Chemical mechanical polishing (CMP) has become the key planarization technology in ultralarge scale integration (ULSI) silicon device manufacturing to fabricate sub-quarter-micrometer metal and dielectric lines. [1][2][3][4] In the CMP process, planarization is achieved by polishing a wafer with uneven topography using a colloidal slurry consisting of sub-micrometer sized abrasive particles. The particles are dispersed in an aqueous solution containing various chemicals, which make up the slurry. These chemicals, depending on their identity, play different roles such as oxidizers [Fe(NO 3 ) 3 , (NH 4 )S 2 O 8 , H 2 O 2 ], passivating agents (benzotriazole, benzimidazole), slurry stabilizers [poly(ethylene glycol), arabic gum], etc. While several slurry chemistries are available for Cu CMP and significant progress has been made in utilizing them in manufacturing, very little fundamental information about them is available in the published literature.Copper CMP in a highly acidic pH regime leads to corrosion problems, while Cu CMP in alkaline conditions is faced with an unfavorable polish rate selectivity with respect to SiO 2 , leading to interlayer dielectric (ILD) erosion. Thus, an intermediate pH range 4-7 appears to be a better choice for Cu CMP. 5-7 One of the more attractive slurries in this intermediate pH range consists of hydrogen peroxide, glycine (an amino acid), and an abrasive. Hirabayashi et al. 8,9 demonstrated that slurries containing hydrogen peroxide, glycine, and silica particles (abrasive) can be used for Cu CMP. They successfully fabricated inlaid copper wiring using the damascene process with very little dishing (less than 60 nm in the linewidth range of 0.5-100 m) with the above slurry. The mechanism, as proposed by them, consists of the oxidation of Cu to copper oxide by H 2 O 2 in the recessed region, thereby preventing the dissolution of Cu from the recessed areas of the wafer. The oxide formed in the protruded regions, on the other hand, is removed by the abrasives exposing the underlying metallic Cu surface to the slurry. According to Hirabayashi et al. copper is then converted into Cu(H 2 O) 4 2ϩ by the hydrogen peroxide in the slurry which in turn reacts with glycine (also in the slurry) to form a Cu 2ϩ -glycine chelate that is soluble in water. Thus, while the formation of copper oxide prevents the direct etching of Cu in the low lying regions, Cu in the protruded regions is removed by both direct dissolution as well as by the removal of the oxide formed. However, they did not describe the interaction between the Cu 2ϩ -glycine complex and hydrogen peroxide and its effect on Cu removal or the role of hydroxyl radicals. It has been well established that the decomposition of hydrogen peroxide leads to the formation of hydroxyl radicals (*OH) which are a much stronger oxidizing agent than hydrogen peroxide itself. 10,11 There is also a rich collection of information concerning the catalytic generation of *OH from hydrogen peroxide, with various metal ions and metal ion complexes acting ...
The effects of chemical mechanical planarization ͑CMP͒ process parameters and consumables on the polish rate and defects generated in various low-k materials with k values ranging from 2.2 to 3.0 were studied in detail. The process consumables and conditions evaluated include slurry material ͑alumina and silica͒, pad type ͑soft and hard͒, polish pressure, and polish time. Atomic force microscopy ͑AFM͒ images and roughness numbers were used to evaluate the post-CMP defect generation under various process conditions and revealed nano/microscratches, pits, voids, and film delamination. For a given material, the removal rate increased with increasing pressure. The dependence of defects on pressure appeared highly driven by the slurry material, pad type, and low-k material properties. Defects increased with increasing pad hardness and decreasing k values. The increasing defects with decreasing k value can be attributed to the lower elastic modulus observed with low-k materials. Removal rates exhibited both a decrease and an increase in conjunction with polish time, depending on the type of low-k material used. AFM analysis showed an improvement in global surface roughness with increasing polish time; however, an increase in localized defects such as pits was also observed. Fourier transform infrared and X-ray photoelectron spectroscopy analyses showed no change in film chemistry under the conditions studied here.With advancements in ultralarge-scale integration, integrated microelectronic device dimensions are continually being scaled down. At present, manufacturers are running 0.11 m technology in their production lines, which increases the number of dies per wafer and reduces manufacturing costs. In addition, higher device speed requirements have forced the industry to include material changes in the device interconnection array. For example, in an attempt to reduce the resistance-capacitance time delay, manufacturers have moved toward a copper damascene process, with copper replacing aluminum and low-k dielectric materials replacing SiO 2 .Chemical mechanical planarization ͑CMP͒ is an important step in the copper damascene process. Copper deposited in the patterned dielectric trenches is planarized by CMP. The primary objective of the Cu-CMP process is to planarize and remove excess copper and the barrier layer. During this process, the underlying dielectric material is exposed. With industry shifts toward ultra low-k materials, very low k values can usually be achieved by using porous/organic materials, which also exhibit low hardness and modulus. However, these relatively weak materials cannot withstand the rigorous CMP process. Additionally, low-k films are less moisture resistant than conventional oxides produced by chemical vapor deposition ͑CVD͒; thus, extended exposure to CMP slurries and the environment affects their properties.Numerous low-k materials are being investigated. The compatibility of these materials with the CMP process is a significant factor in determining the appropriate low-k material to us...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.