Reactive Liquids for Non–Prestonian Chemical Mechanical Polishing of Polysilicon Films

Abstract

Slurries which yield non-Prestonian polysilicon (poly-Si) removal rates (RRs) are desirable during the fabrication of semiconductor and microelectromechanical systems (MEMS) devices. Here, we show that aqueous solutions of 250 ppm poly(diallyldimethylammonium chloride) (PDADMAC), DADMAC (monomer of PDADMAC) or guanidine carbonate (GC) at pH 10 give high poly-Si polish rates. Addition of 25 ppm cetylpyridinium bromide (CPB), a cationic surfactant to any of these solutions modifies their polish performance to non-Prestonain behavior. The underlying mechanism for the non-Prestonain removal of poly-Si films with the RL containing 250 ppm PDADMAC + 25 ppm CPB solution at pH 10, chosen as the model system, was investigated using ultraviolet-visible (UV-Vis) spectroscopy, zeta potential measurements and fourier transform infrared (FTIR) spectroscopy. These analyses suggest that CPB adsorbs strongly on poly-Si and suppresses RR at 1 and 1.5 psi to ∼0 nm/min. At higher pressures, the binding between CPB and poly-Si is ruptured due to pressure applied by the polishing pad during CMP process allowing the nitrogen atom of PDADMAC to access and bind to the negative dipole of underlying polarized -Si-Si-bonds. The polarized and weakened backbone Si-Si bonds are broken during polishing, yielding high RRs.

Control of the Micro-Defects on the Surface of Silicon Wafer in Chemical Mechanical Polishing

“…Under low pressure conditions, there is basically no removal rate in the nonabrasive polishing liquid. 26 Compared with non-abrasive polishing liquid, a certain removal rate can be obtained under low pressure (1.5 psi).…”

Selective Polishing of Amorphous Silicon Carbonitride (a-SiCN) Films Over Silicon Dioxide and Silicon Nitride Films for Hardmask Applications

“…Furthermore, several studies have shown that the adsorption of the additive molecules present in CMP slurries on the film being polished can be influenced by resonance, polarizability, chain length, and inductive, dispersion and steric effects. 13,14,[21][22][23] For example, Patri et al 14 found that the copper film RRs obtained with glycine and β-Alanine as additives in silica slurries were significantly different from each other in the pH range 2-10 and argued that the carbon chain length controlled the reactivity of the additives. In another study on polishing of SiN films, Penta et al 13 showed that picolinic acid was more effective for nitride suppression than proline and suggested that the differences were due to steric effects.…”