Effect of Cerium Oxide Particle Sizes in Oxide Chemical Mechanical Planarization

Abstract

This study explored the effect of different cerium oxide abrasive particle sizes in chemical mechanical planarization of 200 mm blanket plasma-enhanced tetraethylorthosilicate wafers. All polishing experiments were done with a polisher and tribometer capable of measuring shear force and down force in real-time. Coefficient of friction and removal rate were found to correlate well with the slurry median particle size distribution. Removal rate modeling based on particle size was explored to support the interpretation of the experimental results.Shallow trench isolation ͑STI͒ is the preferred method of isolation for high-volume integrated circuit ͑IC͒ manufacturing with active device features in the deep submicrometer regime. During IC fabrication, STI incorporates silicon dioxide as the insulator and silicon nitride as the stopping layer. 1 In STI chemical mechanical planarization ͑CMP͒, cerium-oxide-based slurries are commonly used due to their high oxide-to-nitride removal rate selectivity. 2 Recent studies on silicon dioxide and silicon nitride CMP using ceria abrasive slurries have focused on the chemical aspect of the polishing process. 3-7 Generally speaking, cerium oxide slurry is a relatively new formulation compared to the more well-known silicaand alumina-based slurries. Compared to the extensive effort spent on the chemical aspects of the ceria abrasive slurry polishing, there is minimal published work on the effect of cerium oxide particle size in dielectric CMP.This paper investigates the coefficient of friction ͑COF͒ and removal rate of 200 mm blanket plasma-enhanced tetraethylorthosilicate ͑PETEOS͒ wafers using cerium-oxide-based slurries having four different particle size distributions. Wafers were polished using Araca APD-500 polisher and tribometer to measure the shear force and down force in real-time. Removal rate and COF of each slurry are then analyzed. Furthermore, removal rate modeling based on elastic-plastic microcontact mechanics and abrasive wear theory is explored to help interpret the experimental results. ExperimentalFour cerium-oxide-based slurries with different particle sizes ͑summarized in Table I͒ were used to polish 200 mm blanket PETEOS wafers. D 50 and D 99 are defined as the size of the particle on the 50th and 99th percentiles of the total solid weight fraction, respectively. In all cases, the fraction of solid weight content of all slurries, X w , was 0.01. Polishing was performed on the Araca APD-500 polisher, which was capable of measuring friction force and actual down force in real-time during polishing. The polisher and its associated accessories have been described in detail elsewhere. 1,8 In this study, the data acquisition rate for shear force and down force measurement was set at 1000 Hz. For each slurry, five blanket PETEOS wafers were polished for 1 min on a 500 mm IC1000 A2 concentric groove pad manufactured by Rohm and Haas Electronic Materials. Polishing pressure and sliding velocity were kept constant at 3 psi and 1.37 m/s ͑a combination of 93 rpm for the ...

Microscopic Removal Function and the Relationship Between Slurry Particle Size Distribution and Workpiece Roughness During Pad Polishing

“…These approaches also account for the pad roughness (often referred to as pad microasperities), pad hardness, as well as the slurry's particle size distribution (PSD). The focus of these approaches has been largely on understanding and predicting the average material removal rate and how it scales with change in the slurry's PSD and pad properties …”

“…The focus of these approaches has been largely on understanding and predicting the average material removal rate and how it scales with change in the slurry's PSD and pad properties. 12,13 On the other hand, little effort has been devoted to quantitatively understanding and predicting the surface roughness and texture of the workpiece as a function of the PSD and pad properties. Some studies have examined the effect of the slurry's PSD on scratches either by spiking the slurry with larger particles [14][15][16] or by changing the overall PSD.…”

Study on PVA Brush Loading and Conditioning during Shallow Trench Isolation Post-CMP Cleaning Process

“…Ceria actively removes overburdened SiO 2 and inhibits its action at the nitride layer. [15][16][17][18] However, the most serious issue with ceria is its contamination to the oxide surface after polishing. Many researchers have investigated the effect of cleaning solution pH during oxide post-CMP cleaning to remove ceria particles from oxide wafers.…”

Tribological, Thermal and Kinetic Characterization of SiO2and Si3N4Polishing for STI CMP on Blanket and Patterned Wafers

“…Previous studies have been done with SiO 2 and Si 3 N 4 substrates using slurries with ceria nano-particles which have focused on understanding the chemical aspect of the polishing process [2][3][4][5][6][9][10][11]28 as well as the effect of particle size. 29 Investigations on the effects of the ceria-based slurry surfactant concentration and their molecular weight on polish performance have been done by Park et al 2,[9][10][11] Lee et al investigated the effect of chemical additives on the removal rate of SiO 2 and proposed that the resulting planarization without dishing was due to the slurry additives that coated the ceria particles which allowed the oxide layer (present at the "down" regions) of the wafer to be passivated. 4 Chandrasekaran et al proposed that the direct chemical interaction between the ceria nano-particle and the silicon dioxide surface led to high SiO 2 removal rates and that the removal of Si 3 N 4 required the surface layer to become hydrated before it could be mechanically removed by the abrasive ceria nano-particles.…”

“…28 Studies have also been conducted with focus on the tribological and thermal aspects of STI wafer polishing process as well the effect of particle size. 16,29,30 Additionally, Boning et al has provided a number of computational modeling methods for characterization of pattern wafer polishing processes. 19,21,[31][32][33][34][35][36][37] However, the basis of these models is on contact between the pad-wafer interface and are built off of Preston's equations.…”