High planarization efficiency chemical mechanical polishing pads and methods of making

Abstract

A chemical mechanical polishing pad for polishing a semiconductor substrate is provided containing a polishing layer that comprises a polyurethane reaction product of a reaction mixture comprising a curative and a polyisocyanate prepolymer having an unreacted isocyanate (NCO) concentration of from 8.3 to 9.8 wt. % and formed from a polyol blend of polypropylene glycol (PPG) and polytetramethylene ether glycol (PTMEG) and containing a hydrophilic portion of polyethylene glycol or ethylene oxide repeat units, a toluene diisocyanate, and one or more isocyanate extenders, wherein the polyurethane reaction product exhibits a wet Shore D hardness of from 10 to 20% less than the Shore D hardness of the dry polyurethane reaction product.

Claims

1. A chemical mechanical (CMP) polishing pad for polishing a substrate chosen from at least one of a magnetic substrate, an optical substrate and a semiconductor substrate comprise a polishing layer adapted for polishing the substrate which is a polyurethane reaction product of a reaction mixture comprising a curative and a polyisocyanate prepolymer having an unreacted isocyanate (NCO) concentration of from 8.3 to 9.8 wt. % of the polyisocyanate prepolymer, the polyisocyanate prepolymer formed from a polyol blend of polypropylene glycol (PPG) and polytetramethylene ether glycol (PTMEG) and containing a hydrophilic portion of polyethylene glycol or ethylene oxide repeat units, a toluene diisocyanate, and one or more isocyanate extenders, and wherein the polyurethane reaction product in the polishing pad has a Shore D hardness according to ASTM D2240-15 (2015) of from 65 to 80 and exhibits a wet Shore D hardness according to ASTM D2240-15 (2015) after soaking in deionized (DI) water for a period of 7 days of from 10 to 20% less than the Shore D hardness of the dry polyurethane reaction product.

2. The CMP polishing pad as claimed in claim 1, wherein the polyisocyanate prepolymer has an unreacted isocyanate (NCO) concentration of from 8.6 to 9.3 wt. %.

3. The CMP polishing pad as claimed in claim 1, wherein the amount of toluene diisocyanate (TDI) used to form the polyisocyanate prepolymer ranges from more than 35 wt. % to 45 wt. %, based on the total wt. % of the reactants used to make the polyisocyanate prepolymer, wherein, further, the amount of the one or more isocyanate extenders used to form the polyisocyanate prepolymer ranges from 3 to 11 wt. %, based on the total weight of the reactants used to make the polyisocyanate prepolymer, and wherein, still further, the amount of the polyol blend used to form the polyisocyanate prepolymer ranges from 44 to less than 62 wt. %, based on the total wt. % of the reactants used to make the polyisocyanate prepolymer.

4. The CMP polishing pad as claimed in claim 1, wherein the polyol blend used to form the polyisocyanate prepolymer contains a hydrophilic portion and is chosen from (i) a polyol blend of PTMEG and PPG in a ratio of PTMEG to PPG of from 1:1.5 to 1:2 and a hydrophilic portion in the amount of from 20 to 30 wt. %, based on the total weight of reactants used to make the polyisocyanate prepolymer or (ii) a polyol blend of PTMEG and PPG in a ratio of PTMEG to PPG of from 9:1 to 12:1 wt. ratio and a hydrophilic portion in the amount of from 1 to 10 wt. %, based on the total weight of reactants used to make the polyisocyanate prepolymer.

5. The CMP polishing pad as claimed in claim 1, wherein the polyurethane reaction product is formed from a reaction mixture containing from 70 to 81 wt. %, based on the total weight of the reaction mixture, of the polyisocyanate prepolymer, from 19 to 27.5 wt. %, based on the total weight of the reaction mixture, of the curative and from 0 to 2.5 wt. %, of one or more microelements, based on the total weight of the reaction mixture.

6. The CMP polishing pad as claimed in claim 1, wherein the curative in the reaction mixture is chosen from a diamine or a mixture of a diamine and a polyol curative and the molar ratio of polyamine NH.sub.2 groups to polyol OH groups ranges from 40:1 to 1:0.

7. The CMP polishing pad as claimed in claim 6, wherein the stoichiometric ratio of the sum of the total moles of amine (NH.sub.2) groups and the total moles of hydroxyl (OH) groups in the curative in the reaction mixture to the total moles of unreacted isocyanate (NCO) groups in the reaction mixture ranges from 0.91:1 to 1.15:1.

8. The CMP polishing pad as claimed in claim 1, wherein the polishing pad or polishing layer has a density of 0.93 to 1.1 g/cm.sup.3.

9. The CMP polishing pad as claimed in claim 1, wherein the polishing pad further comprises microelements chosen from entrapped gas bubbles, hollow core polymeric materials, liquid filled hollow core polymeric materials, and boron nitride.

10. A method for making a chemical mechanical (CMP) polishing pad having a polishing layer adapted for polishing a substrate comprising: providing one or more polyisocyanate prepolymer as claimed in claim 1 at a temperature of from 45 to 65 C.; forming a reaction mixture containing from 70 to 81 wt. %, based on the total weight of the reaction mixture, of the polyisocyanate prepolymer, from 0.0 to 2.5 wt. %, based on the total weight of the reaction mixture, of one or more microelements, wherein the microelements and the polyisocyanate prepolymer are blended together, cooling the polyisocyanate prepolymer and microelement mixture to from 20 to 40 C.; providing, as a separate component, from 19 to 27.5 wt. %, based on the total weight of the reaction mixture, of a curative; combining the components of the reaction mixture, preheating a mold to from 60 to 100 C.; filling the mold with the reaction mixture and heat curing the reaction mixture at a temperature of from 80 to 120 C. for a period of from 4 to 24 hours to form a cast polyurethane; and, forming a polishing layer from the cast polyurethane.

Description

EXAMPLES

(1) The present invention will now be described in detail in the following, non-limiting Examples:

(2) Unless otherwise stated all temperatures are room temperature (21-23 C.) and all pressures are atmospheric pressure (760 mm Hg or 101 kPa).

(3) Notwithstanding other raw materials disclosed below, the following raw materials were used in the Examples:

(4) V5055HH: Multifunctional polyol (OH Eq. wt 1900), also sold as Voralux HF505 high molecular weight polyol curative having a number average molecular weight, M.sub.N, of 11,400 (The Dow Chemical Company, Midland, Mich. (Dow)).

(5) Expancel 551 DE 40 d42 beads: Fluid filled polymeric microspheres with nominal diameter of 40 m and true density of 42 g/I (Akzo Nobel, Arnhem, NL); and,

(6) Expancel 461DE 20 d70 beads: Fluid filled polymeric microspheres with nominal diameter of 20 m and true density of 70 g/I (Akzo Nobel).

(7) The following abbreviations appear in the Examples:

(8) PO: Propylene oxide/glycol; EO: Ethylene oxide/glycol; PTMEG: Poly(THF) or polytetramethylene glycol; TDI: Toluene diisocyanate (80% 2,4 isomer, 20% 2,6 isomer); BDO: Butanediol (1,3 or 1,4 regioisomers); DEG: Diethylene glycol; MbOCA: 4,4-Methylenebis(2-chloroaniline).

(9) TABLE-US-00001 TABLE 1 Polyisocyanate Prepolymers PO EO PTMEG TDI BDO DEG NCO Molecular Prepolymer Backbone (wt. %) (wt. %) (wt. %) (wt. %) (wt. %) (wt. %) (wt %) weights A PTMEG 0 0 58 38 0 3 ~9.0 Mn 900; Mw 1,350 B PPG 26 20 0 41 0 12 ~9.0 Mn 650; Mw 1,300 C PPG 54 15 0 24 5 2 ~5.8 Mn 900; Mw 2,320

(10) NMR Spectroscopy:

(11) Was performed on homogeneous solutions of 3 g sample and 1.2 mL of a 0.025M chromium (III) acetoacetate Cr(AcAc).sub.3 solution of Acetone-d.sub.6 in 10 mm NMR tubes (Cr(AcAc).sub.3 was added as a relaxation agent for quantitative .sup.13C NMR spectra). .sup.13C NMR experiments were carried out at room temperature on an AVANCE 400 spectrometer equipped with a 10 mm broadband observe (BBO) probe head (Bruker Instruments, Billerica, Mass.). Table 2, below, provides peak assignments which were integrated to give the contents of the indicated species.

(12) TABLE-US-00002 TABLE 2 13C NMR Spectra and Peak Assignments for Polyurethane Prepolymers 13C NMR Peaks (in ppm).sup.1 Assigned 68.2, 68.5, 70.2, EO 70.5 15.5, 17.0, 18.1, PO 72.4, 72.9, 74.6, 74.8 PO 63.8, 69.3, 69.4 DEG 20.0, 35.9, 60.5, BDO 68, 2 26-28, 64-65, 69- PTMEG 70, 69-72.5 11.9, 15.5, 16.6, TDI (2,4 and 2,6 109.1, 109.9, 110- regioisomers) 142, 151.1, 152.3 .sup.1Peak locations vary; so, all observed peak assignments from several samples are reported and ranges are given to indicate areas in which multiple peaks are clustered.

(13) As shown in Table 3, below, formulations of various reaction mixtures were cast into polytetrafluorethylene (PTFE-coated) circular molds 86.36 cm (34) in diameter having a flat bottom to make moldings for use in making polishing pads or polishing layers. To form the formulations, the indicated polyisocyanate prepolymer heated to 52 C. to insure adequate flow and having in it the indicated microelements, as one component, and the curative, as another component were mixed together using a high shear mix head. After exiting the mix head, the formulation was dispensed over a period of 2 to 5 minutes into the mold to give a total pour thickness of 7 to 10 cm and was allowed to gel for 15 minutes before placing the mold in a curing oven. The mold was then cured in the curing oven using the following cycle: 30 minutes ramp from ambient temperature to a set point of 104 C., then hold for 15.5 hours at 104 C., and then 2 hour ramp from 104 C. to 21 C.

(14) To cast the reaction mixture formulations as cakes with a high post-skiving yield, the inventive Examples 2, 6, and 10 were cast using a prepolymer line heat exchanger to reduce the prepolymer casting temperature to the indicated temperature from 52 C. to 27 C. (80 F.), and the molds were preheated to 93 C., this enables control of the high exotherm to mitigate variation within the mold. In the Comparative Examples 1, 3-5 and 7-9, as indicated in Table 4, below, cooling of the reaction mixture or mold preheating was varied. The reaction mixture was cooled in Comparative Example 1 because of its highly reactive reaction mixture. Porosity is proportional to microsphere loading and inversely proportional to SG; porosity was limited in the inventive Examples 2, 6 and 10 because the high exotherm would otherwise have led to uneven or uncontrolled microsphere expansion during molding.

(15) TABLE-US-00003 TABLE 3 Example Formulations Prepolymer Blend Prepolymer Approx. Pore Blend Ratio wt. % Wt. % Pore Level Size Example 1:2 1:2 NCO.sup.1 Curative Stoich.sup.2 (Wt. %) SG (m) 0* L325.sup.3 9.05-9.25 20.1 0.87 1.7 0.80 40 1* A n/a 8.75-9.03 22.9 1.05 1.4 0.96 20 2 A:C 9:1 ~8.6 22.3 1.05 0.8 1.04 20 3* A:C 4:1 8.03-8.36 18.7 0.89 1.1 1.00 20 4* A:C 1:1 7.12-7.41 18.3 0.97 2.7 0.82 20 5* C n/a ~5.7 14.0 0.90 1.2 0.91 40 6 A:B 1:4 8.67-9.05 22.8 1.05 1.1 1.02 20 7* A:B 1:1 8.70-9.04 21.1 0.95 5.4 0.64 20 8* B n/a 8.65-9.05 20.4 0.91 0.8 1.07 20 9* B n/a 8.67-9.05 20.4 0.91 0.4 1.07 40 10 A:B 1:4 8.65-9.05 26.1 1.05 1.5 0.97 20 *Denotes Comparative Example; .sup.1Unreacted free NCO content; .sup.2Stoichiometry refers to a ratio of (OH + NH.sub.2 groups) to free NCO groups; .sup.3IC1000 pad (Dow) made using ADIPRENE L325 prepolymer (Chemtura).

(16) In Examples 0 to 9 above, the polyamine curative was MbOCA and in Example 10 it was MbOCA+V5055HH polyol (5 wt. % of the total reaction mixture).

(17) TABLE-US-00004 TABLE 4 Casting Parameters E Elbow Temp Mold Example ( C.) Temp Porosity 0* 52 RT 0.30 1* 27 RT 0.19 2 27 93 C. 0.12 3* 46 RT 0.15 4* 52 RT 0.29 5* 44 RT 0.22 6 27 93 C. 0.15 7* 52 RT 0.47 8* 52 93 C. 0.11 9* 52 93 C. 0.11 10 27 93 C. 0.19 *Denotes Comparative Example.

(18) The cured polyurethane cakes were then removed from the mold and skived (cut using a stationary blade) at a temperature of from 70 to 90 C. into approximately thirty separate 2.0 mm (80 mil) thick sheets. Skiving was initiated from the top of each cake. Any incomplete sheets were discarded.

(19) The ungrooved, polishing layer materials from each example were analyzed to determine their physical properties. Note that the pad density data reported were determined according to ASTM D1622-08 (2008); the Shore D hardness data reported were determined according to ASTM D2240-15 (2015); and, the modulus and elongation to break data reported were determined according to ASTM D412-6a (2006). Test results are shown in Tables 5, 6 and 7, below.

(20) As determined by the proportion or amount of useful pad materials made from a single cast polyurethane cakes compared to the total amount of the cake, the resulting inventive polishing pads in Examples 2, 6 and 10 gave high casting yield for polishing pads. For example, relative to Comparative Example 7, the casting conditions for Examples 6 and 10 produce a higher casting yield while offering slightly improved polishing performance without the porosity of the pad in Comparative Example 7.

(21) Test Methods:

(22) The following methods were used to test the polishing pads: Chemical mechanical polishing pads were constructed using polishing layers. These polishing layers were then machine grooved to provide a groove pattern in the polishing surface comprising a plurality of concentric circular grooves having dimensions of 70 mil (1.78 mm) pitch, 20 mil (0.51 mm width) and 30 mil (0.76 mm) depth. The polishing layers were then laminated to a foam sub-pad layer (SUBA IV available from Rohm and Haas Electronic Materials CMP Inc.). The resulting pads were mounted to the polishing platen of the indicated polisher using a double sided pressure sensitive adhesive film.

(23) A Mirra CMP polishing platform (Applied Materials, Santa Clara, Calif.) was used to polish 200 mm diameter TEOS (oxide) blanket wafers (Novellus Systems, Tualatin, Oreg.) with the indicated pads. The indicated polishing medium used in the polishing experiments was a CES333F (Asahi Glass Company) ceria slurry, KLEBOSOL II K1730 (Rohm and Haas Electronic Materials CMP Inc.) colloidal silica slurry, or ILD 3225 (Nitta Naas Inc.) fumed silica slurry. The polishing conditions used in all of the polishing experiments included a platen speed of 93 rpm; a carrier speed of 87 rpm; with a polishing medium flow rate of 200 mL/min and a down force of 31.0 kPa (KLEBOSOL and ILD slurries) or 20.7 kPa (CES333F slurry). An AM02BSL8031C1-PM (AK45) diamond conditioning disk (Saesol Diamond Ind. Co., Ltd.) was used to condition the chemical mechanical polishing pads. The chemical mechanical polishing pads were each broken in with the conditioner using a down force of 3.2 kg (7 lbs) for 40 minutes. The polishing pads were further conditioned in situ using a down force of 3.2 kg (7 lbs). The removal rates were determined by measuring the film thickness before and after polishing using a FX200 metrology tool (KLA-Tencor, Milpitas, Calif.) using a 49 point spiral scan with a 3 mm edge exclusion.

(24) Planarization Efficiency (PE):

(25) To assess the ability of an indicated pad to remove material in the step height reduction from a non-level and non-uniform substrate, a substrate pattern wafer with a step height of 8000 (CMP Characterization Mask Set, MIT-SKW7) was formed by chemical vapor deposition of TEOS in a lined pattern that includes rectangular sections of varying pitches (from 10 to 500 m at 50% pattern density) and pattern densities (from 0% to 100% at a 100 m line pitch). Planarization efficiency ratio was evaluated by optical interference using a RE-3200 Ellipsometric Film Thickness Measurement System (Screen Holdings Co). Planarization efficiency is defined as 1-RR.sub.low/RR.sub.high. The planarization efficiency ratio was calculated by integrating under the curve of planarization efficiency vs. step height and dividing the result by the initial step height. Results are shown in Tables 5, 6 and 7, below.

(26) PE (Norm):

(27) In Table 7, this refers to planarization efficiency relative to Example 0 as a standard.

(28) Defectivity:

(29) The creation of defects during polishing was measured using a Hitachi High-Tech LS6600 metrology tool (Hitachi High Technologies Corporation, Tokyo, Japan) wherein the substrate was cleaned with HF (2 wt. % in water) to an etching amount of 400 TEOS. Target remaining TEOS thickness was 6000 . Defect count was determined in a wafer substrate which is not a pattern wafer by an LS6600 wafer surface inspection system with 0.2 m resolution. Results are shown in Table 4, below.

(30) Subtractive defects are scratches and chatter marks (not additive defects) counted using the metrology tool and confirmed by manual inspection by SEM (KLA-Tencor eDR5210 Review SEM) and are normalized to a pad of Comparative Example 1) which is assigned a value of 1.0. A lower number means less defects in the substrate after polishing.

(31) Matrix Dry Hardness:

(32) The matrix hardness was determined by taking a lab-cast plaque of the indicated polyurethane reaction product. Six samples were stacked and shuffled for each hardness measurement; and each pad tested was conditioned by placing it in 50 percent relative humidity for five days at 23 C. before testing and using methodology outlined in ASTM D2240-15 (2015) to improve the repeatability of the hardness tests.

(33) Matrix Wet Hardness:

(34) The matrix wet hardness was determine by cutting to samples from a lab-cast plaque and subjecting it to the same ASTM hardness analysis as in Matrix Dry Hardness after soaking in DI water for a period of 7 days.

(35) TABLE-US-00005 TABLE 5 Planarization Efficiency and Defectivity with ILD3225 Fumed Silica Slurry.sup.1 Matrix Matrix Tan Subtractive Dry Wet Delta Defects Example Hardness Hardness (50C) PE (norm.) 0* 66.3 65.6 0.111 0.877 1* 72.3 67.4 0.160 0.915 1.0 2 73.2 64.7 0.176 0.908 0.2 3* 65.8 62.7 0.099 0.885 4* 64.5 61.6 0.125 0.817 0.1 5* 53.8 41.7 0.081 0.761 6 71.5 60.0 0.145 0.911 0.4 7* 68.4 63.5 0.883 8* 71.3 64.0 0.112 0.854 9* 71.3 64.0 0.112 0.894 10 66.6 57.6 0.133 0.895 0.2 .sup.1ILD3225 fumed silica slurry; *denotes comparative Example.

(36) TABLE-US-00006 TABLE 6 Planarization Efficiency and Defectivity with K1730 Colloidal Silica Slurry.sup.1 Subtractive Defects Example PE (norm.) 0* 0.773 1* 0.874 1.0 2 0.877 0.2 3* 0.840 4* 0.765 0.2 5* 0.592 6 0.896 0.4 7* 8* 9* 0.837 10 0.888 0.2 .sup.1K1730 colloidal silica slurry; *denotes comparative Example.

(37) TABLE-US-00007 TABLE 7 Planarization Efficiency and Defectivity with CES333 ceria slurry Subtractive PE Defects Example (norm.) (norm.) 0* Medium Medium 1* High Very High 2 High 3* 4* Medium Low 5* 6 Very High Low 7* 8* 9* 10 1. CES333 ceria slurry, mean particle size 170 nm; *denotes comparative Example.

(38) As shown in Tables 5, 6 and 7, above, the pads of inventive Examples 2 and 6 maintain similar PE as a high quality prior art planarizing pad (Comparative Example 1) while exhibiting significantly attenuated defectivity with ILD3225 (fumed silica), K1730 (colloidal silica), and CES333 (conventional ceria) slurries compared to the same pad. The inventive Examples 2, 6 and 10 all gave improved PE as compared to IC1000 commercial pads (Comparative Example 0).

(39) As shown in Tables 5, 6 and 7, above, the pads in inventive Examples 2, 6, and 10 offer similar, if not higher, planarization efficiency than a high quality prior art planarizing pad (Comparative Example 1) while exhibiting significantly decreased defectivity. This combination makes these formulations ideal for front-end-of-line polishing applications.

(40) As shown in Table 5 and in Tables 6 and 7, by correlation of the same pad materials used in all three tables, the performance of the inventive Examples 2, 6 and 10 relates to the drop from dry hardness of the materials to the wet hardness of the materials while in use, their high flexural rigidity (EI), and their high damping component in the relevant polishing method as shown by tan delta similar to the good planarizing pad of Comparative Example 1. The inventive pads exhibit a unique decrease in hardness between their dry and wet states. Further, the Shore D hardness of the pads in Examples 2, 6, and 10 drops significantly (>10%) when they are wet. By comparison, the pad of Comparative Example 1 maintains high dry and wet hardness leading to high subtractive defects in substrates.