JP4746540B2 - Functionally graded pad assembly for chemical mechanical planarization - Google Patents

Description

(Citation of related application)
We claim the following benefits under 35 USC 119:
a) US Provisional Application No. 60 / 475,305, Title of Invention MODULATION OF POLISH PAD PROPERIES THROUGH A RAMIFFED ALKYL AND ARYL FUNCTIONAL GROUP OF PROCURSORS, filed Jun. 3, 2003 (57718-30003).
b) US Provisional Application No. 60 / 475,374, Title of Invention SYNTHESIS OF A GRADED PAD STRUCTURE FOR CHEMICAL-METHONIC PLANARIZATION USING REACTIVE INJECTION MOLDING, filed Jun. 3, 2003;
c) U.S. Provisional Application No. 60 / 475,283, Title of Invention REINFORCED CHIP-CUSTOMIZED PADS FOR CHEMICAL-MECHANICAL PLANARIZATION PROCESSES, filed Jun. 3, 2003 (57718-30005.00); and d) U.S. Provisional Application No. 60 / 475,307, Title of Invention INCORRATION OF MULTI-FUNCTIONAL CURRENT AGENTS IN PADS USED FOR A CHEMICAL-MECHANICAL PLANARIZATION PROCESS, filed on June 3, 2003 (57718-30006.00)
The entire contents of these are hereby incorporated by reference.

(Field)
We describe a polishing pad suitable for chemical mechanical planarization (CMP) of a semiconductor wafer. Specifically, our specification relates to a polishing pad that is tuned to provide a spatially gradual change in material / tribological properties. These differential gradual changes can be used to achieve custom polishing of various dielectric and metal films during silicon integrated circuit (IC) processing.

(background)
Chemical mechanical planarization (CMP) is used to planarize individual layers (dielectric layers or metal layers) during the manufacture of integrated circuits (ICs) on semiconductor wafers. CMP removes unwanted topological features of the IC on the wafer (eg, metal deposition following a damascene process). CMD can be used to remove excess oxide that forms during the shallow trench isolation process and includes interlevel dielectric (ILD) and intermetallic, including those with low dielectric constant (low k) materials The dielectric (IMD) can be planarized.

  CMP typically uses a combination of a reactive liquid medium and a polishing pad surface to provide appropriate mechanical and chemical control in achieving flatness. Either or both of these reactive liquids and the polishing surface of the pad contain inorganic particles (often dimensions are nano-sized) to allow chemical reactivity and / or mechanical activity of the CMP process on the wafer. Can be enhanced. Pads commonly used today are often made from a substantially rigid microporous polyurethane material. This material achieves a simultaneous function of providing uniform slurry transport, providing distribution and removal of the resulting particle product, and providing uniform distribution of pressure applied across the wafer. Selected.

  In the CMP process, the chemical interaction between the slurry fluid and the wafer results in the formation of a chemically modified layer at the polishing surface. At the same time, the abrasive in the slurry mechanically interacts with this chemically modified surface layer, thereby resulting in material removal. The abrasive particles are generally associated with mechanical polishing in the material removal process. When viewed at the nanoscale level, the kinetics of thin surface layer formation and subsequent removal controls CMP results (ie, removal rate (RR), surface flatness, surface defects, and slurry selectivity). . Thus, the local material / tribological / mechanical properties of the pad are important for both local and global planarization during the CMP process.

  The material removal rate (RR) in the CMP process is a function of a number of factors, but in particular a function of the abrasive concentration of the slurry and the average coefficient of friction ("f") at the pad / slurry / wafer interface region. The degree of normal and shear forces during CMP, as well as “real time f”, depends on both the tribology of the pad and the hydrodynamics of the slurry. Recent studies (refs. 1 and 2) have shown that pad material compliance, contact area between the wafer and slurry abrasive particles, and the degree of lubricity of the system, all during any CMP process. , Show a significant role.

A Stribeck curve can be used as a way to quantify the role of these factors in the CMP process. As a background, the Stribeck curve is a plot of the average friction coefficient “f” against the Sommerfield number (So [So = μV / p · d _eff ]), and the wafer, rotating pad, and contained abrasive particles (Where μ is the viscosity of the slurry, V is the relative pad-wafer velocity, and d _eff is α · R _a + [1-α] d. _groove ). Where R _a is the average pad roughness, d _groove is the depth of the pad groove, and “α” (a dimensionless area parameter according to the wafer pressure factor) = A _up-features. / A _{flat pad} ; P _eff = p / α.

  The general Stribeck curve shown in FIG. 1 shows three different regimes: a “boundary lubrication regime”, a “transition regime”, and a “hydrodynamic lubrication regime”. First, in the “boundary lubrication regime”, all solids are in intimate contact with the abrasive particles of the slurry and “f” remains constant with respect to So. In this regime, larger values of “f” and “RR” occur.

  Second, at an intermediate “So” value, ie “partial lubrication regime” or “transition zone regime”, the pad and wafer are either due to the presence of a fluid film layer, or perhaps in part Due to the roughness of the, there is no direct contact. In this transition regime, the slope of “f” is negative.

Finally, the more viscous fluid layer present in the “hydrodynamic lubrication” regime or at higher values of So yields smaller “RR” and “f” values. As So increases, any slight increase in “f” is the fluid flow field (f = k · So ^β , where k is a constant and β is the tribological mechanism indicator of the lubrication regime. ) Seems to be due to the presence of spirals. When β> 0, the boundary lubrication mechanism dominates. If β is a negative number, the partial lubrication regime is conspicuous and the lower RR (RR = K _Pr · p _eff · V, where K _Pr is the Preston constant, which is the chemical mechanical of the process Depending on the specific situation).

The partial lubrication regime (or transition zone regime) results in increased pad life. However, operation in the boundary lubrication regime provides greater stability, control, and predictability of RR, and wafer uniformity. Both “f” and “β” have a linear relationship with K _Pr and thus have a linear relationship with RR. K _Pr is inversely proportional to the pad storage modulus. Thus, the softer the pad as the compressibility increases, the greater the RR, and the harder the pad as the compressibility decreases, the smaller the RR. Softer pads experience greater shear forces at the wafer tip during polishing. This is because softer pads are relatively more compressed and thus create obstacles that the wafer must overcome continuously as the wafer moves. At the microscale, pad roughness at the wafer / pad interface is reduced, thereby increasing both the shear, f and β components.

Pad material compliance, microtribology, and nanotribology change the shape of a particular Stribeck curve, along with slurry rheology, variously, its lubricity, groove or perforation configuration, abrasive concentration, pH, and temperature, respectively. And the relative degree of different lubrication regimes and the values of K _Pr , f and β can be varied.

  The pads that we describe herein may have different areas of pad tribology. By adjusting the pad tribology and choosing the appropriate operating regime of the Stribeck curve, the lubricity for these adjusted or different regions, the simultaneous local and overall planarization, can be achieved with a stack of different materials ( Wafers with metal / barrier, such as found in shallow trench isolation (STI), which can be found in oxide / nitride constructs); or materials bent for sub-90 nm technology (eg, , Wafers with low-k and strained Si materials and silicon-on-insulator (SOI) structures, and complex device designs and structures (eg, “system-on-a-chip”) (So ) And various vertical gate structures (such as those apparent in FinFETs) can be achieved.Our tuned or functionally gradual pads have a wide range of varying patterns Can be designed for wafers with density and chip size.

  Conventional open cell and discontinuous cell polymer pads with substantially uniform tribological, chemical, and frictional properties have been suitable for use in CMP. However, when 250 nm CMOS technology was introduced, it was investigated that there was a limit to the compatibility of these pads. For technologies below 250 nm, for example, technologies with increased design complexity (eg, SoC), process (eg, SOI, FinFET) or material (eg, Cu or low-k materials), especially chip pattern density and When combined with substantial variations in increased chip size, the use of these conventional open hole and closed hole uniform polymer pads allows for chip yield, compared to the same measurements on previous generation wafers, Accompanied by degradation of device performance and device reliability.

  Various attempts to change pad thickness (laminated and non-laminated) and pad surface structure (eg, perforations, K-grooves, XY grooves, and K-groove / XY groove combinations) include chip pattern density, chip It did not specifically address the impact that size, structural complexity, and dielectric / metal process flow have on final chip breakdown, device performance, and integrated circuit reliability.

  The functionally graded pads described herein and having spatial variations of pad tribological / material properties, even for those using sub-130 nm technology nodes, have a CMP process for these new technologies. Suitable for

(Summary)
Described is a functionally graded polishing pad for a CMP process that has a polishing surface for polishing silicon-containing wafers that are configured for use in a CMP procedure. Made from polishing pad. Here, the polishing surface is a single piece, is substantially flat and includes at least two areas having different physical properties. The at least two areas may have separate boundaries or boundaries formed with a mixture of constituent polymers. The at least two areas may each comprise a polymeric material having a different composition, and the region between the areas may comprise a mixture of polymeric materials having a different composition.

  The shape of at least one of the at least two areas may be an annular shape, an island shape, or an arbitrary shape. The at least two areas each contain different polymeric materials having different physical parameters, and the change in composition between the at least two areas may be a step change or a continuous change. It may be a combination thereof. The pad may have an edge and a center, the edge comprising one area, the center comprising another area, and a composition change between the edge and the center. May be continuous, stepped, or in another form.

  This functionally graded polishing pad can be made by a process of continuous injection molding or co-injection molding of at least two polymer compositions to form the at least two areas described above. The functionally incremental polishing pad comprises the at least two polymers comprising a block copolymer (perhaps the block copolymer has a different constitutive polymer composition in the block copolymer as the distance increases). It may have at least one of the compositions.

  In progressively changed pads, important mechanical properties of the pad, such as hardness, module, etc., and important physical properties, such as porosity, vary spatially. Several types of gradual changes can be envisaged, the gradual changes including patterns that can be classified as circular gradual changes, island gradual changes and step gradual changes. Many other gradual change patterns can be implemented depending on the type of actuation. The most appropriate pad gradual change for a specific operation is developed using a simulation method based on parameters including wafer sweep during CMP and its residual time distribution, friction coefficient and other physical properties Is done.

(Detailed explanation)
(Polishing pad with functional gradual change)
Described below is a polishing pad suitable for use in a CMP procedure for polishing the silicon-containing wafer described above. Those pads having at least two areas on the polishing surface adjacent to the silicon wafer to be polished have different material properties. The areas may be separate and may have these well defined clear boundaries. These areas can be of a type whose material properties vary with distance. Our polishing pads can include one or more polymeric materials, each of which can include one or more of the following polymeric materials: neat polymers having a specific molecular weight or molecular weight distribution, one Mixtures or alloys of the above polymers, two or more copolymers, and two or more block copolymers. The polymer composition may be filled with other polymeric or non-polymeric materials (eg, polymer fibers, natural fibers, particulate materials (eg, discrete “crumbs” of polymer material).

  The surface of our described polishing pad adjacent to the wafer surface to be polished can be substantially flat after completion of the synthesis process described herein. The surface having the above characteristics of the present invention was grooved and hilled.

  Suitable materials for the above CMD polishing pads include a wide variety of polymers (eg, a variety of polymers such as: polyurethane, polyurea, polycarbonate, nylon, various other polymers, polysulfone, various polyacetals). Etc.). These polymeric materials and their chemically related likes can be used for the manufacture of CMP pads. Other polymer chemistries can of course be used. Formulations using these materials necessarily entail some understanding of the relationship between the structure of the polymeric material and the resulting physical properties. Processing characteristics of this various components and composites (eg, interpolymer compatibility, reactivity, and viscosity between various areas).

  One polymer system with significant scientific, engineering and commercial history is the polyurethane and polyurea chemical system. These polymer products often contain isocyanates, polyols, polyamines, chain extenders and the like. Commercially, more than 90% of the isocyanates are toluene-diisocyanate (TDI) or diphenylmethane-diisocyanate (MDI) and their derivatives. Others include polymethylene polyphenyl isocyanate (PAPI). Isocyanate functional groups are important because they provide cross-linking and thus hardness and other pad compliance. The size and molecular weight of the polyamine / polyol reactant contributes to the flexibility, low temperature properties, hydrophilicity, light stability and processing properties of the resulting polymer.

  Chain extenders are often low molecular weight diamines or diols that are used to increase the urea / urethane content of the final polymer. These chain extenders react with the isocyanate and become part of the “hard segment” of the resulting polymer and often have a significant impact on hardness and elastic compliance. Many available chain extenders also modify processing characteristics such as increased gel time and viscosity. The strength of the final product is also often affected. Crosslinkers are characterized by molecular weight and functionality. Low molecular weight molecules, on a molar basis, are effective in crosslinking the polymer matrix and provide increased resistance to swelling, low temperature flexibility, and processing kinetics.

  There are two well known approaches for formulating polyurea / polyurethanes. These approaches are known as: (i) “one-shot” technology, and (ii) “two-shot” technology. FIG. 2 shows an outline of the one-shot technique. In summary, the components (eg, long chain diol, diisocyanate, and chain extender, if necessary) are mixed and reacted together. However, such a process is difficult to control. Local concentrations of reactants in this reaction mixture and random thermal control sometimes results in widely varying polymer product characteristics.

  FIG. 3 shows a schematic representation of the two-shot technique. The diisocyanate is pre-reacted with a long chain diamine / diol in the first step to form a high molecular weight isocyanate (typically known as a “prepolymer”). This functionalized polymer is then further reacted with a diamine / diol curable material or chain extender to complete the formation of the polyurea / polyurethane. This process is more easily controlled but requires a higher processing temperature (often near 100 ° C.).

  Our polishing pads can contain a neat polymer (eg, the polyurethane material or polyurea material discussed immediately above) or contain two or more different molecular weight product polyurethane or polyurea materials. It should be clear that it can be contained in different regions on the polished surface.

  This description of two methods for producing polyurethane or polyurea materials is both an example for producing these polymers and as an example for placing other polymers in the synthesis of incrementally modified pads. Should be considered. Such examples should not be considered as limitations on the scope of the present disclosure.

(Multi-stage injection molding)
Another variation of the process for making functionally graded pads includes a process known as multi-point injection molding (also referred to as co-injection sandwich molding). Multi-stage injection molding is a continuous process in which two or more polymeric materials are utilized, each of which is injected into a mold at a different time. This process is different from the in situ multi-stage injection molding process described below.

  This is one example of a procedure for synthesizing a gradually changing pad of two areas where two different polymer compositions (one for each area) are used. As shown in FIG. 4, the first outer annular ring of this pad is molded using an injection molding process. The completed outer ring is then placed in a second mold and the center of the pad ring is then filled with a second polymeric material. Once again, in the first molding step and the second molding step, two different materials are used so that the resulting pad has two different regions or areas with different (mechanical and physical) properties. Is done. Proper bonding at the interface between these two materials may require the selection of materials that are compatible with each other. Such information is readily available in published literature. The interface compatibility between the two polymers is usually good.

  This method can be used to form a gradual changed pad having more than two areas and more than two stages. Furthermore, this method can be used to achieve a gradual change pattern of any stage, from the simplest, best defined circular pattern to the most complex and random pattern.

  Figures 5 and 6 show more complex patterns that can be made using this process. The cruciform shown in FIG. 5 (as region 2) is surrounded by a cruciform object or partition, so that no polymer enters the other space. The cross-shaped mold or object is removed from the partially molded pad after one or other of these areas is filled.

  FIG. 6 shows an irregular set of patterns defining various selected areas (ellipses (202, 203, 204) and flags (205)) on the CMD polishing pad (200). In each of the areas described, the respective polymer can each be a different polymer type discussed above, or can be at least two different polymers. Again, such a pattern can be achieved using an appropriate molding geometry.

(Multistage live feed (or in situ) injection molding)
A mold with multiple in-situ injection ports can be used to create a progressively altered pad. In this method, a mold of at least two ports (generally independent) is selected for polymer injection. At least two different polymers are injected through these ports during the same injection process, often simultaneously, filling the mold. Depending on the gradual change requirements required for this pad, using normal polymer engineering calculations, to select the appropriate injection point and injection flow rate for different polymers fed into the mold The flow rate and heat transfer required for the calculation can be calculated.

  FIG. 7 shows a continuously gradual change made using this process by injecting a first polymer from the outer perimeter of the mold (212) and simultaneously injecting a second polymer material from the center. A polymer pad (210) is shown.

(Block copolymer)
Block copolymer systems can be used to produce incrementally modified pads. FIG. 8 shows a phase diagram expressing the relationship between the block copolymer composition of the final product (as a component of relative%) and the crystal structure (BCC, HCP, etc.) in a functional manner. In order to achieve a step change in properties, the composition of this block copolymer is stepwise. Thus, in a controlled manner, geometrical differences are achieved in this block copolymer by spatial dispersion. Since the geometric shape and% of (A or B) vary spatially and due to the fact that A and B are different units, this dispersion is a physical property (eg hardness, Provides a step change in modulus, porosity (by A or B solvent removal), roughness, and roughness.

  Specifically, a gradual change in hardness can be achieved by changing the molecular unit concentration of the block copolymer over distance. This is because the two molecular units A and B used to make this pad are of different hardness values. Furthermore, the change in crystal structure that results from the compositional relationship shown in FIG. 9 results in further controllable variations in physical parameters.

  Suitable molecular units for these procedures, including block copolymers, include materials such as styrene, isoprene, butadiene, urethane / urea, long chain diols and diamines.

(Injection molding assisted by gas)
A method for producing a gradually modified pad with microporosity includes a gas during the injection molding process to achieve a functional gradual change in porosity in a polymer polishing pad. The gas can be dispersed into the mold and injected into the mold from different ports at different flow rates in order to achieve gradual changes. The resulting pad contains different amounts of mixed gas at different points and achieves a difference in hardness or density.

(Reaction injection molding (RIM))
Certain polymer systems (eg polyurethane) are compatible with molding processes using RIM technology. In this molding process, instead of injecting a pre-synthesized polymer, the component monomer materials and a suitable cross-linking agent (eg glycerol), as well as an initiator and chain extender are added, and the resulting mixture is molded into During the polymerization. To create a progressively changed pad, multiple ports are used to inject two or more types of monomer units (and corresponding chain extenders) to achieve a gradual change in chemical structure . A step change in chemical structure results in a step change in mechanical and physical properties.

  By differentially adding monomer to the template, this method can be used to produce pads that are incrementally graded and continuous as discussed above.

(Lamella injection molding)
By using a mixture of pre-extruded (possibly layered) polymers, a polishing pad with progressively altered characteristics can be produced in an injection molding procedure as discussed above. This mode of making a simple physical mixture of polymers is straightforward and easily applied to change the demands on the manufacturer. The step change in properties obtained follows the mechanical and physical characteristics of the individual polymer.

  This method can result in a step change of the microdomain.

(Micro foaming (microcell molding))
In this technique, the polymer fluid to be molded is mixed with a gas for the purpose of forming a solution mixture. Utilizing two or more such solutions with different chemical properties results in a gradual change in both mechanical properties and porosity.

  The purpose of using incrementally modified pads for the CMP process is to allow different regions of the wafer to be polished to be exposed to selected different regions of the polymer pad in a selected manner. It is to be. For example, it may be advantageous for the outer periphery of the wafer to be exposed to a softer area of the pad, where the frictional shear force is higher than the harder area of the pad. In contrast, it may be advantageous for the inner region of the wafer to be exposed to a harder region of the pad to achieve a higher planarization length. During any CMP process using these incrementally modified polymer pads, both the wafer and pad are moving (rotation, vibration), so when combined with the dwell time distribution model, local planarization and overall At the same time. The material removal rate, defectivity, erosion, and wrinkle deformation of this wafer, as well as the effective planarization length of the CMP process, can determine the local tribology [hardness, compliance] and physical properties of the pad material [pore size and density, roughness ]. Our functionally graded polishing pads allow for the selection of local physical pad parameters and global physical pad parameters that allow these CMP shortcomings to be improved. To do.

FIG. 1 shows an exemplary general Stribeck curve, which shows the degree of contact between rotating wafers, rotating pads, and contained abrasive particles in various lubrication regimes. FIG. 2 shows a schematic one-shot one-shot technique for producing a progressively modified polishing pad in which isocyanates, polyamines / polyols, chain extenders, and other additives are added. All blended together. FIG. 3 shows a schematic prepolymer or “two-shot” technique for producing a progressively modified polishing pad. FIG. 4 shows a schematic step gradual change pad having an outer ring of one formulation and an inner region of a second formulation having different tribological properties. FIG. 5 shows a gradual change pad with an island of one polymer formulation and a peripheral region of the polymer matrix with different tribology. FIG. 6 shows a gradual change pad having a gradual change composite with materials having varying shapes, sizes, and tribological properties. FIG. 7 shows a continuous gradual change pad, in which the center contains the first formulation and the edge contains the second formulation. In this variation, the two formulations are completely miscible. FIG. 8 shows a gradual change pad where the gradual change is a microdomain morphological organization using injection of block copolymer or grafted copolymer. FIG. 9 schematically shows the phase diagram (temperature-composition) of block copolymers having different equilibrium structures (cubes (bcc), hexagons (hcp) and lamelae).

Claims (7)

A functionally graded polishing pad for a CMP process, the polishing pad comprising a polishing pad having a polishing surface, the polishing surface for polishing a silicon-containing wafer; is configured for use in the CMP steps, the polishing surface is one piece, a substantially flat, includes the two areas, each of the two areas, first, the long chain diamine isocyanate A polyurethane or polyurea of different molecular weight formed by reacting with a diol / diol to form an isocyanate polymer and then reacting the isocyanate polymer with a diamine / diol curable material or chain extender, the two areas The change in composition during the process is a continuous change, a functionally incremental polishing pad. The functionally graded polishing pad of claim 1, wherein a region between the two areas comprises a mixture of polymeric materials having different compositions. Wherein at least one of the two areas, which is a cyclic, functionally polishing pads gradations of claim 1. Wherein at least one of the two areas one, but an island-shaped, functionally polishing pads gradations of claim 1. The pad is circular with an edge and a center, the edge comprises one area, the center comprises another area, and the composition change between the edge and the center is: The functionally incremental polishing pad of claim 1, wherein the polishing pad is the continuous change. The functionally graded polishing pad of claim 1 manufactured by a continuous injection molding process of at least two polymer compositions so as to form the two areas. The functionally graded polishing pad of claim 1 manufactured by a process of simultaneous injection molding of at least two polymer compositions so as to form the two areas.

Images