Ultrafine silica, alumina and doped alumina/silica powders have been produced by Chemical Vapor Synthesis (CVS). Primary particle sizes from 3.5 to 12.4 nm have been achieved. Agglomerate sizes ranging from approx. 10 to 150 nm have been achieved in aqueous dispersions.

To develop high signal speed semiconductor LSIs, Cu interconnection is one of the most important requirements. In the fabrication of Cu interconnections using CMP method, minimized dishing, erosion and reduction in micro-scratches are large issues to be realized. We performed a research for superior properties of Cu CMP. Finally, we succeeded in developing Abrasive Free (AF) Cu slurry suitable for these requirements.

We also developed slurry for barrier (TaN) with a high selectivity between TaN and SiO2 of 50 to reduce oxide (SiO2) loss. This reduced oxide loss is directly related to obtaining a controlled circuit resistivity.

By applying these two kinds of slurries, ULSIs with multilevel Cu interconnects and excellent reliabilities were obtained.

The aim of this study has been to synthesize the microporous silica spheres and to coat as-synthesized SiO2 with CeO2 for CMP applications. First, spherical microporous silica powders with a narrow size distribution have been prepared by a precipitation technique involving the hydrolysis reaction of a silicon alkoxide in ethanol. The interparticle microporosity has been created by adsorption of an organic compound (glycerol) as the porogen. The presence of glycerol during the synthesis affects considerably the precipitation mechanism and its effect on the particle size will be discussed. The synthesis of silica microporous spheres of narrow size distribution yielded the preparation, by varying particle size and porosity, of a wide range of aqueous silica slurries. The influence of particle size, particle size distribution, porosity and particle concentration will be discussed in chemical mechanical polishing applications. Although silica particles show large plastic deformation than the bulk material, very good glass polishing rates are obtained due to the plastic deformation of the silica layer during CMP. Silica particles are suitable candidates for application in CMP because silica can be directly precipitated as monodispersed spheres, their narrow size distribution being an important requirement in CMP applications. Secondly, as-synthesized silica particles were coated with the cerium dioxide particles having hexagonal shape, which were precipitated by decomposition from the cerium precursors. For this study, three coating processes were introduced to investigate the best coating parameters. Improvements in CMP of glass were also obtained by coating silica particles with cerium oxide nanoparticles.

In a previous CMP model, the effects of the slurry concentration of chemicals and abrasives on the polishing rate were quantitatively explained, for constant polishing pressure and velocity. That model is extended here to consider specific descriptions of mechanical abrasion, involving polishing pressures and speeds. Predictions of behavior are compared to literature data, leading to an improved understanding of the abrasive process in CMP.

The evolution from Cu to Ta surface during Cu CMP has been investigated. As a first step of Cu CMP, removing excess Cu on field area to Ta often gives rise to topography issues such as dishing or erosion. The exposure of Ta is detected by in-situ reflectance measurement, where the downward slope-start signifies the beginning of Ta exposure. For blanket wafers, the maximum remaining Cu thickness at slope-start increases with initial Cu thickness. Applying lower polishing pressure and table speed, so-called soft landing, near the Cu-Ta transition stage results in thicker remaining Cu. In patterned wafers, Ta exposure starts near high pattern density areas such as wide lines, and contrary to blanket wafers, remaining Cu thickness at slope-start cannot be correlated simply with initial Cu thickness. In this experiment, with increasing initial Cu thickness, remaining Cu thickness at slope-start decreased. The amount of dishing is largely proportional to the initial Cu thickness.

More than 4 years have passed since IBM made the official announcement approving copper dual damascene (Cu2D) technology, which drew the rest of the semiconductor world into the Cu2D race. And the technology appeared to be a hard nut to crack. Nowadays, after more than three years of extensive R&D work, the EOL yields are still low, and a lot of process integration work has to be done to make it mass production ready.

It has been observed that CMP removal rate of copper varies for films from different sources. While the film hardness and static wet etch rate in the absence of inhibitors are similar for various films, the static etch rate in the presence of inhibitors is significantly different. Analysis by AFM, XPS, and ion beam sputtering showed the film roughness and surface composition were different for different Cu films within 160 angstroms of the surface. Thus the different in Cu CMP removal rates can be explained by a synergetic effect of copper film roughness (grain size), surface composition, and effective adsorption of inhibitors. This explanation suggests that if polishing is initiated by an aggressive polishing step, removal rate of the remaining bulk copper films become more consistent. Based on these findings, the polishing slurry and process were further optimized.

This work investigates the retention and transport of chemical species and abrasive particles during chemical-mechanical polishing (CMP) of copper (Cu). “Slurry step-flow” experiments, in which the concentrations of the chemicals and abrasives in the slurry are altered in steps during polishing were conducted with hydrogen peroxide (H2O2)/glycine based slurries. Two different pads, Suba-500 and IC 1400 (with k grooves), were compared in terms of their slurry retention and transport characteristics. With these experiments, it has been shown that both the abrasives and chemicals are constantly replaced during a typical CMP process. Better polishing performance of the IC 1400 over Suba 500 is a result of improved transport of the chemicals and the abrasives between the wafer/pad interface.

We present a new model for dishing and erosion during chemical-mechanical planarization. According to this model, dishing and erosion is controlled by the local pressure distribution between features on the wafer and the polishing pad. The model uses a contact mechanics analysis based on the work by Greenwood to evaluate the pressure distribution taking into account the compliance of the pad as well as its roughness. Using the model, the effects of pattern density, line width, applied down-force, selectivity, pad properties, etc. on both dishing and erosion can be readily evaluated. The model may be applied to CMP used for oxide planarization, metal damascene or shallow trench isolation.

The model is implemented as an algorithm that quickly calculates the evolution of the profile of a set of features on the wafer during the polishing process. With proper calibration of the process parameters, it can be used as a tool in optimizing the CMP process and implementing CMP design rules.

In this paper a new numerical model combining die scale and feature scale polishing characteristics for oxide CMP is proposed. This work is based on the effective density model [1-3] which takes into account a die scale pad deformation. We have extended this model to include the description of material removal at the feature scale. A pseudo topography discretization allows us to describe the polishing in down areas more accurately depending on the down area width. The physical parameters considered include the deformation of the polishing pad (die scale and feature scale), the removal rate of a blanket wafer polished under the same process conditions, and the removal rate in down areas for small polishing times.

Comparison of simulated results with MIT testreticle experiments has been performed and evaluated.

We investigate the post-CMP cleaning process with a tribological approach. A cleaning process involves three components: brush, wafer/disk, fluid undergoing three-body sliding contact between the brush, wafer, and particles from slurry. Having this in mind, we investigated cleaning mechanisms through experimental measurement of friction force and analyzed the contact condition for particle removal. Our investigation leads to the conclusions that the cleaning process is a boundary to elastohydrodynamic lubricating process that involves a constant contact between a brush and the wafer or disk surface. The motion of the brush nodule is such that the surface forces between the brush and workpiece change from an initial adhesion to sliding abrasion. These analysis leads to insight of particle removal mechanisms.

Newly designed abrasives with various particle morphology and component have been prepared and evaluated in terms of polish efficiency for CMP application. Polymer sphere with uniformed particle size and special functional group were prepared by soap-free emulsion polymerization. CMP slurry including polymer abrasives indicated characteristic properties such as dishing, erosion and low defectivity for metal and dielectric surface. This study has been also conducted on the preparation of composite particles consisting of polymer core covered with inorganic composition such as zirconium compound, titanium compound, SiO2 and Al2O3. The controlled hydrolysis and heterocoagulation systems were proposed as preparation methods of composite particles. 1st step Cu-CMP Slurry with composite particles as abrasion indicated an excellent dishing, erosion and their over-polish-margin. Particularly these composite particles were useful to prevent increasing scratch on the TEOS and low-k dielectric materials surface in 2nd step CMP process.

It was found in a Cu-CMP process using EP-C 5001 slurry and IC 1000 pad that Cu removal rate, being extremely low without H2O2 in the slurry, increases up to a maximum with the addition of H2O2, and then decreases again. Analysis of polarization curves and Eh-pH diagrams shows that without H2O2 Cu has the lowest electric potential, as a result, the highest thermodynamic stability in the Cu/slurry system. Addition of H2O2 shifts the potential up and induces the formation of Cu2O, resulting in a high removal rate. At high H2O2 concentration, a CuO passivation film is formed. In this case, only mechanical removal of the passivating oxide film allows the process to proceed. It is speculated that the moving pad surface adheres the oxidized species via the formation of hydrogen bonds with oxygen atoms of copper oxide molecules, thus detaching them from the wafer surface. Each oxygen atom is capable of pulling out two Cu atoms if Cu2O is formed on the surface and only one Cu atom if CuO is formed. This would explain why the removal rate is high at low H2O2 concentration and low at high H2O2 concentration.

pH has a strong effect on the polish rates of copper (Cu) and tantalum (Ta) [1]. In this paper, removal rates of Cu and Ta using aqueous slurries containing alumina and silica abrasives in H2O2-glycine solution are studied at varying pH values. It is observed that variation in the Cu and Ta removal rates is a direct result of the change in surface characteristics of the films. Surface characteristics such as presence/absence of a passivating layer and hardness of such layer vary with pH and hence result in removal rate variation. It is also shown that a favorable Cu/Ta polish rate selectivity can be obtained by adjusting the pH of the slurry.

At present, several different competing mechanical configurations are used in chemical mechanical polishing. These range from classical rotary designs to orbital and linear systems, as well as the more recent web format. Invariably, wafer rotation is used in all of these systems to average out gross material removal rates and improve within wafer thickness uniformity. Although this might appear to be trivial and unworthy of serious investigation, in reality the interaction of polishing pad surface structure, along with whatever abrasive particles and chemical agents that might be present is quite complex. In addition, fluid transport between wafer and pad surfaces is strongly affected by pad macrostructure and relative motion of the two surfaces.

We discuss applications of optoacoustic film thickness metrology for characterization of copper chemical-mechanical polishing (CMP). We highlight areas where the use of optoacoustics for CMP characterization provides data complementary to that obtained by other techniques because of its ability to directly measure film thickness with high spatial resolution in a rapid, non-destructive manner. Examples considered include determination of planarization length, measurement of film thickness at intermediate stages of polish, and measurement of arrays of metal lines.

A long-standing CMP related yield/reliability puzzle - film thickness spike, frequently appearing at the wafer center after it is processed by CMP, was studied. Optical film thickness measurements for pre- and post-CMP film were compared with surface topography profiles. Wafer surface profiles were also studied after the oxide film was HF-etched out. It was found that the post-CMP film thickness spike actually is not caused by CMP malfunctioning, but originates to a specific bare silicon wafer defect and occurs when the bare silicon wafer has a narrow dip-like defect (depression) in its center. The dielectric film, deposited over the wafer, follows the underlying center-dipped topography, and, being planarized by CMP, is transformed into a top-flat film with a thick pattern in the wafer center, filling the silicon dip. In-line and end-of-line yield related issues are discussed.

Shallow trench isolation (STI) requires a high quality oxide with superior fill capability provided by High Density Plasma (HDP) oxide. Unfortunately, the HDP deposition process can create large within die topographies that are difficult to polish directly using conventional silica slurries. As a result, etch back integration schemes have traditionally been incorporated for STI polish. A more revolutionary approach is the use of Fixed Abrasive (FA) CMP [1]. FA CMP allows direct STI polish with good planarization/process stability, eliminating the need for prior etch back. The planarization efficiency is strongly dependent on the shape of the pad composites that hold the CeO2 mineral. Fixed abrasive pads with pyramid and pole shapes are available. In this work, three different fixed abrasive pads supplied by 3M corporation were evaluated for STI CMP polish performance using the Obsidian 8200C web format CMP tool. Basic polish characteristics such as planarity (dependence on sub-pad/pattern density), selectivity to topography, oxide dishing and nitride erosion are presented. The FA pads discussed here have been classified as “slow”, “medium” or “fast” depending on blanket oxide removal rate.

The slow rate pad had a very high selectivity to topography and very low dishing of the down area oxide. The removal rate of blanket oxide was less than 100 Å/min. The pad was best suited to the polish of isolation trench structures with small, controlled overfill (> 200 Å) across the wafer. A large process window was demonstrated. The removal rate of the “medium” pad also decreased significantly at the onset of planarization with a blanket oxide removal rate of ca. 200 Å/min. Unlike the slow rate pad, the medium rate pad did not provide a suitable overpolish process window required for a manufacturable STI process. It is believed this pad would be a good choice for BPSG polish. In contrast to the slow and medium rate pads, the blanket oxide removal rate of the fast pad was ca. 2000 Å/min with no self-stopping capability at the onset of planarization. The removal rate was extremely center fast, such that it could not be compensated by adjustment of tool parameters. Use of a modified process developed within the DRAM development alliance (DDA) at East Fishkill (IBM/ Infineon) enabled the fast pad to polish deep STI structures that would otherwise be impossible using the slow or medium rate pads.

Wehaveusedalargepatterntestmaskandaspecificarrangementofstructuresonawaferfor direct measurement of an average planarization length for copper chemical mechanical polishing (CMP)processes.Weproposenewminimum,maximum,andaverageplanarizationlengthdefini-tions, based on up and down area measurements as a function of trench width. The average pla-narizationlengthisusefulforqualitativelycomparingtheplanarizationcapabilityofcopperCMP processes. We have also performed several experiments that show how the average planarization length depends on polish process settings such as down force and relative speed, as well as on consumables such as pad and slurry.

Model alumina slurries for chemical-mechanical planarization (CMP) of copper have been studied using fundamental electrochemical concepts together with modification of surface tension and zeta-potential. A model slurry was established containing 3 wt% alumina (50 nm nominal size), 2 wt% potassium dichromate (K2Cr2O7) and 1 vol % DOWFAX, as abrasive, oxidizer and anionic surfactant, respectively, which resulted in blanket copper removal rates of 130 nm/min with smooth, low-particulate defect density surfaces without aggressive post-CMP cleaning. When used with a model silica-abrasive slurry to remove the Ta liner, well-defined damascene-patterned structures were achieved with low-particulate defect densities.

Open circuit potential measurement, potentiodynamic polarization, surface tension measurement, and zeta-potential measurement were used in developing the model slurries. The variations obtained with alternative abrasives, alternative oxidizers and alternative surfactants illustrate the principles of synthesizing slurries for Cu and Ta.