Two-Additive Electrolytes for Superplanarizing Damascene Cu Metals

Abstract

Two-additive electropolishing ͑EP͒ electrolytes exhibit an extremely high planarization-efficiency in Cu damascene schemes, independent of pattern sizes ͑1-50 m͒. This electrolyte is demonstrated by adding alcohols and organic acids to the H 3 PO 4 electrolyte. The high wetting ability of alcohols allows such additives to easily access the damascene bottom. This mechanism, assisted by the reduced polishing rate associated with the high surface viscosity caused by alcohol additives, greatly passivates the damascene bottom from elecropolishing. Accordingly, the superpolishing functionality reported by Chang et al. ͓Electrochem. Solid-State Lett., 6, G72 ͑2003͔͒ in the two-additive electrolyte outperforms in the one-additive electrolyte.Cu damascene schemes continue to dominate the next generation of interconnect preparation in the integrated circuit ͑IC͒ industry as feature sizes are scaled down below severaltens of nm. Chemical mechanical polishing ͑CMP͒ of copper and barrier metals is the current method for defining a planarized interconnection area. Recently, due to the introduction of fragile low-k films ͑such as porous dielectrics͒ as intermetal dielectrics ͑IMD͒, minimization of the mechanical components from the Cu-CMP becomes more important. 1,2 Copper electropolishing ͑Cu-EP͒ has recently been introduced to replace Cu-CMP because of its stress-free nature ͑no mechanical force component͒, high polishing rate, low volume of waste stream, no requirement of abrasives and no scratching.Many studies have reported the feasibility of Cu-EP for planarizing Cu damascenes. [3][4][5][6] Our group also demonstrated a one-additive ͑organic acid species͒ EP electrolyte, a so-called superpolishing electrolyte, which can improve the planarization-efficiency ͑PE͒ of Cu-EP. 7 This one-additive EP electrolyte provides the functionality of a selective Cu dissolution rate within damascene features. However, the PE of superpolished Cu still depends on the feature size; the best value is ϳ34% for 50 m patterns. With reference to two ͑or three͒-additive electroplating electrolytes, 8,9 the introduction of different EP functionalities from the various additives is a possible solution. Wetting additives for Cu electroplating, like polyethylene glycols ͑PEG͒, alcohols were found to access easily the reacting surfaces in EP. 10,11 Moreover, the solution viscosity for alcoholcontaining electrolytes is higher than for alcohol-free electrolytes. 12 Because Cu dissolution is suppressed by increased solution viscosity near polished interfaces, using alcohol-containing EP electrolytes for planarizing damascene features leads to numerous alcohol additives at the damascene bottom, thus passivating that area from EP. This proposed inhibitor, assisted by the superpolishing functionality generated by the second additive of organic acids, 7 hence constitutes the formula of two-additive EP electrolytes.This paper demonstrates an extremely efficient two-additive EP electrolyte, which consists of organic acid and alcohols in the conventional Cu EP e...

Electropolishing—A Practical Method for Accessing Voids in Metal Films for Analyses

“…Already in the 20th century, many publications released contributed to obtaining a fundamental understanding of its underlying mechanisms [29,30,[32][33][34][35][36][37][38][39]. As of today, recent review publications [40,41] demonstrate that still a lot of research on electropolishing is performed, especially towards the direction of fine tuning process parameters [42] and the chemical composition of the used electrolyte [43,44]; often in the application of specialized industrial manufacturing processes [45][46][47][48][49][50]. Within the present publication, a simple experimental setup for electropolishing single microelectronic chips with a copper metallization is presented and optimized in terms of micro smoothing results.…”

Roles of Additives in Damascene Copper Electropolishing

“…3a and b. With reference to the glycerol-containing one-additive electrolyte, 5 all organic acids, except acetic acids, considerably reduced the glycerol-induced inhibition of the removal of Cu at the damascene bottom. Therefore, the difference between the PE of these twoadditive electrolytes is closely related to the change in the alcoholrelated inhibition of Cu removal rate at the feature bottom in the presence of various organic acids.…”

“…The patterned wafer used in PE measurements consisted of a 30 nm thick ionized metal plasma ͑IMP͒-TaN layer as the diffusion barrier, a 200 nm thick IMP-Cu film as the seed layer, and a 1.7 m thick electroplated Cu as the conduction layer, deposited into trenches with various widths ͑1-50 m͒ and a depth of 1 m. Experiments on Cu electroplating and electropolishing were performed as described elsewhere. 5 In Cu EP, the EP electrolytes were phosphoric acid ͑85% H 3 PO 4 ͒ and contained various organic acids and glycerol alcohols; the films were polished under potentiostatic control at room temperature.…”

“…The selective Cu dissolution rate, the acidity, the resistance of the passivation layer, and even the interfacial compositions within features cannot be easily determined experimentally, so alternative approaches should be developed. Our previous work stated that electrolytes with additives can be used at various concentrations to electropolish Cu metals, and the associated polishing rates are analyzed to determine the gradient of the Cu removal rate, 5,16 ͑R up − R down ͒/R up . Herein, R up represents the Cu removal rate at the damascene opening and is obtained using two-additive EP electrolytes with additives at optimal concentrations.…”

Study of Cu Electrochemical Polishing Mechanism With Observation of Water Acceptor Diffusion

“…This Cu ECP method has several advantages, such as a solution-based process that uses simple equipment, selectivity for the conductive substrate, and causing no mechanical damages. Due to these unique properties, the Cu ECP process can be used in various applications that treat surfaces for cosmetic purposes ( Du and Suni, 2004 ), for substrates for graphene growth ( Zhang et al, 2012 ), for TEM and EBSD samples ( Sun et al, 2005 ; Lapeire et al, 2013 ), and for Cu planarization in semiconductor interconnections ( Chang et al, 2002 ; Chang et al, 2003 ; Padhi et al, 2003 ; Liu et al, 2005 ; Suni and Du, 2005 ; West et al, 2005 ; Liu et al, 2006 ; Liu et al, 2006 ). Despite the many potentials of Cu ECP, its actual mechanisms are still controversial.…”