AQUEOUS SILICA SLURRY COMPOSITIONS FOR USE IN SHALLOW TRENCH ISOLATION AND METHODS OF USING THEM

Abstract

The present invention provides aqueous CMP polishing compositions comprising a from 0.5 to 30 wt. %, based on the total weight of the composition of a dispersion of a plurality of elongated, bent or nodular colloidal silica particles which contain a cationic nitrogen atom, and from 0.001 to 0.5 wt. %, preferably from 10 to 500 ppm, of a cationic copolymer of a diallyldimethylammonium salt, such as a diallyldimethylammonium halide, wherein the compositions have a pH of from 1 to 4.5. Preferably, the cationic copolymer of a diallyldimethylammonium salt comprises a copolymer of diallyldimethylammonium chloride (DADMAC) and sulfur dioxide. The slurry compositions demonstrate good oxide selectivity in the CMP polishing of pattern wafers having nitride and silicon patterns.

Claims

1. An aqueous chemical mechanical planarization polishing composition comprising a dispersion of a plurality of elongated, bent or nodular colloidal silica particles which contain a cationic nitrogen atom, and from 10 to 500 ppm of a cationic copolymer of a diallyldimethylammonium salt comprising a copolymer of a diallyldimethylammonium halide salt and sulfur dioxide, wherein the composition has a pH of from 1 to 4.5 and, further wherein, the amount of the dispersion of the elongated, bent or nodular colloidal silica particles, ranges from 0.5 to 30 wt. %, all weights based on the total weight of the composition.

2. The aqueous chemical mechanical planarization polishing composition as claimed in claim 1, wherein the dispersion of elongated, bent or nodular colloidal silica particles have for the average particle an aspect ratio of longest dimension to the diameter which is perpendicular to the longest dimension from 1.8:1 to 3:1.

3. The aqueous chemical mechanical planarization polishing composition as claimed in claim 1, wherein the amount of the dispersion of elongated, bent or nodular colloidal silica particles ranges from 80 to 99.9 wt. %, based on the total solids weight of the colloidal silica particles in the composition.

4. (canceled)

5. The aqueous chemical mechanical planarization polishing composition as claimed in claim 1, wherein the cationic copolymer of a diallyldimethylammonium salt comprises a copolymer of diallyldimethylammonium chloride and sulfur dioxide.

6. The aqueous chemical mechanical planarization polishing composition as claimed in claim 1, wherein the cationic copolymer of a diallyldimethylammonium salt comprises a copolymer of 45 to 55 mole % of the diallyldimethylammonium halide salt and from 45 to 55 mole % of the sulfur dioxide.

7. The aqueous chemical mechanical planarization polishing composition as claimed in claim 5, wherein the cationic copolymer of a diallyldimethylammonium salt comprising a copolymer of diallyldimethylammonium halide and sulfur dioxide has a weight average molecular weight of from 1,000 to 15,000.

8. The aqueous chemical mechanical planarization polishing composition as claimed in claim 1, wherein the composition has a pH of from 2.5 to 4.3.

9. A method of using the aqueous chemical mechanical planarization polishing composition as claimed in claim 1, comprising: polishing a substrate with a chemical mechanical polishing pad and the aqueous chemical mechanical polishing composition.

10. The method as claimed in claim 9, wherein the substrate comprises both silicon dioxide and silicon nitrides, and the polishing results in an oxide:nitride removal rate ratio of from 3:1 to 25:1.

11. The aqueous chemical mechanical planarization polishing composition as claimed in claim 5, wherein the cationic copolymer of a diallyldimethylammonium salt comprises a copolymer of diallyldimethylammonium chloride and sulfur dioxide in amounts of 10-20 ppm.

Description

EXAMPLES

The Following Examples Illustrate the Various Features of the Present Invention

[0046] In the Examples that follow, unless otherwise indicated, conditions of temperature and pressure are ambient or room temperature and standard pressure.

[0047] The following materials, including those listed in Table A, below, were used in the Examples that follow:

TABLE-US-00001 TABLE A Silica and Other Abrasive Particles Aqueous Particle size Raw Concentration Silica Slurry Source pH.sup.5 (CPS, nm) Morphology Materials (wt. % solids) Slurry A Klebosol?.sup., 1 7.7 38 Spherical Na Silicate 30 1598-B25 Slurry B Klebosol?.sup., 1 2.5 75 Spherical Na Silicate 30 30H50i Slurry C HL-3?.sup., 3 7.8 55 Elongated, TMOS 20 cationic.sup.4 particle Slurry D BS-3?.sup., 3, 7.3 53 Elongated, TMOS 20 cationic.sup.4 particle Slurry E Ceria (see separate listing, below) Slurry F (see Slurry C, above, and separate listing, below) Slurry G (see Slurry A and B, above and separate listing, below) .sup.1 Merck KgAA, Lamotte, France; .sup.3 Fuso Chemical, Osaka, JP; .sup.4Charge determined at pH of 3.0 and cationic particles formed with TMOS and an amine containing alkaline catalyst, such as tetramethylammonium hydroxide; .sup.5pH as delivered from source. Diquat? additive: N,N,N,N,N,N-hexabutyl-1,4-butanediammonium dihydroxide, 98 wt. % (Sachem, Austin, TX); Slurry G: A 24 wt. % solids formulation at a pH of 2.39 and containing 20 wt. % of Slurry A solids, 4 wt. % of Slurry B solids, 0.2 wt. % Diquat? additive and 0.112 wt. % HNO.sub.3. At POU (6%, 4x dilution), the pH was ~pH 3. Copolymer 1 is a 1:1 copolymer of DADMAC and sulfur dioxide, having a weight average molecular weight (MW) (GPC using polyethylene glycol standards) of 5,000 as reported by manufacturer (PAS-A-1, Nitto Boseke Co. Ltd, Fukushima, JP); Polymer 2 is a homopolymer of DADMAC having a weight average molecular weight of (MW, GPC) of 8,500 (Nitto Boseke Co.). Slurry E: Ceria slurry, pH 5.2, polyacrylic acid dispersant, 0.75 wt. % ceria solids undiluted, 1:3 dilution as used. Slurry F: A 2 wt. % solids composition of Slurry C/67 ppm poly(acrylic acid) (PAA), of weight average MW 1800/citric acid/pH 3.3; and, Slurry C is positively charged below pH 4.5. The various colloidal silica particles used in the Examples are listed in Table A, above. The following abbrevations were used in the Examples that follow: POU: Point of use; RR: Removal rate;

[0048] The following test methods were used in the Examples that follow:

[0049] pH at POU: The pH at point of use (pH at POU) was that measured during removal rate testing after dilution of the indicated concentrate compositions with water to the indicated solids content.

[0050] Post CMP (SP2xp) Defect Counts: Four TEOS wafers were used as defect monitor wafers for each slurry. Each defect wafer was polished for 60 s at 3 psi, 93/87rpm and 150 ml/min slurry flow rate. After polish, wafers were scanned on a Surfscan? SP2xp metrology tool (KLA-Tencor, Milpitas, Calif.) to obtain Post CMP defect wafer maps, followed by automatic SEM review of 100 random defects. Klarity Defect software (KLA-Tencor, Milpitas, Calif.) with a wide open channel setting (i.e. no defect size limit) was used to extract Post CMP total defect counts for each wafer. Defect counts should be as low as possible.

[0051] Post HF (hydrofluoric acid) Defect Counts: After post CMP defect analysis, wafers were exposed to a 1.92 wt. % HF solution for a time sufficient to remove 200 ? of a given substrate, using an M3307-2949 Veeco? HF cleaner (Veeco, Horsham, Pa.). Wafers were re-scanned on the Surfscan SP2xp (KLA-Tencor) to get Post HF defect wafer maps, followed by automatic SEM review of 100 random defects.

[0052] Removal Rate: In a removal rate test, a Mirra? (200 mm) polishing machine or Mirra RR (Applied Materials, Santa Clara, Calif.) polishing device with an IC1010? or other indicated CMP polishing pad (The Dow Chemical Company, Midland, Mich. (Dow)) was used to polish an STI pattern wafer substrate having a specified feature % (which corresponds to the area of active or high areas in the wafer relative to the total area thereof) with an MIT mask (SKW-3 wafers, SKW, Inc. Santa Clara, Calif.) using the CMP polishing compositions defined in Table 1, below, at a 20.7 kPa (3 psi) down-force, slurry flow rate of 150 mL/min, a 93 rpm platen speed and an 87 rpm carrier speed. During polishing, the pad was conditioned with a Kinik? AD3CS-211250-1FN conditioning disk (Kinik Company, Taiwan) at a 3.17 kg (7 lbf) pressure, using 100% in situ conditioning.

[0053] Multi-step CMP polishingP1 (first step) and P2 (subsequent steps): CMP polishing was conducted such that, in the first step or P1 process, the overburden high density plasma oxide (HDP) film was removed. The film was polished using a VP6000? polyurethane CMP polishing pad (Dow, Shore D (2 second) hardness: 53) and Slurry E and by applying a polishing down-force of 20.7 kPa (3 psi) and platen speed of 93 rpm. P1 polishing was stopped when complete planarization was achieved on the 50% pattern density (PD) feature on the middle die of the wafer. At this point, ?500 ? of HDP film remained on the 50% feature. On the smaller features, such as the 10% and 20% PD features, however, the HDP film was completely removed and the underlying nitride film was exposed. Features with >50% PD still had significant dielectric film over the nitride film. Before moving to P2, the patterned wafer was cleaned using SP100 cleaning chemistry (TMAH containing) on a OnTrak DSS-200 Synergy? tool (Lam Research, Fremont, Calif.) to remove ceria particles from the wafer. P2 polishing was performed using an IC? polyurethane polishing pad (Dow, Shore D (2 second) hardness: 70) with 1010? groove design (Dow) and the indicated slurry composition, using a polishing down-force of 20.7 kPa (3 psi) and a platen speed of 93 rpm. For the 50% pattern density feature, the polishing endpoint was defined as the time at which the HDP was cleared and the nitride film was exposed. Trench oxide loss was monitored on the 50% pattern density feature for each step-polishing event. The HDP oxide removal on the 100% pattern density feature was also measured. Overpolish is defined as the amount of HDP film removed on the 100% feature after silicon nitride was exposed on the 50% pattern density feature. Selectivity was calculated as the ratio of silicon nitride removal rate to the ratio of HDP oxide removal rate on the 100% feature. All dielectric film thicknesses and removal rates were determined by measuring the film thickness before and after polishing using a KLA-Tencor? FX200 metrology tool (KLA Tencor, Milpitas, Calif.) using a 49 point spiral scan with a 3 mm edge exclusion. Further polishing details are set forth in Table B, below.

TABLE-US-00002 TABLE B Polishing Parameters Pads P1: VP6000 2 mm (0.080) SIV 508 mm (20); D18AR; SG 0.8 P2: IC1010 2 mm (0.080) SIV 508 mm (20); 1010; SG 0.8 Slurry P1: ceria Slurry E (1:3) P2: Silica STI formulations Polishing 20.7 kPa (3 psi), 93/87 rpm, 150 mL/min Process Polishing Tool Applied Materials Mirra? Thin Film KLA-Tencor? FX200, 49 point spiral scan w/3 mm Metrology edge exclusion Break-in Recipe P1: 3.17 kg (7 lbf) for 40 min P2: 3.17 kg (7 lbf) for 40 min Conditioning P1: 100% in situ at 3.17 kg (7 lbf) P2: 100% in situ at 3.17 kg (7 lbf) Slurry Drop Point ~9.53 cm (~3.75) from pad center

[0054] Polishing was continued for the indicated time intervals or to the extent of the indicated overpolish amount. In each of Tables 3, 4, and 5, below, Performance Criterion A is trench oxide loss (?): Acceptable trench oxide loss is less than 250 ? at a 500 ? overpolish amount, preferably, less than 215 ? at 500 ? overpolish amount; Performance Criterion B is SiN loss (?): Acceptable SiN loss is less than 200 ? at a 500 ? overpolish amount, preferably, less than 150 ? at a 500 ? overpolish amount; and Performance Criterion A is dishing (?): Acceptable dishing is less than 200 ? at 500 ? overpolish amount, preferably, less than 175 ? at 500 ? overpolish amount.

[0055] Where otherwise indicated, the polished substrate was a recycled tetraethoxylsilicate (TEOS) wafer (TENR) used for blanket wafer studies.

TABLE-US-00003 TABLE 1 Slurry Formulation Details Slurry/ Amount pH (wt. % Copolymer 1 (adjusted with Slurry solids) (ppm) HNO.sub.3) 1 C/1 10 3.3 2 C/3 20 3.3 3 D/4.5 20 3.3 4* (Slurry F) C/2 (none) 67 3.3 ppm PAA 5 C/3 (with 20 3.3 75 ppm Diquat) 6* C/3 20 of Polymer 3.3 2 *Denotes Comparative Example.

Example 1

Defect Counts

[0056] In Table 2, below, the substrate was an oxide wafer from tetraethyl orthosilicate (TEOS). Polishing was conducted for 60 seconds using the indicated slurry.

TABLE-US-00004 TABLE 2 Defect Counts Example 1A 1B 1C Slurry 2 4* Slurry E* Post CMP Wafer 1 241 245 22831 SP2xp Defect Wafer 2 119 496 23637 Counts Wafer 3 207 124 26888 Wafer 4 139 136 21826 Average 177 250 23796 Std. 57 173 2191 Dev. Post HF Defect Wafer 1 1354 966 911 Counts Wafer 2 1447 1126 846 Wafer 3 1361 938 647 Wafer 4 1049 1061 442 Average 1303 1023 712 Std. 174 87 212 Dev. *Denotes Comparative Example.

[0057] As shown in Table 2, above, the defect counts after CMP polishing dropped dramatically compared to defect counts of the same wafer polished in the same manner with a ceria slurry (Slurry E) or Slurry 4 which lacks the cationic copolymer of the present invention.

Example 2

Performance in Polishing a Pattern Wafer

[0058] In Table 3, below, the substrate was an STI Wafer having a 50% PD feature. Polishing was conducted in multiple steps using the indicated slurry.

TABLE-US-00005 TABLE 3 Polishing With The Cationic Copolymer and Without Example 2A* 2B 2C Slurry Slurry C (1 wt. % solids) 1 2 Performance Parameter A B C A B C A B C Oxide 121 33 20 13 Overpolish 125 Amount, ? 214 263 123 36 68 270 84 57 27 302 171 51 121 372 380 492 154 107 46 522 201 55 146 530 303 95 208 605 663 209 150 59 672 701 246 74 175 804 446 147 299 816 276 184 92 832 870 272 75 232 1071 609 208 401

[0059] As shown in Table 3, above, the trench oxide loss A, SiN loss B and dishing C of the inventive slurry compositions is dramatically improved over time versus just elongated cationic silica Slurry C in Comparative Example 2A.

Example 3

More Performance on a Feature Wafer

[0060] In Table 4, below, the substrate was an STI Wafer having a 50% PD feature. Polishing was conducted in multiple steps using the indicated slurry.

TABLE-US-00006 TABLE 4 Polishing With The Cationic Copolymer Example 3A 3B Slurry 5 3 Performance Parameter A B C A B C Oxide 121 Overpolish 125 26 12 14 Amount, ? 214 78 51 27 263 270 302 372 154 97 57 380 110 109 1 492 522 530 139 153 ?14 605 252 164 88 663 672 205 213 ?7 701 804 816 832 382 249 132 870 1071 * Denotes Comparative Example.

[0061] As shown in Table 4, above, at a slight overpolish, the trench oxide loss A, SiN loss B and dishing C of the inventive slurry compositions are acceptable.

Example 4

Polishing With Various Pads

[0062] In Table 5, below, the slurry 2 was used to polish with two different pads. The substrate was an STI Wafer having a 50% PD feature. Polishing was conducted in multiple steps using the indicated slurry.

TABLE-US-00007 TABLE 5 Polishing with Various CMP Polishing Pads Example 4A 4B Slurry 2 2 Pad IC1010? pad VP6000/K7-R32 pad Performance Parameter A B C A B C Oxide 263 123 36 68 Overpolish 289 144 103 41 Amount, ? 522 201 55 146 544 287 209 79 701 246 74 175 788 409 299 110 870 272 75 232 1009 566 412 155

[0063] As shown in Table 5, above, at a slight overpolish, the trench oxide loss A, SiN loss B and dishing C of the inventive slurry composition 2 are all acceptable with an IC1010? pad (Dow); with a slightly softer pad in Example 4B, dishing C is improved.

Comparative Example 5

Polishing as in Examples 2, 3 and 4 was Performed, Except Using a DADMAC Homopolymer Additive (Polymer 2)

[0064]

TABLE-US-00008 TABLE 6 Comparative Polymer 2 Performance Oxide Overpolish Amount, ? Slurry Performance 344 637 947 1282 1604 6* Trench Oxide 215 386 542 696 879 Loss, ? SiN Loss, ? 39 81 122 176 251 Dishing, ? 176 304 420 520 628 *Denotes Comparative Example

[0065] As shown in Table 6, above, the homopolymer of DADMAC fails to provide anywhere near the polishing performance of the compositions of the present invention having a DADMAC copolymer. Compare results with those in Tables 2, 3 and 4, above.