Chemical mechanical polishing pad

Abstract

The present invention provides methods of CMP polishing a metal surface, such as a copper or tungsten containing metal surface in a semiconductor wafer, the methods comprising CMP polishing the substrate with a CMP polishing pad that has a top polishing surface in a polishing layer which is the reaction product of an isocyanate terminated urethane prepolymer and a curative component comprising a polyol curative having a number average molecular weight of 6000 to 15,000, and having an average of 5 to 7 hydroxyl groups per molecule and a polyfunctional aromatic amine curative, wherein the polishing layer would if unfilled have a water uptake of 4 to 8 wt. % after one week of soaking in deionized (DI) water at room temperature. The methods form coplanar metal and dielectric or oxide layer surfaces with low defectivity and a minimized degree of dishing.

Claims

1. A CMP polishing pad useful for polishing a metal surface on a semiconductor wafer, the CMP polishing pad comprising: a porous polishing layer, the porous polishing layer having a top polishing surface, the CMP polishing pad and porous polishing layer consisting essentially of: a cured reaction product of an isocyanate terminated urethane prepolymer, the isocyanate terminated urethane prepolymer being a blend of toluene diisocyanate, polytetramethylene glycol and 4,4-methylene dicyclohexyl diisocyanate having 8.5 to 9.5 wt. % of unreacted NCO groups; and a curative component consisting essentially of from 60.3 to 70 wt. %, based on the solids weight of the curative component, of a polyol curative having a number average molecular weight of 6,000 to 15,000 and having an average of 5 to 7 hydroxyl groups per molecule; and from 30 to 39.7 wt. %, based on the solids weight of the curative component, of 4,4 methylene bis (2 chloroaniline), wherein a cured unfilled formulation of the CMP polishing pad lacking porosity, has a water uptake of 4 to 8 wt. % after one week of soaking in deionized (DI) water at room temperature and wherein the polishing layer having the porous top polishing surface has a Shore D hardness ranging of from 28 to 64 according to ASTM D2240 (2015), the top polishing surface being adapted for planarizing and polishing the metal surface of the semiconductor wafer.

2. The CMP polishing pad as claimed in claim 1, wherein the curative component of the polyol component is 60.6 to 65 wt %, based on the solids weight of the curative component.

3. The CMP polishing pad as claimed in claim 1, wherein the CMP polishing pad has a Shore D hardness of from 30 to 60.

4. The CMP polishing pad as claimed in claim 1, wherein the CMP polishing has a Shore D hardness of from 35 to 55.

5. The CMP polishing pad as claimed in claim 1, wherein the CMP polishing pad comprises CMP polishing stoichiometric ratio of reactive hydrogen groups (defined as sum of NH.sub.2 and OH groups) in the curative component to the unreacted NCO groups in the isocyanate terminated urethane prepolymer ranges from 0.85:1 to 1.15:1.

6. The CMP polishing pad as claimed in claim 1, wherein the curative is substantially free of any diol.

7. The CMP polishing pad as claimed in claim 1, wherein the CMP polishing pad has a specific gravity (SG) of from 0.4 to 1.15 and the cured unfilled formulation of the CMP polishing pad has an SG of from 1.1 to 1.2.

8. The CMP polishing pad as claimed in claim 1, wherein the CMP polishing pad and porous polishing layer comprise gas filled or hollow microelements.

Description

EXAMPLES

(1) Table 1 below summarizes compositions of a polishing layer of the present invention and a prior art example.

(2) Prepolymer 1 comprises Adiprene L325 prepolymer having 8.95 to 9.25 wt. % unreacted NCO groups, made from toluene diisocyantae (TDI), polytetramethylene glycol (PTMEG), and 4,4-methylene dicyclohexyl diisocyanate (H.sub.12MDI) (Chemtura, Philadelphia, Pa.);

(3) MbOCA is 4,4-methylene-bis-(2-chloroaniline); and,

(4) Voranol 5055HH polyol (Dow) has an average of 6 hydroxyl groups per molecule and a number average molecular weight of 11,400.

(5) TABLE-US-00002 TABLE 1 CMP Polishing Layer Compositions Curative Component Aromatic EX. amine curative Polyol Curative Molar ratio NO Prepolymer (% NCO) (wt. %) (wt. %) (Curative/NCO) 1C* 1 9.1 MbOCA 53.5 Voranol 46.5 1.05 5055HH 1 1 9.1 MbOCA 39.2 Voranol 60.8 1.05 5055HH *Denotes Comparative Example.

(6) The formulations in Table 1 were cured at 104 C. for a period of 15.5 hours to make a bulk cured polishing layer material. Unfilled bulk materials (38 mm by 38 mm squares of 2 mm thick) were soaked in deionized (DI) water for one week. Water uptake was calculated by the percent weight change of each material. Pad hardness was measured before and after one week of water soaking, referred as dry and wet hardness, respectively. The wet hardness was measured on soaked samples after surface water had been removed. Sample hardness was recorded at both 2-seconds (2 s) and 15-seconds (15 s). Table 2, below summarizes water uptake and Shore D hardness measured before and after one week of water soaking.

(7) TABLE-US-00003 TABLE 2 Polishing Layer Properties Shore D Shore D Shore D Shore D Hard Water Ex SG, hardness, hardness, hardness, hardness, segment uptake, No. unfilled 2 s dry 15 s dry 2 s, wet 15 s, wet content wt. % 1C* 1.15 61.7 58.8 54.7 49.2 42% 3.4 1 1.14 54.1 50.1 46.0 37.2 36% 5.0 *Denotes Comparative Example.

(8) The hard segment content was calculated from the percentage hard segment in the polishing layer contributed mainly by the isocyanate component in the isocyanate terminated urethane prepolymer and the polyfunctional aromatic amine curative in the curative component. The balance was considered the soft segment, comprising mainly the prepolymer polyol used in making the isocyanate terminated urethane prepolymer and the polyol curative in the curative component.

(9) To evaluate the polishing layers shown in Table 1, above, the polyurethane materials were skived to a thickness of 2 mm (80 mil) thick and were grooved in a circumferential K7 pattern (concentric circular groove pattern, with a groove pitch of 1.78 mm (70 mils), groove width of 0.51 mm (20 mils) and groove depth of 0.76 mm (30 mils)) to form a 775 mm (30.5 inch) diameter polishing layer; the polishing layer was then adhered to a SP2310 sub-pad, commercially available from The Dow Chemical Company, Midland, Mich. The resulting CMP polishing pads were used with a commercial bulk copper slurry (CSL9044C colloidal silica abrasive particle slurry, Fujifilm Planar Solutions, Minato, Tokyo, Japan) at 300 ml/min to polish both copper sheet and pattern wafers. Polishing was conducted in an Applied Materials Reflexion LK 300 mm polisher. The CMP polishing pad was broken in with a Kinik I-PDA31G-3N conditioning disk (Taipei City, Taiwan, R.O.C.) at 2.27 kg (5 lbs) down force for 30 mins with DI water, with the platen rotating at 93 rpm and the conditioning disk rotating at 81 rpm. Following break in, the CMP polishing pad was used to polish 20 pcs of tetraethoxysilane (TEOS) oxide dummy wafers, before performance data was collected on three copper sheet wafers and one copper pattern wafer. Polishing the copper sheet and pattern wafers was performed with a polishing down force of 1.135 kg (2.5 psi), with the platen rotating at 93 rpm, and the wafer/carrier rotating at 87 rpm. After polishing, the substrate was cleaned with CX-100 cleaning solution (Wako Pure Chemical Industries, Ltd, Osaka, Japan). After cleaning, defectivity and dishing performance was determined.

(10) Defectivity on Cu Sheet Wafers:

(11) Defectivity was determined using Surfscan SP2 unpatterned wafer surface inspection tool (KLA-Tencor Corporation, Milpitas, Calif.) at a threshold 0.049 um. At 1.135 kg (2.5 psi) polishing down force, the average number of scratches in the polished copper sheet wafers are shown in Table 3, below.

(12) TABLE-US-00004 TABLE 3 Polishing Defectivity Results Ex No. Average scratch/chatter mark defect count 1C* 50.3 1 14.3 *Denotes Comparative Example.

(13) The method of the present invention has demonstrated significantly improved defectivity performance than the methods of the Comparative Example 1C.

(14) Dishing on Pattern Wafer:

(15) Dishing during the polishing of a pattern wafer was evaluated using an end-point-detection (EPD) system to detect the endpoint for copper (Cu) film clearing (the end point of polishing). The EDP system detects a signal change as different film materials are exposed by polishing. An over-polish step was performed after the endpoint was detected to ensure the Cu film was cleared except inside the oxide trenches of the pattern wafer. The polish time of the over-polish step was set as a fixed percentage of the previous EPD time to achieve a similar over-polish amount in different wafers or different pads. After polishing, a Bruker AFM Probes tool (Billerica, Mass.) was used to measure dishing and topography at different feature sizes on the polished pattern wafer and to calculate total-indicated-range (TIR) using the tool's software. The dishing performance from Cu pattern wafer polishing is shown in Table 4, at feature sizes of both 100100 m and 5050 m.

(16) TABLE-US-00005 TABLE 4 Dishing Performance Dishing improvement Dishing improvement Ex No. at 100 100 m at 50 50 m 1C* Control Control C 18% 12% *Denotes Comparative Example.

(17) Table 4 shows substantial dishing improvement from the methods of the present invention when compared to the method of polishing using a CMP polishing pad made using less of the polyol curative of the present invention in Comparative Example 1C. It was unexpected that methods of polishing with a softer pad from in accordance with the present invention provides similar or better dishing than the methods of polishing with a harder pad as in Comparative Example 1C.