shows an exemplary peak envelope fitted to the XPS signal. The peaks were identified as (1) Si—O, (2) silicate, (3) Si—N, and (4) Si—C bonds. Each spectrum was deconvoluted into two major peaks at about 102.3 and about 103.4 eV, corresponding to bonds in the silicate and Si-O species, respectively.

  These data indicate that if the coating time is short (eg, less than about 30 minutes), the pad surface is rich in silanol and is consistent with the TEOS film being deposited at low process temperatures. As further shown in FIG. 13, the intensity ratio of oxygen to Si calculated from XPS data is high at the beginning of the coating, indicating that the concentration of silanol in the deposited coating is high. The silanol concentration decreases with the coating time up to 30 minutes and then begins to rise. For example, as shown in FIG. 9, after coating for about 10, 20, 30, 40 and 45 minutes, the O: Si ratio is about 7.4, 4.8, respectively. Equal to 3.6, 7.1 and 9.9.

  Referring again to FIG. 12, for the coating times of 30, 40 and 45 minutes, a small peak corresponding to the Si—N bond is observed at about 102.1 eV. During this same period of time, the ratio of Si to N decreases rapidly, indicating that the nitrogen species on the surface has increased.

These observations suggest that PECVD-based coating involves competition between several processes. The PECVD-based coating forms silica and silicate on (SiO _x and SiO ₂ ) on the foam surface. Furthermore, the surface chemistry of the substrate varies as a function of coating time. In a coating time of less than 30 minutes, net etching of the deposit occurs due to the impact of Ar ions on the surface. Also, the sample is heated by the plasma, and a thermal process occurs. As the coating time increases, the silicate content on the substrate surface begins to decrease, the pad becomes dense and the pad hardness increases.

With a coating time of 30 minutes or more, the substrate temperature becomes sufficiently high, causing release of the nitrogen gas used to foam the substrate or decomposition of the residual blowing agent left in the foam, producing nitrogen gas. Nitrogen reacts with Si-containing intermediates in the gas phase or Si species on the pad surface to form nitrides such as Si ₃ N ₄ while consuming SiO ₂ . Such nitrides are incorporated into the abrasive at a concentration of up to 10 mol%. Furthermore, at such coating times, a significant amount of carbon radicals are generated on the pad surface due to ion bombardment of the foam surface. These radicals react with silicon species to form carbides such as, for example, silicon carbide (SiC), and the formed carbides are incorporated into the abrasive at a concentration of up to 10 mol%.

  Figure 14 compares the relative blanket-tungsten removal rate (W-RR) and static coefficient of friction (COF) for thermoplastic foam substrates coated at different times with TEOS. is there. Both W-RR and COF increase up to 30 minutes with increasing coating time, meaning that the abrasive thickness increases. At coating times from 30 to 60 minutes, both W-RR and COF decrease and then increase. These results show that because the micromechanics and chemical properties of the surface are different, the pad is coated on the surface coated in more than 30 minutes, with one mechanism for the surface coated in time up to 30 minutes. Suggests that it looks like an abrasive with a different mechanism.

It is sectional drawing of the polishing pad of this invention. FIG. 6 is a cross-sectional view of selected stages in the method of the present invention for manufacturing a polishing pad. FIG. 6 is a cross-sectional view of selected stages in the method of the present invention for manufacturing a polishing pad. FIG. 6 is a cross-sectional view of selected stages in the method of the present invention for manufacturing a polishing pad. 2 is a near infrared spectrum of a thermoplastic foam polishing pad sample after coating at various times with an abrasive precursor comprising tetraethoxysilane (TEOS). 2 is a graph showing the change in the near infrared signal of a representative thermoplastic foam polishing pad after exposure to TEOS for various times. 2 is an exemplary indentation curve of a thermoplastic foam polishing pad after being coated with TEOS. FIG. 6 is a graph showing a representative change in (Xa), ie, “pop-in” of a thermoplastic foam polishing pad, as a function of TEOS coating time. 2 is a graph showing a typical change in hardness of a thermoplastic foam polishing pad as a function of TEOS coating time. 2 is a graph showing a typical change in elastic modulus of a thermoplastic foam polishing pad as a function of TEOS coating time. 2 is a graph showing a typical change in storage and loss modulus of a thermoplastic foam polishing pad as a function of TEOS coating time. 2 is a representative XPS spectrum of a polishing pad after coating with TEOS at various times. It is a graph which shows the change of the value of the ratio of the intensity | strength of oxygen with respect to Si calculated from the XPS spectrum of the polishing pad after coat | covering with TEOS for various time. 6 is a graph showing the change in relative blanket tungsten removal rate (WRR) and static coefficient of friction (COF) of a thermoplastic foam polishing pad as a function of TEOS coating time.

Claims (10)

A chemical body comprising a polishing body comprising a thermoplastic foam substrate having a surface containing concave cells, and an abrasive comprising an inorganic metal oxide containing carbide or nitride covering the inner surface of the concave cells Polishing pad for polishing.   The polishing pad of claim 1, wherein the carbide or nitride is incorporated in the inorganic metal oxide lattice.   The polishing pad according to claim 2, wherein the nitride includes silicon nitride or titanium nitride, and the carbide includes silicon carbide or titanium carbide.   The polishing pad according to claim 1, wherein the polishing pad has a hardness of 70 KPa or more.   The polishing pad according to claim 1, wherein the polishing pad has an elastic modulus of 5 MPa or more. A method for producing a polishing pad for chemical mechanical polishing comprising:
Exposing the independent cells within the thermoplastic foam substrate to provide a substrate surface comprising concave cells, and coating the inner surface of the concave cells with an abrasive comprising an inorganic metal oxide, the coating comprising: A method in which carbides and nitrides are incorporated into the inorganic metal oxide during operation. A coating exposes the substrate surface to an initial plasma reactant in a plasma assisted chemical vapor deposition (PECVD) method to produce a modified surface thereon, and the modified to a secondary plasma reactant in the PECVD method. The method of claim 6, comprising exposing an exposed surface to form the abrasive.   8. The method of claim 7, wherein exposure to the secondary plasma reactant continues for 30 to 60 minutes.   The method of claim 7, wherein nitrogen gas is contained inside the independent cell, and the nitrogen gas reacts with the secondary plasma reactant to form the nitride. The method of claim 7, wherein the thermoplastic foam substrate reacts with the secondary plasma reactant to form the carbide.

Images