Search

We extend our recent 2D trajectory (x–y plane) and diffusion coefficient data of ceria particles near a glass surface obtained at pH 3, 5, and 7 using evanescent wave microscopy and video imaging to 3D trajectories by analyzing the separation distance between the particles and the glass surface in the vertical z‐direction. Mean squared displacement (MSD3D) of ceria particles was calculated to quantify 3D trajectories. Three‐dimensional diffusion coefficients were obtained from the MSD3D curves and were compared with two‐dimensional diffusion coefficients. By analyzing the MSD curves, we found that ceria particles exhibited only confined motion at pH 3 and 5, while both confined and Brownian motion were showed at pH 7. We also evaluated the cleaning ability of DI water adjusted to pH 10 and 12 to remove ceria particles from glass surfaces and related the results to the calculated trajectory, diffusion coefficient, and interaction potential data.

Using evanescent wave (EW)–based optical detection methods coupled with video microscopy, we investigated in situ trajectories, diffusion, and interaction energies of ∼140 nm ceria particles near a glass surface at pH 3, 5, and 7. Trajectories of a single ceria particle in a 2D (x–y) plane were obtained by linking its time-sequenced positions. Diffusion coefficients of several single ceria particles were calculated from their respective mean-square displacement (MSD) versus time curves, and the results were interpreted based on the interaction potential energy curves obtained from Boltzmann statistics of the EW scattering intensity fluctuations of the particles. The types and characteristics of particle motions were determined by analyzing the MSD curves. Whereas both confined or subdiffusive and Brownian motions of the particles were observed at pH 7, only confined motion was seen at pH 3 and 5, and their corresponding diffusion coefficients are similar to those reported by several authors.

Chemical mechanical planarization (CMP) is a widely used technique for the manufacturing of integrated circuit chips in the semiconductor industry. The process generates large amounts of waste containing engineered particles, chemical additives, and chemo-mechanically removed compounds. The environmental and health effects associated with the release of CMP materials are largely unknown and have recently become of significant concern. Using a zebrafish embryo assay, we established toxicity profiles of individual CMP particle abrasives (SiO2 and CeO2), chemical additives (hydrogen peroxide, proline, glycine, nicotinic acid, and benzotriazole), as well as three model representative slurries and their resulting waste. These materials were characterized before and after use in a typical CMP process in order to assess changes that may affect their toxicological profile and alter their surface chemistry due to polishing. Toxicity outcome in zebrafish is discussed in relation with the physicochemical characteristics of the abrasive particles and with the type and concentration profile of the slurry components pre and post-polishing, as well as the interactions between particle abrasives and additives. This work provides toxicological information of realistic CMP slurries and their polishing waste, and can be used as a guideline to predict the impact of these materials in the environment.

Temperature is one of the parameters that needs to be continuously monitored and controlled during handling, storage, and transportation to achieve repeatable and consistent CMP performances. Here, we investigated the effect of three different storage temperatures (15, 25, and 45oC) on several slurry health parameters (particle size, zeta potential, pH, total dissolved solids, conductivity, and dissolved oxygen (DO) concentration) of a ceria-based slurry and also on removal rates of SiO2 films during polishing. The changes in all the parameters of slurries stored at 15oC and 45oC for three weeks, except for DO concentration, were reversible and returned to their values at 25oC when the slurries were brought back to 25oC. DO concentration increased by ~14 % and decreased by ~18 % in the slurries stored for three weeks at 15oC and 45oC, respectively, compared to that of ceria slurry stored at 25oC. They did not return to the original values even after keeping them at 25oC for 6 hours. The increase in storage temperature causes an increase in the adsorption of dissolved oxygen onto the ceria surface, which can transform Ce3+ species into less reactive Ce4+-superoxo or a Ce3+-peroxo species, resulting in lower SiO2 removal rates.

Brush scrubbing is commonly employed for cleaning contaminated polished wafers, especially after chemical mechanical polishing. Here we report the results from real-time video imaging of the brush cleaning of ∼90 nm ceria particles from thin oxide films on transparent glass substrates using evanescent wave microscopy to identify the interactions among the particles, brush, film and cleaning liquid. Two cleaning liquids, DI water (pH ∼ 6) and 0.1 M NH4OH solution at pH ∼ 11, were used. It was found that purely hydrodynamic forces are unable to remove the ceria particles from the oxide film surface and direct contact between brush asperities and film is crucial for particle detachment to occur. However, such a direct contact also causes significant particle loading of the brush as well as redeposition of some of the already dislodged particles. Preventing such redeposition is crucial for efficient surface cleaning. The dislodged particles that are transferred into the fluid flow on the film continued to move without redepositing. Since the 0.1 M NH4OH solution at high alkaline pH created an environment for charge repulsion among the ceria particles, PVA brush and oxide film, particle removal from the film is more rapid and somewhat more efficient compared to DI water.

A combined experimental and computational study of fibrous filters for removal of large hydrosols, essential to minimize defects during chemical mechanical polishing, was performed. Dilute aqueous suspensions of colloidal ceria particles, of known size distribution, were filtered and the filter efficiencies were measured for different particle sizes and pH, then converted to single fiber efficiencies. The particle size distributions were also measured for the influent and effluent streams. In a series of numerical simulations, the Navier-Stokes equation was solved for a single fiber using the ANSYS-FLUENT computational fluid dynamics commercial package. For dilute suspensions, the motion of the dispersed particles in the size range of 35–600 nm was tracked in the Lagrangian reference frame including the effects of hydrodynamic drag, lift, gravity, hydrodynamic retardation, Brownian, van der Waals and electric double layer forces. The electric double layer and van der Waals forces were incorporated in the calculations by developing a user defined function. Particular attention was given to the effect of Brownian excitations, as well as the electric double layer and van der Waals forces - which have been neglected in many of the previous models - on the overall fiber collection efficiency for different particle sizes and charges. The simulated fiber efficiencies and particle size distributions compared very well with experimental results.

Development of systems for capture, sequestration, and tracking of nanoparticles (NPs) is becoming a significant focus in many aspects of nanotechnology and environmental research. These systems enable a broad range of applications for evaluating concentration, distribution, and effects of NPs for environmental, clinical, epidemiological, and occupational exposure studies. Herein, we describe the first example of a ligand-graft multifunctional platform for capture and detection of cerium oxide (CeO2 or ceria) NPs. The approach involves the use of redox-active ligands containing o-dihydroxy functionality, enabling multivalent binding, surface retention, and formation of charge transfer complexes between the grafted ligand and the NPs. Using this strategy, paper-based and microarray-printed platforms with NP-capture ability involving either catechol or ascorbic acid as ligands were successfully fabricated. Surface modification was determined by infrared spectroscopy, electron microscopy, X-ray spectroscopy,...

We demonstrated that the movement of the single ceria particles located at the glass interface could be imaged in situ using evanescent wave (EW) microscopy. (1,2) Here, the flat glass surface of the lens became our model for the SiO2 film surface of interest. 3D trajectories, diffusion coefficients, and interactions of ceria particles with a glass surface at pH 3, 5, and 7 were investigated using EW microscopy. We also defined the types and characteristics of particle motions by analyzing the mean-square displacement (MSD) versus time curves. (1) The removal of single ceria particles from the glass films in the cleaning solution was monitored in situ using EW microscopy coupled with a brush set-up. In situ measurement data collected during cleaning are likely to provide more important insights into the underlying mechanisms of various complex reactions that occur at the particles/brush–wafer interface. (3) This talk will also introduce lateral force microscopy (LFM) mode in the atomic force microscopy (AFM) to obtain the removal force of single ceria particles from the SiO2 film. LFM is a powerful tool to monitor the torsion bending (or twisting) of the cantilever with contact mode, which is corresponding to the removal force. Types of particle removal mechanism depending on the characterization of single particles (size, shape, and surface) will be discussed. Our results will provide a better understanding of how the abrasive particles, the additives, and the film surfaces interact together during cleaning. (1) Seo, J., Gowda, A., Khajornrungruang, P., Hamada, S., Song, T., & Babu, S. (2020). Trajectories, diffusion, and interactions of single ceria particles on a glass surface observed by evanescent wave microscopy. Journal of Materials Research, 35(3), 321-331. (2) Seo, J., Gowda, A., Khajornrungruang, P., Hamada, S., & Babu, S. (2020). 3D trajectories and diffusion of single ceria particles near a glass surface and their removal. Journal of Materials Research, 1-10. (3) Seo, J. (2020). A review on chemical and mechanical phenomena at the wafer interface during chemical mechanical planarization. Journal of Materials Research, 1-23.

It has been very difficult to remove smaller ceria particles from oxide surfaces after polishing. As the particle size decreases from 90 to 10 nm, cleaning efficiency of SC1 solution decreases from ~94 % to ~19 %, presumably because smaller ceria particles with a higher surface concentration of Ce3+ are more strongly coupled with silicon dioxide surfaces via strong Ce-O-Si bonding. (1) As was reported earlier, where the cleaning of oxide surfaces of ceria particles from slurries containing no additives, we reported that equimolar solutions containing H2O2 and NH4OH (4.2 mol l−1 each of H2O2 and NH4OH), aided by ultrasonic cleaning, and at high pH could almost completely remove even 10 nm sized ceria particles from silicon dioxide film surfaces. (1) However, a cleaning solution containing H2O2 and NH4OH at equal molar (4.2 mol l−1) was not effective in cleaning oxide and nitride surfaces contaminated with ceria particles in the presence of additives such as proline and citric acid. (2) Recently, we reported a cleaning solution comprising 1 wt% ascorbic acid, 1 wt% ammonium carbonate and 50 ppm Triton X-100 at pH 12, aided by ultransonic energy, to clean ~30 nm ceria particles from both oxide and nitride films with ~ 99% cleaning efficiencies. The binding between the additive-covered ceria particle particles and the oxide/nitride surfaces could be broken by the nucleophilic attack of hydroxyl ions. Ascorbic acid and ammonium carbonate prevent ceria particle redeposition by adsorbing on the surface of the removed particles and blocking the active Ce3+ species. Triton X-100 reduces adhesion between the particle and the film surfaces during cleaning. (1) Seo, J., Gowda, A., & Babu, S. V. (2018). Almost complete removal of ceria particles down to 10 nm size from silicon dioxide surfaces. ECS Journal of Solid State Science and Technology, 7(5), P243. (2) Gowda, A., Seo, J., Ranaweera, C. K., & Babu, S. V. (2020). Cleaning Solutions for Removal of∼ 30 nm Ceria Particles from Proline and Citric Acid Containing Slurries Deposited on Silicon Dioxide and Silicon Nitride Surfaces. ECS Journal of Solid State Science and Technology, 9(4), 044013.

Chemical mechanical polishing (CMP) behavior of Cu/Ru/TiN barrier liner stack was investigated with a slurry comprising of silica abrasives, potassium permanganate (KMnO4), guanidine carbonate (GC) and benzotriazole (BTA) in the alkaline region. The corrosion and polishing behavior of the Cu, Ru and TiN films in the solution consisting of the above additives were characterized by open circuit potential and potentiodynamic measurements, polishing rates, dissolution rates and contact angle measurements. A slurry comprising of 10 mM KMnO4, 1 wt% GC and 5 wt% Silica at pH 10 has shown adequate polish rates as well as low )

A previously developed aqueous cleaning solution (4.2 mol l−1 each of H2O2 and NH4OH) was found to be ineffective in cleaning oxide/nitride surfaces after contamination with ceria particles from slurries containing proline or citric acid. However, a cleaning solution consisting of 1 wt% ascorbic acid, 1 wt% ammonium carbonate and 50 ppm triton X-100 at pH 12, aided by ultrasonic cleaning, removed these ceria particles, even those as small as ∼30 nm, from both oxide and nitride surfaces with efficiencies >99% as determined by AFM imaging. Fourier transform infrared (FTIR) spectroscopy results indicated that ceria particles treated with these additives can also bind with oxide/nitride surfaces through Si–O–C and Si–O–H bonds, in addition to any Ce–O–Si, where the C and H atoms are from the additives adsorbed on the ceria particles. All these bonds are broken effectively by the nucleophilic attack of hydroxyl anions in the cleaning solution while triton X-100 in the cleaning solution reduces adhesion between the particles and the film surface and facilitates cleaning via a wetting mechanism. More importantly, ascorbic acid and ammonium carbonate prevent particle redeposition by complexing with the removed particles and blocking the active Ce3+ species on their surface.

Following our earlier analysis of amino acids, we investigated the effectiveness of acetic acid, pyridine, and sorbitol that were chosen to represent carboxylic acid, amine and alcohol functional group families, respectively, as additives in ceria dispersions for polishing SiO2 and Si3N4 films. By comparing the speciation of the additives available and the material removal rates with respect to pH, we identified that the removal rates of nitride, but not oxide, are suppressed to

A colloidal AFM-based method was used to understand the role of diallyldimethyl ammonium chloride (DADMAC) molecules in non-Prestonian silicon dioxide removal obtained using ceria-based slurries. A series of force-distance measurements between silicon dioxide films, polyurethane IC1000 polishing pad and ceria abrasives in liquid media with and without DADMAC molecules were made. It was shown that at 15 mM concentration, DADMAC molecules yield non-Prestonian behavior by interfering between ceria particles and the oxide film. The effects of the DADMAC additive on the removal rates of the oxide films and their corresponding coefficients of friction (COFs) values were also investigated. Finally, the correlation between AFM force measurements, removal rates and COF values measured during polishing are discussed. Interestingly, below a threshold COF value of ∼0.23 to 0.24, corresponding to a threshold down pressure of ∼2psi for both 15 mM and 25 mM DADMAC concentration, no oxide film removal was observed. © The Author(s) 2015. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0031511jss] All rights reserved.

Real time visualization of the in situ growth of ∼40 μm long high aspect ratio silver ribbons, synthesized by template-free galvanic displacement using ∼100 nm Cu particles and Ag ions in aqueous conditions, is demonstrated using evanescent wave microscopy. The observed growth rate from several ribbons is 1.5 ± 0.5 μm s−1.

A mathematical model, a continuation of the work by Zhao et al. [1] that describes particle trajectories during chemical mechanical polishing, was extended to account for the effects of larger particles, particle location changes due to slurry dispensing and in-situ conditioning. Material removal rate (MRR) and within wafer non-uniformity (WIWNU) were determined based on the calculated particle trajectory densities in the absence of chemical activity from the slurry. The effect of pad-wafer rotary dynamics and reciprocating motion of the wafer carrier on MRR uniformity were included. It was also shown that in-situ conditioning improves the MRR uniformity of the polished wafers. Using the model, we also investigated the effect of particle size distribution and large particles (>200 nm in diameter) on WIWNU and scratch growth. It was shown that the presence of even 1 wt.% of larger particles can deteriorate the WIWNU.

A slurry with a non-Prestonian dependence on the polishing pressure can help in minimizing dishing and erosion during shallow trench isolation chemical mechanical planarization. Here, we show that ceria-based slurries containing diallyldimethylammonium chloride (DADMAC) yield a non-Prestonian blanket film polish rate with a low threshold pressure (1–2 psi) when polishing plasma-enhanced chemical vapor (PECVD) tetraethylorthosilicate (TEOS) deposited oxide as well as thermal oxide films. The polishing mechanism of this non-Prestonian slurry was investigated by a series of experiments involving zeta potential measurements, thermogravimetric analysis (TGA) and UV-vis spectroscopy and it was shown that more DADMAC molecules are adsorbed on silica particles (as oxide film representatives) than on ceria particles and the binding strength between DADMAC and silica is much higher than that with ceria surface. © The Author(s) 2014. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0091502jss] All rights reserved.

The fabrication of an InGaAs n-channel in complementary metal oxide semiconductors (CMOS) requires an intervening layer of Indium Phosphide (InP) that should be planarized using chemical mechanical planarization (CMP) before deposition of the InGaAs layer. The InP CMP process has to provide good InP removal rates (RRs) (100-200nmmin^-^1), very low oxide loss, low dishing and minimal or no phosphine evolution. In this work, it is shown that aqueous slurries containing 3wt% silica, 1wt% hydrogen peroxide and 0.08M oxalic acid or citric acid satisfy these requirements. Cross-sectional scanning electron microscope (SEM) images showed that InP in the shallow trench isolation structures was planarized and scratches, slurry particles and smearing of InP on the wafers were absent. Additionally, wafers polished at pH 6 showed very low dishing values of about 12-15nm, determined by cross-sectional SEM.

Silicon carbide thin films, though of widespread use in microelectronic engineering, are difficult to process by chemical mechanical polishing (CMP) because of their hardness and chemical inertness. This report discusses the development of slurries based on silica abrasives that resulted in high amorphous SiC (a-SiC) removal rates (RRs). The ionic strength of the silica dispersion was found to play a significant role in enhancing material removal rate, while also providing very good post-polish surface-smoothness. For example, the addition of 50 mM potassium nitrate to a pH-8 aqueous slurry consisting of 10 wt% of silica abrasives and 1.47 M hydrogen peroxide increased the RR from about 150 nm h −1 to about 2100 nm h −1 . The role of ionic strength in obtaining such high RRs was investigated using surface ζ-potentials measurements and X-ray photoelectron spectroscopy (XPS). Evidently, H 2 O 2 promoted the oxidation of Si and C to form weakly adhered species that were subsequently removed by the abrasive action of the silica particles. The effect of KNO 3 in increasing material removal is attributed to the reduction in the electrostatic repulsion between the abrasive particles and the SiC surface because of screening of surface charges by the added electrolyte.

This work describes the response of Ru/TiN barrier films to solutions of H2O2, guanidine carbonate (GC) and other additives in terms of open circuit potential measurements and potentiodynamic polarization data. It was found that an aqueous solution containing 1wt% H2O2, 0.25 wt% GC and 5 mM BTA is effective in maintaining a low corrosion potential gap of Ru and Cu at ∼15 mV at pH 9. The slurry prepared by adding 5 wt% silica to this mixture yields a Ru removal rate (RR) of 10 nm/min and a Cu RR of 12 nm/min at 2 psi polishing pressure. The effect of carbonate and guanidinium ions as well as BTA on the corrosion behavior of Ru and Cu is discussed. Dodecyl-benzene-sulfonic acid (DBSA) was also found to be an effective corrosion inhibitor of Ru in the presence of H2O2 and GC based solutions at pH 9 with a good post-polished surface finish but reduced the Ru RRs to ∼ 2 nm/min. © The Author(s) 2014. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND, http://creativecommons.org/licenses/by-nc-nd/4.0/)

We show here that transition metal compounds when used as additives to silica dispersions enhance a-SiC removal rates (RRs). Silica slurries containing KMnO4 gave RRs as high as 2000 nm h−1 at pH 4. Addition of CuSO4 to this slurry further enhanced the RRs to ∼3500 nm h−1 at pH 6. Furthermore, addition of a low concentration of 250 ppm Brij-35 to this slurry suppressed the RRs of SiO2 to zero, while retaining the RRs of a-SiC at ∼2700 nm h−1, a combination of RRs that is appropriate for hard mask polishing. The underlying mechanisms causing the enhancement in the RRs of a-SiC in the presence of transition metal compounds is discussed based on the RR data and XPS analysis of post-polished wafer surfaces. It is shown that a mixed redox system consisting of the oxidation of a-SiC by KMnO4 enhanced by the catalytic activity of the Cu(II) salt is responsible for the rapid oxidation of a-SiC and the observed enhancement in polish rate with silica based slurries. The adsorption characteristics of Brij-35 on the oxide and carbide surfaces are described based on thermogravimetry data and in achieving the desired polish rate selectivity. © The Author(s) 2014. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND, http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is not changed in any way and is properly cited. For permission for commercial reuse, please email: oa@electrochem.org. [DOI: 10.1149/2.021406jss] All rights reserved.

Four different anionic surfactants, sodium dodecyl sulfate, dodecyl benzene sulfonic acid (DBSA), dodecyl phosphate and Sodium lauroyl sarcosine, selected from the sulfate, phosphate, and carboxylic family, were investigated as additives in silica dispersions for selective polishing of silicon dioxide over silicon nitride films. We found that all these anionic surfactants suppress the nitride removal rates (RR) for pH ≤4 while more or less maintaining the oxide RRs, resulting in high oxide-to-nitride RR selectivity. The RR data obtained as a function of pH were explained based on pH dependent distributions of surfactant species, change in the zeta potentials of oxide and nitride surfaces, and thermogravimetric data. It appears that the negatively charged surfactant species preferentially adsorb on the positively charged nitride surface below IEP through its electrostatic interactions and form a bilayer adsorption, resulting in the suppression of nitride RRs. In contrast to the surfactants, K2SO4 interacts only weakly with the nitride surface and hence cannot suppress its RR.

During the chemical mechanical polishing (CMP) of an indium phosphide buffer layer for the fabrication of InGaAs n-channels in CMOS devices, there is interest in controlled removal of In and P atom...

Picolinic acid, nicotinic acid, proline, and γ-amino butyric acid were investigated as additives in ceria dispersions for polishing of SiO2 and Si3N4 films. Removal rates (RRs) data obtained as a function of pH were correlated with pH dependent distributions of amino acid species. We discovered that species containing protonated amino group, when available in sufficient amounts, can suppress the nitride RRs. Hydrogen bonding drives the adsorption of these species on the oxide and nitride surfaces. The adsorption of two of the amino acids, picolinic acid and proline, on SiO2 and Si3N4 surfaces was investigated using thermogravimetry and differential thermogravimetric measurements which identified the formation of peptides activated by the hydrogen bonding. Protonated amino groups act as strong hydrogen bond donors and can form strong hydrogen bonds with hydrogen bond acceptor sites on the Si3N4 surface while the corresponding bonding on SiO2 surface sites is weak. The stronger hydrogen bonding hinders Si3N4 hydrolysis and effectively suppresses Si3N4 removal. In contrast, the weakly bound additive species are easily removed from the SiO2 surface by the polishing pad and ceria abrasives and hence the SiO2 removal is not affected significantly. This leads to high SiO2 to Si3N4 removal rate selectivity.

We describe the effectiveness of depositing an amorphous silicon thin film to cover the existing defects on the surface of extreme ultraviolet (EUV) lithography mask substrates and then polishing the film using abrasive-free liquids. Preliminary investigations on a-silicon coated Si wafers showed that the surface roughness and removal rates (RRs) are largely dependent on the polishing pressure. Higher polishing pressures led to higher removal rates and surface roughness while a lower polishing pressure of 0.5 psi led to negligible RRs but achieved a surface roughness of ~0.1 nm RMS (root-mean square). A two-step CMP process, consisting of polishing at 1 psi followed by polishing at 0.5 psi on an IC-1000 pad gave acceptable surface roughness (~0.1 nm) as well as adequate material removal. The same polishing conditions when transferred to a-silicon films deposited on 6”x6”x0.25” quartz EUV substrates also resulted in a RMS surface roughness of ~0.1 nm.

The fabrication of InGaAs n-channels in CMOS devices requires an intervening buffer layer of GaAs or InP between a Si substrate and the InGaAs channel layer, which has to be planarized to enable the growth of a channel that is free of defects due to uneven surface topography. This report discusses the material removal rates (RRs) and AsH3 evolution observed during chemical mechanical polishing (CMP) of GaAs using aq. solutions of hydrogen peroxide (H2O2) and silica slurries containing H2O2. GaAs RRs were negligible with deionized water or with silica slurries alone. They were relatively high in aq. solutions of H2O2 alone and showed a strong pH-dependence, with significantly higher RRs in the alkaline region. The addition of silica particles to aq. H2O2 did not increase the GaAs RRs significantly. The evolution of arsenic trihydride (AsH3) during the dissolution of GaAs in aq. H2O2 solution was similarly higher in the basic pH range than in neutral pH or in the acidic pH range. However, no AsH3 was measured during polishing, evidently because of the relatively high water solubility of AsH3. In this report, GaAs removal rates and AsH3 formation are analyzed using data from contact angle measurements, X-Ray photoelectron spectroscopy (XPS), and X-Ray fluorescence (XRF) spectroscopy. Reaction pathways leading to material removal are proposed.

Ruthenium (Ru) films deposited on either TiN or TaN/Ta have been proposed as a barrier stack in advanced interconnects. Here, we investigated their polishing behavior using colloidal silica-based slurries containing guanidine carbonate (GC) or hydrogen peroxide (H2O2) or both. Neither GC nor H2O2 alone enhanced the Ru removal rates (RRs) but their combination did, presumably, due to the formation of Ru oxide-guanidinium complexes which can be polished by silica abrasives suggesting that the oxidation of Ru to its oxides is a crucial first step. Ethylenediamine and 2, 2‐bipyridine that were reported to form complexes with halides of Ru in various oxidation states also enhanced the RRs of Ru films, similar to GC. Furthermore, even though both types of Ru films were deposited at same conditions, RRs of Ru on TiN were enhanced more compared to those on TaN/Ta, likely a consequence of the difference in the crystalline structure of the oxide films formed due to a difference in the structure of the Ru films themselves. Using X-ray diffraction, X-ray photoelectron spectroscopy, nanoindentation, zeta potential measurements, thermo gravimetric analysis and contact angle measurements, the role of GC and crystalline structure in enhancing the RRs of the films is discussed.

In the replacement metal gate (RMG) scheme 1 there are several challenges such as the conformal deposition of metal layers in the gate stack, precise gate height control and adequate removal rate selectivity, minimum galvanic corrosion during metal gate chemical mechanical planarization (CMP), etc. During the RMG process, once the polysilicon dummy gate has been etched away, work function metal films are deposited followed by subsequent patterning steps and deposition of a suitable metal layer into the trench vacated by poly-Si. The overburden is then removed by CMP. Aluminum, tungsten and TiN 2 are some of the commonly used gate materials. Recently Kwon et al. 3 proposed a novel Co/Al-based metal fill scheme for the RMG process that can be scaled down to 11 nm. In this scheme, Co would replace Ti as the wetting layer for better reflow and fill of Al into the trench vacated after the dummy gate etch devices. After annealing at 400 ◦ C it forms a thin Al/Co alloy layer (presumably CoAl/Co2Al9. 12 ) However, with these two metals (Al and Co) in contact with each other, Al will galvanically corrode in a CMP environment due to its lower standard electrode potential (E o ) compared to that of Co. Shi et al. 4 investigated galvanic corrosion between Al and Co in ascorbic acid-based chemistry at pH 4 and the lowest corrosion potential (Ecorr) gap they could achieve was around a still very high 0.35 V. In contrast, we recently identified a potas

One of challenge associated with a self-forming Mn barrier for Cu interconnects in sub-22 nm devices is galvanic corrosion that can occur at the Cu-Mn interface during chemical mechanical planarization. In the present work, it is shown that an aqueous solution of sucrose, BTA and KIO4 is able to reduce the corrosion potential gap between Cu and Mn to ∼0.01 V at pH 10 and lower the galvanic currents significantly. We discuss the role of the additives and the inhibiting film that can be formed at the interface of the bimetallic system in this solution. Polishing results for Cu and Mn using a silica-based slurry formulated with this solution are also presented.

This work addresses the corrosion and galvanic corrosion challenges that can arise from the use of various novel barrier materials Ru/TiN, Co and Mn during barrier CMP in the back end of the line (BEOL) interconnects for advanced nodes.

Using a well-known galvanic displacement reaction, ∼25-40 μm long silver ribbons grown after mixing ∼50 nm copper particles with AgNO3 solution were observed as a function of Ag(+) concentration and their growth was characterized in real-time and in situ by evanescent wave (EW) microscopy. At low Ag(+) concentration, chain-like structures consisting of both Ag and Cu were observed. When the sequence of mixing these two reactants was reversed, different Ag microstructures (platelets and dendrites) were formed and were also characterized by EW microscopy. Dependence of the morphology of all these microstructures on silver ion concentration was determined by EW microscopy in conjunction with scanning and transmission electron microscopy.

Germanium (Ge) shallow trench isolation (STI) structures were polished with two different slurries: (A) 12wt.% fumed silica particles +0.5wt.% H"2O"2 at pH 2 and (B) 3wt.% colloidal silica particles +1wt.% H"2O"2+0.2M KNO"3+0.1mM cetyl trimethyl ammonium bromide at pH 8, as both were reported earlier to be good candidates to achieve planarization. Two dimensional before and after polish profiles of 160nm wide Ge STI structures were used to characterize planarization performance. A two step CMP process, where the Ge overburden was first polished with slurry A followed by polishing with fumed silica particles at pH 11, resulted in the desired slightly recessed oxide but also a high root mean square roughness (RMS) surface roughness value (~0.7nm) and some oxide loss. Planarization with a dishing value of ~15nm and an RMS value of ~0.3nm was achieved when polished with slurry A in a single step for 37s, which also showed self-limiting planarization behavior. The dishing may be reduced with tight end point control. Slurry B produced higher dishing values.

In the copper CMP process, organic residues that are related to Benzotriazole (BTA) adsorbed on copper surface after Cu CMP process have to be removed during the cleaning. In order to address this organic defect issue, we present here the study of the performance of BTA removal by post-CMP cleaners using copper particles as substrates instead of copper wafers. In this work, different copper particles including Cu(0), Cu(I) and Cu(II) oxide particles with high surface area, were chosen to study the removal of BTA adsorbed on different copper states. TGA and UV-Vis spectra were used to detect and quantify the BTA removal efficiency. The results by different post CMP cleaners on various Cu particles will be presented.

Results of a detailed study of different parameters of the polishing process and their influence on the removal rates (RRs) of InP are presented. RRs with silica-based slurries as well as generation of PH3 were much higher in the acidic pH regions compared to those in basic pH regions. RRs were also higher than those obtained with harder alumina particles. Smaller silica particles led to higher RRs than larger particles at the same wt% loading, presumably due to the higher total particle surface area. These characteristics suggest an affinity between silica particles and InP similar to that observed earlier between Ta and silica. However, unlike with Ta, replacing the surface silanol groups of the silica particles with dimethyl silane did not suppress the RRs completely. It is proposed that InP removal by silica abrasives is facilitated by the formation of ‐Si‐O‐In‐, ‐Si‐In‐ and ‐Si‐P‐ structures due not only to the reactivity of the silanol groups but also due to the combined effect of the electronegativity of Si atoms, the presence of lone pairs of electrons on In atoms on the InP surface and the electronegativity difference between Si and P.

The pH-dependent interactions of five aqueous abrasive-free polycationic solutions, all at a concentration of 250 ppm, with poly-Si, SiO(2), and Si(3)N(4) films and IC1000 polishing pads used in chemical mechanical polishing have been investigated and compared with the interaction of poly(diallyldimethyl ammonium chloride) (PDADMAC) that was investigated recently. Three of the polycationic solutions, poly(dimethylamine-co-epichlorohydrin-co-ethylenediamine), poly(allylamine), and poly(ethylene imine) (PEI) enhance poly-Si removal rates (RRs) to the range of about 500 to 600 nm/min at pH 10. In contrast, poly(acrylamide) (PAA) suppressed poly-Si RRs to about 50 nm/min, whereas with a copolymer of PAA and PDADMAC, the RRs were lower than those obtained with PDADMAC but higher than those obtained with PAA. For all the polycationic solutions, the RRs of both SiO(2), and Si(3)N(4) films were ~0 nm/min. These solutions offer a low-defect option for the processing of emerging FinFET devices. The variation in the RR magnitude and dependence on pH among the different polycations is related to the relative charge density of the polycations as well as the films being polished, consistent with ζ potential data. Based on the ζ potential data and earlier published reports, it is suggested that the strong polycation-mediated bridging interactions between the polarized and weakened Si-Si bonds of the poly-Si surface and the polyurethane IC 1000 pad are responsible for the high poly-Si RRs.

The dramatic differences in the effects of two different polishing pads (IC1000 and Politex) on the removal rates (RRs) of poly-Si, SiO 2 , and Si 3 N 4 films during chemical mechanical polishing using aqueous abrasive-free solutions of a cationic polymer, poly(diallyldimethyl ammonium chloride) (PDADMAC), are described. For example, with a 250 ppm of aqueous PDADMAC solution, poly-Si RR is 2 and Si 3 N 4 films were suppressed to less than 1 nm/min for both pads. Possible mechanisms for the observed differences in the RRs of poly-Si, SiO 2 , and Si 3 N 4 on the two pads are discussed based on elemental analysis of the pads, X-ray photoelectron spectroscopy, and zeta potential measurements (of the pads in the absence and presence of PDADMAC). Also, the results of exposure to eosin Y, a dye that changes color on exposure to PDADMAC provided useful insights into the strength of the bridging attraction between the pads and PDADMAC.

The effects of colloidal silica particles, oxidizer (H2O2), and pH on the removal rates (RRs) of Ge have been evaluated. High removal and dissolution rates were obtained in the basic pH regions, likely caused by *OH radicals and rapid dissolution of the dissociation products of germanium hydroxides, while at pH 2, a RR of ∼420 nm/min was obtained with minimal dissolution. The surface quality of polished wafer coupons, measured using atomic force microscopy, was very good. A probable reaction mechanism for Ge removal as a function of pH is proposed.