The chemical mechanical polishing of copper in several slurry chemistries based on iodate and iodine oxidizers has been investigated. Benzotriazole (BTA) and potassium iodide (KI) were used for preparing the polishing slurry chemistries based iodate. As observed by the anodic electrochemical behavior of copper and the surface analyses of EDS, it was determined that CuI layer formed in the iodate and iodine based solutions. Especially, in I2 slurry in pH 4, CuI layer formed very fast and uniformly, and passivated the copper. In addition, the highest removal rate using this slurry was obtained. These results were compared to H2O2 based slurries. From these experimental results, the slurry containing 0.1M KIO3 and 0.01M KI, and 0.01N I2 give better results than H2O2 based slurry in copper CMP.

Wafers where deposited with BPSG films having phosphorus concentration varying from 3.65 to 6.25% and boron concentration varying from 4 to 5.7%. These wafers were polished using CMP and the rates were found to depend on dopant concentrations. A fit to the data indicated that removal rates were more than 3 times as sensitive to boron concentration compared to phosphorus concentration. For a constant phosphorus concentration of 5%, each percent increase in boron increases CMP removal rate by 340 Å/min. For a constant boron concentration of 5%, each percent increase in phosphorus increases CMP removal rate by 96 Å/min.

A new set of wafer-scale patterns has been designed for analysis and modeling of key CMP effects. In particular, the goal of this work is to develop methods to characterize the planarization capability of a CMP process using simple measurements on wafer scale patterns. We examine means to pattern large trenches (e.g. 1 to 15 mm wide and 15 mm tall) or circles across 4” and 8” wafers, and present oxide polish results using both stacked and solo pads in conventional polish processes. We find that large separation (15 mm) between trenches enables cleaner measurement and analysis. Examination of oxide removal in the center of the trench as a function of trench width shows a saturation at a length comparable to the planarization length extracted from earlier studies of small-scale oxide patterns. Increase in polish pressure is observed to decrease this saturation point. Such wafer scale patterns may provide information on pad flexing limits in addition to planarization length, and promise to be useful in both patterned wafer CMP modeling and studies of wafer scale CMP dependencies such as nanotopography.

In this paper, the experimental results of interfacial fluid pressure and friction measurements during polishing are presented, as well as their dependence on some major process variables. Under simulated conditions, a sub-ambient fluid pressure was observed, and its magnitude was of the same order of magnitude as the applied polishing load. Since this fluid pressure is non-uniformly distributed, the contact stress, obtained by combining the effects of both applied load and the fluid pressure, is not uniform across the wafer and will result in non-uniform material removal. The mechanism of the presence of the fluid pressure was investigated, and an analytical model was developed to predict the magnitude and distribution of this fluid pressure. The effects of the sub-ambient fluid pressure on material removal rate and profile were tested with thermally grown silicon dioxide on 100mm diameter, P-type (100), single crystal silicon wafers. The polishing experiments show the effect of sub-ambient fluid pressure on polishing rate and profile.

Experimental results on chemical mechanical polishing of copper and tantalum in the presence of different chemicals, such as Fe(NO3)3, H2O2 and glycine, using both silica and alumina abrasive particles are presented. The polish rates with alumina slurries vary in accordance with the hardness of the material being polished, suggesting a dominant role for a wear mechanism. However, polish rates with silica slurries do not relate directly to the hardness of the film, indicating a more complex removal mechanism.

A generic model is presented which explains the dependence of chemical mechanical polishing rates on the concentration of reacting chemicals and abrasives in the slurry. The predictions of this model are compared to data from the literature for tungsten CMP.

Silica slurry used as abrasives in wafer polishing process is made by dispersing silica particles in an alkali solution. Since commercially available colloidal or fumed silica particles need some modifications to be directly used as abrasive slurry due to their small sizes, irregular shapes or broad size distribution, we have prepared silica abrasives by particle growth of fumed silica or colloidal silica as seeds by sol-gel method. Silica slurries prepared by this step-wise growth from commercial seeds were tested using one-armed polisher for the comparison with commercial slurries and showed the performance comparable to commercial slurries. Microstructures of polishing slurries were investigated using transmission electron microscopy and ARES rheometer. From the result, stability of the slurry was found to be more important than the primary particle sizes for the polishing performance.

The purpose of this work was to investigate integration of aluminum multi-level damascene devices with the CMP module. Traditionally aluminum is used in the Metal-1 level of device manufacturing and Reverse Ion Etching (RIE) is used to remove the aluminum over layer. However for <0.25-µproportional]m device rules RIE is not a method of choice due to incomplete removal of the over layer leaving stringers that cause shorting, as well as poor With-In Wafer NonUniformity (WIWNU). The result is loss of device yield and process problems for BEOL modules. Integrating aluminum device wafers into the CMP module has its own drawbacks such as immature consumables, i.e., slurry and polishing pads, as well as ease of scratching the soft aluminum interconnect structures. Hence the basis of this work is to highlight the main issues that impede the integration of the aluminum CMP process. Specifically we investigated the origin of deep aluminum scratches and the effect of not completely removing the residue barrier material, which can cause shorting and poor electrical performance. It is argued in this work that the observed deep scratches in the aluminum material are due to the polished debris emanating from the titanium glue layer. Recommendations are made to help to reduce this effect by design modification to the die layout on the patterned wafer and using electrical testing methods to help ascertain the minimum Over Polishing (OP) time required in order to ensure maximum die yield.

In semiconductor device production, wafers are treated through many cleaning processes. Usually, several chemicals are used so as to match for several purposes like RCA cleaning method. As wafer size becomes larger, large amount of cleaning chemicals usage and waste are necessary, which becomes now a big problem. Considering the above, we developed the Electrolyzed D.I.water with chemicals supply system in order to minimize running cost of chemicals and waste treatment. It is feature that; 1) this system can generate the anode water of the acidity/high ORP (Oxidation-Reduction Potential) and the cathode water of the alkalinity/low ORP by electrolyzing D.I.water adding with a small quantity of chemicals; 2) this system can generate anode water and cathode water at the same time; 3) if necessary, the anode water can be diluted with D.I.water at the optional density, and it is possible with the cathode water that hydrogen peroxide is added. The anode water which shows acidity/high ORP has the effect of removing metal and organic contamination, and the cathode water which shows alkalinity/low ORP has the effect of particle removal. In this report, we explain the outline of this system and the basic characteristic of the Acid and Alkaline water made with this system and its performances.

A novel approach for a single step lapping and chemical-mechanical polishing of antimonide-based III-V compounds using agglomerate-free alumina slurries is presented. Relatively high removal rates, minimal scratching, and low surface roughness have been obtained. The effects of slurry preparation cycle on the slurry properties and chemomechanical polishing results are discussed.

In this paper, results of studies on the addition of salt to a polishing slurry, in terms of its effect on slurry stability, SiO2 polishing rate and surface roughness of the polished surface are presented. Three salts, viz. LiCl, NaCl and KCl were selected, and three concentrations were tested. Polishing rate measurements using these slurries show that adding salt leads to increased removal rate without affecting surface roughness significantly. Based on these results, we can say that the agglomerates formed by adding salt to the slurry are fairly soft and easily broken during the polishing process. In addition, turbidity and particle size measurements show that significant coagulation of the particles in the slurry occurs only at the highest salt concentration, and is fastest for LiCl and NaCl, with KCl showing the slowest coagulation. From these results, it can be concluded that the enhancement in polish rate is due to increased contact at the wafer-pad-slurry interface, and not due to formation of larger agglomerated particles in the slurry. This is because of reduced electrostatic repulsion between these three surfaces, due to the screening of their negative surface charge by the metal ions in solution, resulting in a higher wear rate.

In the semiconductor industry, there is a need to establish fundamental, mechanism-based, correlation(s) between process conditions, consumables (e.g., pads and slurries), and observed process performance in Chemical-Mechanical Polishing (CMP). In this paper, we present recent results of studies on polyurethane-based CMP pads in static and dynamic conditions using Dynamic Mechanical Analysis (DMA) to monitor modulus and energy damping changes. Two-layered, composite IC1000 on Suba IV pads, [IC1000 (cast and cured polyurethane elastomer) / Suba IV (polyurethane impregnated polyester felt)], were analyzed: prior to use (fresh); after a 24-hr soak in silica-containing oxide slurry (basic); and after oxide polishing (used). Upon comparison it was observed that a characteristic transition feature due to water is present at sub-ambient temperatures in both the slurry soaked and used pads. Exposure of as-received pads to basic oxide slurry results in a broad, high temperature transition thought to be the result of chemical-induced disruption of macrostructural units. Polishing (load-enhanced chemical exposure) introduces further changes to the polymer network represented by an apparent degradation to the structural species responsible for the high temperature transition in Suba IV.

Chemical-Mechanical Polishing has been revealed as an attractive technique for poly-Si of trench planalization. Major issue of the process integration is pattern erosion after over polishing. A new process with silica slurry containing organic surfactant is reported in this paper. A patterned wafer after conventional CMP process is eroded by over polishing, however, the new process conducts small erosion for wide trenches. The organic surfactant is well known as a inhibitor for the protection of poly-Si from alkaline, and the new slurry shows a large pH dependency of the viscosity. The experimental work has been focused on the viscosity, and the mechanism of the small erosion is discussed. This new process should be useful for recessing poly-Si by CMP, because it keeps the erosion level very low.

Chemical mechanical planarization (CMP) has been widely used for planarization of ILD, STI, plug and wiring processes. Wafer has several surfaces of materials, such as wiring materials, barrier materials, dielectric materials etc., that must be cleaned at the same time. In post metal CMP cleaning processes, in addition to cleaning several surfaces, it is very important that the oxidization level of metal materials, such as wiring, is held and controlled to maintain its resistance. Especially copper, that is began to use for wiring, is very easy to be oxidized. We have confirmed that the Electrolyzed D.I.water is effective in post Cu CMP cleaning for controlling the surface condition of Cu during cleaning and leaving a robust surface after CMP. We describe the Electrolyzed D.I.water system and present some result of analysis of Cu surface by treated with the Electrolyzed D.I.water.

Chemical mechanical polishing (CMP) of copper using alumina-based NH4OH slurry containing benzotriazole (BTA) has been evaluated in terms of polish efficiency and viability. Dishing of damascene copper patterns can result from a combination of chemical dissolution and mechanical abrasion due to the deformed polishing pad bending into the recessed copper regions. The addition of at least 0.1 wt.% BTA to the slurry leads to the formation of a thin Cu(I)-BTA polymer on the copper surface during CMP. This polymer reduces the amount of dishing by an order of magnitude. At the same time, however, the CMP polish rate falls sharply with the addition of 0.1 - 0.25 wt.% BTA to the slurry. Above 0.25 wt.% BTA, the polish rate falls no further. Stability of alumina particles in the NH4OH slurry is found to deteriorate with the addition of BTA. Integrated copper/barrier electromigration resistance test structures with large contact areas (2×2mm) have been successfully patterned using a 2-step CMP/etching process scheme, using a BTA-containing slurry to minimise dishing.

Chemical mechanical polishing is an essential process for achieving a high degree of planarization. The planarity after CMP sensitively depends on pattern scales, pattern densities and mechanical properties of polishing pads. In order to simulate the topography after CMP, a numerical model for the polishing pad is proposed. In this model, the surface roughness layer of the polishing pad is assumed as a flat soft layer. The distribution of the contact pressure between the patterned wafer and the polishing pad is calculated with finite element method, and the pattern topography is modified based on the pressure dependency of the polishing rate. The iterations of the contact pressure analyses and the topography modifications give the progress of the polishing process numerically. The model is applied to oxide CMP process with silica slurry and stacked pad of polyurethane and non-woven fabric. The compressive elastic moduli of polyurethane layer and non-woven fabric layer are measured dynamically. The elastic modulus of the soft layer is treated as a fitting parameter between the experimental results and the numerical model. The models with the elastic modulus of 10 MPa for the soft layer show good agreements with the experimental results in both of a short range, where the compressive deformation of the pad is dominant, and a long range, where the bending deformation is dominant. Static measurements for the surface elasticity of the polyurethane layer also give a good agreement with the model. The proposed pad model should be useful for the topography simulation, and it also guides the development of new polishing pads.

Understanding the time-dependent behavior and size of particulate systems, specifically, particles in abrasive slurries, is a key part of chemical mechanical polishing (CMP). Microscopy and image analysis enables both qualitative and quantitative analysis of particle interaction behavior in diluted samples of such particulate suspensions. A technique using microscopy and image analysis has been developed specific to the analysis of suspensions of, micron size or larger, abrasive particle slurry systems. This technique has been applied to aluminum oxide (Al2O3) slurries and measurements of particle size and stability have been obtained with good accuracy and precision.

The fluid film thickness and drag during chemical-mechanical polishing are largely dependent on the shape of the wafer polished. In this study we use dual emission laser induced fluorescence to measure the film thickness and a strain gage, mounted on the polishing table, to measure the friction force between the wafer and the pad. All measurements are taken during real polishing processes. The trends indicate that with a convex wafer in contact with the polishing pad, the slurry layer increases with increasing platen speed and decreases with increasing downforce. The drag force decreases with increasing platen speed and increases with increasing downforce. These similarities are observed for both in-situ and ex-situ conditioning. However, these trends are significantly different for the case of a concave wafer in contact with the polishing pad. During ex-situ conditioning the trends are similar as with a convex wafer. However, in-situ conditioning decreases the slurry film layer with increasing platen speed, and increases it with increasing downforce in the case of the concave wafer. The drag force increases with increasing platen speed as well as increasing downforce. Since we are continually polishing, the wafer shape does change over the course of each experiment causing a larger error in repeatability than the measurement error itself. Different wafers are used throughout the experiment and the results are consistent with the variance of the wafer shape. Local pressure measurements on the rotating wafer help explain the variances in fluid film thickness and friction during polishing.

The dependence of IC fabrication on the Chemical Mechanical Planarization (CMP) process increases as the device features go down to 0.25 micron or beyond. Due to the tighter CMP process spec, it is very important to reduce the within wafer non-uniformity (WIWNU%) to achieve higher process yield. The symmetrical increment of linear velocity at wafer edge is not sufficient to change wafer edge profile by breaking the matched speed rule. A better solution is through the change of head design for a fixed platen from the polisher design point of view. This study demonstrates the improvement of the CMP process performance, especially at the wafer edge, from the modification of the floating type polish head. The best WIWNU% from a single air chamber head is about 5.12% at 6-mm edge exclusion (EE). In order to obtain better pad deformation control, the retaining-ring pressure chamber is separated from that of the sub-carrier. The average WIWNU% is about 4% for 3-mm and 5-mm EE from two-pressure-chamber head. Due to the limitation of retaining-ring pressure effect, a third pressure chamber is further added that can be extended the edge control up to 1 inch from the wafer edge. The WIWNU% is about 3.8% at 5-mm edge exclusion with low down forces. The slurry and insert types also show effect on the wafer edge profile. It has been also proven that this three-pressure-chamber head is able to reduce the post-CMP thickness variation from the ILD production wafer, especially at wafer edges. More detailed information and CMP mechanism will be discussed in this paper.

In this work, we used surface analysis techniques, such as a field-emission highresolution analytical TEM, X-ray spectroscopy, and XPS to analyze abrasive particles after polishing. Results showed evidence of copper oxide (Cu2O) in the polished slurry. However, there was no metallic crystalline copper detected. After comparing these data with the results obtained from our electro-chemical experiments, we propose two possible chemical wear mechanisms in Cu CMP.