Effect of Slurry Injection Position on Slurry Mixing, Friction, Removal Rate, and Temperature in Copper CMP

Abstract

In this study, the extent of mixing of old and new slurry on the polishing pad is varied by the use of three different points of injection. Influences of the conditioner and bow wave on slurry mixing can be inferred from the experimental results, which include coefficient of friction data and pad and substrate thermal data. Results measured under identical lubrication mechanisms show that the slurry injection position can play a significant role in slurry mixing and slurry utilization efficiency. Slurry injection positions that induce lower slurry mixing are found to increase copper removal rate. Simulations of the bow wave and slurry puddle support the interpretation of the mixing phenomenon. This work underscores the importance of optimum slurry injection geometry and flow for obtaining a more cost effective and environmentally benign copper chemical mechanical polishing ͑CMP͒ process.The chemical mechanical polishing ͑CMP͒ process has a large environmental and economic effect due to the high use of process consumables, especially slurry. In certain CMP processes, slurry cost may represent up to 50% of the total cost of ownership ͑COO͒ of the CMP module. 1,2 Along with COO issues, environmental concerns can be significant because CMP slurry can contain nonenvironmentally benign and hazardous chemicals. The recovery and reuse of spent oxide CMP slurry has been previously investigated. 3,4 For copper, it is more difficult to reclaim spent slurry due to the nature of the chemical constituents and the presence of complex reaction products. Therefore, optimizing CMP slurry process parameters to reduce slurry usage is the best solution to achieve a more cost effective and environmentally benign process.Several research groups have investigated individual aspects of the role of slurry in the overall CMP process. Subjects of focus have been slurry chemistry, 5-9 slurry film thickness, 10 numerical simulations of slurry fluid mechanics, 11-14 and slurry fluid dynamics. 1,15,16 There has been, however, little experimental research regarding the slurry injection position.Philipossian and Mitchell 1 employed residence time distribution ͑RTD͒ analysis to show that not all of the slurry introduced onto the pad actually participates in the polishing process. Slurry utilization efficiencies ranged from 2 to 22%, depending on the operating conditions and pad types. Slurry utilization efficiency was defined as the ratio of slurry flow rate beneath the wafer to the total slurry flow rate introduced onto the pad. On a k-grooved pad, it was estimated that the maximum slurry utilization was about 5%.Coppeta et al. 16 used dual-emission laser-induced fluorescence ͑DELIF͒ to quantify slurry transport in the gap between the wafer and the pad. With in situ conditioning, slurry mixing was found to be higher and slurry transport efficiency lower than with ex situ conditioning. For a given input pulse of the new slurry, slurry transport efficiency was defined as the volume of new slurry entrained beneath the wafer divided by the total volu...

Chemical mechanical planarization of copper in alkaline slurry with uric acid as inhibitor

“…It is also noted that the reduction tends to saturate at a concentration of 0.05 M. If the static etch rates are high, then the planarization efficiency would be low. The actual etch rate during polishing may be much higher than the ones measured at room temperature since the local temperature can raise by about 10 • C during CMP [32][33][34]. Thus it is necessary to maintain a very low etch rate for effective planarization.…”

Tribological, Thermal, and Wear Characteristics of Poly(phenylene sulfide) and Polyetheretherketone Retaining Rings in Interlayer Dielectric CMP

“…The pad temperature measurement and its averaging method has been described in detail elsewhere. 11 As the polisher had only one single mechanical control on the down force, it could not apply separate pressures to the retaining ring and wafer. Therefore, no wafer was polished during the retaining ring wear tests.…”

Colloid aspects of chemical–mechanical planarization

“…16. There is a significant amount of additional literature where effects of slurry flow and pad/particle/wafer surface interactions have also been analyzed [87][88][89][90][91][92][93][94][95].…”