Chemical Mechanical Planarization Slurry Research Articles

Next generation barrier CMP slurry with novel weakly alkaline chelating agent

To strengthen the device performance with the pattern wafer by enhancing the Cu polishing rate and improve the surface roughness with the Cu lines, a new weakly alkaline chelating agent with a barrier slurry is developed to meet the process demand of the advanced barrier chemical mechanical planarization (CMP). This new chelating agent has a stronger chelating ability and a lower pH value than the previous generation-FA/O I chelating agent researched before. Without an unstable oxidant agent added in the polishing slurry, it is difficult to enhance the copper polishing rate during the barrier CMP. The stronger chelating ability of the new chelating agent could increase the copper polishing rate along with controlling the Cu/Ta/TEOS removal rate selectivity to meet the requirements of the IC fabrication process. Thus it has solved the problem of excessive roughness due to the lower polishing rate, avoiding reducing the device performance with the pattern wafer. The new chelating agent with its lower pH value could make it possible to protect the low-k dielectric under the barrier layer from structurally breaking. The CMP experiment was performed on the 12 inch MIT 854 pattern wafers with the barrier slurry containing the new weakly alkaline chelating agent. By the DOE optimization, the results indicate that as the new chelating agent concentration in the slurry was up to 2.5 mL/L, the copper polishing rate is about 31.082 nm/min. Meanwhile, the wafer surface has a rather low roughness value of 0.693 nm (10 × 10 μm), the correction ability with the above slurry is adapted to the next generation barrier CMP and the k value of the low-k dielectric seems to have no k-shift. All the results presented show that the new weakly alkaline chelating agent with its superior performance can be used for the advanced barrier CMP.

Shallow Trench Isolation Chemical Mechanical Planarization: A Review

Electrical isolation of the billion or so active components in each integrated device is achieved using shallow trench isolation (STI) which requires chemical mechanical planarization (CMP) involving silicon dioxide removal at a high rate and stopping on an underlying silicon nitride film. Several colloidal slurries with various additives can yield the desired high rate selectivity between the oxide and nitride films during CMP while maintaining an acceptably low nitride rate. Here, many of such high selectivity STI CMP slurries described in the literature are reviewed along with the characteristics of the colloidal dispersions like the abrasives, additives, the interactions between them and with the films being planarized and the associated pH range in which the high selectivity is observed. The mechanisms proposed to explain the high reactivity of ceria with oxide, the role of additives in suppressing the nitride removal rate and resulting high selectivity are discussed. Reduction of a multitude of defects in post-CMP processed STI structures still remains an important challenge, especially as the feature sizes continue to shrink.

Tribo-Electrochemical Characterization of Ru, Ta and Cu CMP Systems Using Percarbonate Based Solutions

Defect-control is a critical requirement for chemical mechanical planarization (CMP) of the ultrathin diffusion barriers considered for the new Cu-interconnects. The challenging task of developing advanced CMP slurries for such systems can be aided by electrochemical evaluations of model CMP schemes under tribological conditions. The present work uses this strategy to characterize an alkaline slurry formulation aimed at minimizing galvanic corrosion in the CMP systems involving Ru, Ta (barrier metals) and Cu (wiring metal). This slurry is based on percarbonate and guanidine additives, and the test metals are polycrystalline disc samples. A particular goal of this study is to explore the essential analytical aspects of evaluating CMP systems in the tribo-electrochemical approach. The CMP specific surface reactions are characterized by potentiodynamic polarization and open circuit voltage measurements, performed both in the presence and in the absence of polishing, and by employing abrasive free as well as abrasive (colloidal SiO2) added solutions. The findings of these experiments are further checked by using impedance spectroscopy. The electrochemical mixed potential steps of the CMP promoting reactions are analyzed, and the removable surface species formed by these reactions are discussed.

Characterization of Nano-Scale Protective Oxide Films: Application on Metal Chemical Mechanical Planarization

This study focuses on the characterization of nano-scale metal oxide films for chemical mechanical planarization (CMP) applications. The protective nature of the self-grown metal oxide layers in the CMP slurry environment enable topographic selectivity required for metallization of interconnects. Tungsten was selected as the model metal film to study the formation and characteristics of the metal oxide nano-layers since tungsten CMP is very well-established in conventional semiconductor manufacturing. The tungsten oxide nano-films were characterized for thickness, density and surface topography in addition to evaluation of their protective nature by calculation of the Pilling-Bedworth (P-B) ratios. It was observed that in addition to controlling the self-protective characteristics, the oxidizer concentration also affects the surface structure of the metal oxide films resulting in significant changes in the CMP process performance in terms of material removal rates and surface finish with a sweet-spot detected at 0.075 M H2O2 concentration.

Stability of alumina, ceria, and silica nanoparticles in municipal wastewater

Inorganic oxide nanoparticles (NPs) are used in semiconductor manufacturing operations such as wafer chemical-mechanical planarization (CMP). Understanding the stability of NPs in municipal wastewater is essential for the evaluation of the fate of NPs released to municipal wastewater treatment plants (WWTPs). This study aimed to evaluate the stability of Al(2)O(3), CeO(2), and SiO(2) NPs and CMP waste effluents containing these NPs in municipal wastewater. Al(2)O(3) and CeO(2) NPs were destabilized by wastewater constituents, as indicated by the formation of large agglomerates. However, the same NPs in the CMP waste slurries showed high stability in wastewater, probably due to additives present in the slurry that modify the surface chemistry of the particles. Likewise, both the commercial SiO(2) NPs and the CMP waste slurry containing SiO(2) NPs showed substantial stability in wastewater since this NP has a very low point of zero charge, which suggests that this NP could be the hardest one to remove in conventional WWTPs by aggregation-sedimentation. In summary, the results indicate that wastewater may destabilize NPs suspensions, which would facilitate NP removal in WWTPs. However, some chemicals present in real CMP slurries may counterbalance this effect. More research is needed to completely understand the surface chemistry involved.

Application of Fluorescence Correlation Spectroscopy in the Characterization of Particle Size Distributions of Colloidal Silica Abrasives Used in Chemical-Mechanical Planarization

Particle size distribution (PSD) analysis of colloidal silica abrasives employed in chemical-mechanical planarization slurries is demonstrated using fluorescence correlation spectroscopy (FCS). FCS allows a quantitative determination of the PSD for a dispersion of abrasive silica particles through fluorescence measurements of a dye, Rhodamine 110 (R110), adsorbed to the particles. Analysis of the autocorrelation function (ACF) of the fluorescence intensity fluctuations from R110-dyed silica confirms the presence of the adsorbed dye on the particles. Application of the method of histograms to the ACF yields a detailed analysis of the shape of the PSD for silica particle dispersions. FCS is shown to reliably measure the hydrodynamic diameter of silica particles less than 20 nm in mean size, and at particle concentrations which are too dilute for PSD analysis using a widely applied method for this purpose, dynamic light scattering.

Current Status of Slurries and Cleans for CMP of III-V Device Fabrications – A Critical Review

With the growth of cloud computing market, the need for logic and memory chips for data centers have been steadily growing which require low power and high performance. Another area of growing semiconductor application is mobile products which require increasingly higher performance without sacrificing power. The above two requirements can only be achieved by scaling supply voltage (Vdd). III-V materials enable Vdd scaling to address power without sacrificing performance due to the very high electron injection velocity at virtual source compared to Si. For example, Vinj (InGaAs) > 2Vinj (Si) at less than half VDD. The development at Intel on InGaAs and at IMEC on Ge based FinFET structures have made it possible to conceive new integration schemes[1]. In order to enable such CMOS fabrication, development of CMP (Chemical Mechanical Planarization) slurry, post CMP clean and modification of the presently available process tools need to be done. While new slurries and cleans will be needed to planarize III-V film layer(s) without compromising the film composition (and functionality), modification of the process tool will be needed to address potential release of toxic waste water into the waste stream after planarization. In this paper, the status of development of CMP processes, slurries and post CMP cleans will be reviewed and the critical needs to enable for the future III-V device integration will be analyzed.

A Cahn-Hilliard Modeling of Metal Oxide Thin Films for Advanced CMP Applications

Chemical mechanical planarization (CMP) process enables topographic selectivity through formation of a protective oxide thin film on the recessed locations of the deposited metal layer, while a continuous chemical oxidation reaction is followed by mechanical abrasion takes place on the protruding locations. This paper demonstrates a new approach to CMP process optimization in terms of analyzing the nano-scale surface topography of the protective metal oxide films and modeling their growth through a Cahn-Hilliard equation (CHE) approach as an alternative to classical nucleation theory. It is observed that the material removal rate mechanisms and the consequent planarization performance depend on the nature of nucleation of the metal oxide films, which is a function of the oxidizer concentration of the CMP slurry.

Benzotriazole as a passivating agent during chemical mechanical planarization of Ni–P alloy substrates

Abstract With the rapid increase of data storage density on computer hard disk drives (HDDs), the operation distance between read/write head and disk surface has fallen to just a few nanometers. Chemical mechanical planarization (CMP) has been selected as the best process to produce high quality surface finish during the manufacturing of Ni–P alloy substrates for HDD applications. Herein we report, for the first time, the use of benzotriazole (BTA) as a passivating agent in CMP slurries to decrease the surface roughness ( R a ). Results show that the average R a of the polished surfaces is decreased to 0.2 nm in a 5 μm × 5 μm scan area with the adding of 2 mM BTA. X-ray photoelectron spectroscopy (XPS) and electrochemical studies results further prove the interaction between BTA and Ni–P surface and the formation of an effective passivating layer on Cu in CMP slurries containing BTA.

Inhibition Mechanism of Benzotriazole in Copper Chemical Mechanical Planarization

During the process of chemical mechanical planarization (CMP) of copper, benzotriazole (BTA) is the most commonly used inhibitor in the slurry. Though the corrosion inhibition mechanism has been studied widely, the mechanism of BTA layer on copper surface in CMP slurries should be further investigated. In this paper, the adsorption mechanisms of BTA were studied by static corrosion tests. Besides, the surface composition was measured by XPS. Combining with CMP experiments, the material removal mechanism of copper CMP depending on pH values was investigated. It was found that the formation of passive film, consisting of Cu-BTA complex, adsorption of BTA and copper oxides, played a dominant role under acidic conditions. While the surface film composed of adsorption layer of BTA and copper oxides under alkaline conditions. The inhibition mechanism of BTA varied with pH values, resulted in corresponding changes of material removal rate and coefficients of friction.

TiO2 Abrasive for Dielectric CMP Application

Ceria (CeO2) was demonstrated to have the highest polishing rate of glass than the other ceramic oxides by Cook [1] in 1990. Since then CeO2 particle was increasingly applied as the abrasive in chemical mechanical planarization (CMP) of the SiO2 and Si3N4 based dielectric materials. At present, the share of the CeO2 slurry has grown to the second largest part in the CMP slurry market just next to the silica slurry. However, the price of precursor of CeO2 synthesis, as one of the rare earth materials, rose significantly as long as the export quota was restricted for the political policy by the rare earth material exporting countries [2]. This rise in price would have direct influence on the cost of ownership and the profit margin of the semiconductor manufacturing companies. Therefore, an alternative is required to totally or partially replace the CeO2 slurry in the present dielectric CMP application. ZrO2 and TiO2 had the second and third high polishing rates in the Cook’s report. However, the extremely high mechanical hardness of ZrO2 prohibited its potential application in the dielectric CMP due to the concern of producing scratches. TiO2 has the moderate mechanical hardness and applied in large number of applications including food, cosmetic, catalyst, and paint [3]. Therefore, the annual production capacity and the price would be more stable than that of the rate earth materials. For this reason, we investigated the SiO2 CMP using the TiO2particle in the study. Figure 1 shows the Transmission Electron Microscopy of the TiO2 particle used in the study. We used commercially available TiO2(P25, Degussa, Germany) nanoparticles. Degussa P25 consists of two crystal forms of rutile and anatase, and is widely used in photocatalytic degradation studies, because of its chemical stability, ready availability, reproducibility, and activity as a catalyst in oxidation processes [4]. Figure 2 indicates the polishing rate of SiO2 film in the TiO2 slurry as a function of the added polyacrylic acid (PAA) concentration. The polishing rate was below 200 Å/min in the TiO2 slurry. However, it increased suddenly approximately to 700 Å/min as the PAA concentration increased to 0.025 wt%. Surprisingly, the polishing rate abruptly dropped below 100 Å/min as the PAA concentration increased by 0.005 wt% to 0.03 wt%. The dispersion stability of the TiO2slurry was good as long as the PAA concentration was over 0.02 wt%. Therefore, the poor dispersion stability can explain the low polishing rate at the PAA concentration that was lower than that of the critical point. However, the dispersion stability was failed to explain the abrupt decrease in the polishing rate over 0.03 wt% of PAA. Figure 3 shows the viscosity of the TiO2 slurry as a function of the PAA concentration. Each slurry shows the shear thining behavior as the sear rate increases due to the structural rearrange of the adsorbed PAA on the TiO2 particle. The magnitude of the viscosity shows a totally inverse trend as that of the polishing rate as a function of PAA concentration. The adsorbed PAA on the TiO2particle probably undergoes a structural change as the PAA concentration over the critical concentration of 0.025 wt%. Therefore, the abruptly decreased polishing rate would be resulted from this structural change. The TiO2 slurry showed a moderate polishing rate of SiO2 with PAA only in a very narrow concentration range. Therefore, we need more investigation to understand the CMP mechanism of the TiO2 slurry and find out a more effective additive to increase the polishing rate of SiO2 in order to satisfy at least the same polishing performance as the CeO2 slurry.

Research of New Concept Dishing-Less Tungsten CMP Slurry Using Using 25 Nm Crystal Zirconium Dioxide Abrasive

As design rule of semiconductor device shrinks, aspect ratio is becoming larger at wiring formation process [1], and the importance of chemical mechanical planarization (CMP) process is gradually increasing [2]. However, dishing and erosion of wiring generated from CMP process induced the fatal problems such as the increase of resistance and heat generation [3]. The solution for fatal defects such as dishing and erosion has been investigated for a long time, but the progress is still slight. Erosion can be controlled by enhancing polishing selectivity using surfactant to rapidly decrease polishing rate of SiO2 film. However, dishing cannot be controlled in the case of the chemical-dominant CMP slurry, which mainly polishes the oxidized W surface. Therefore, CMP process of the adequately mixed mechanical and chemical-concept is necessary to dramatically improve dishing. This research suggests the possibility that the dishing was dramatically improved using 25 nm crystal zirconium dioxide (ZrO2) abrasive.All slurry was composed of an abrasive, a dispersant (except colloidal silica), a pH titrant, an oxidizer (H2O2), mixing with iron(III) nitrate (Fe(NO3)3) as catalyst, and deionized water (DIW). All of slurry was titrated to pH 2.3 with the addition of nitric acid. We used 6 x 6 cm W and oxide wafers to observe polishing performance using a polishing machine (poli-300, G&P Tech. inc., Korea). And we used the line pattern wafer. It has 10 nm width and about 250 nm thickness above SiO2fillm. We deposited about 300 nm thick W film by Chemical vapor deposition method. Before and after the pattern polishing, the cross-sectional analysis of the patterned wafer was fabricated by a Forcused Ion Beam (Quanta 3D 200i, FEI Corp., USA). A High-resolution Transmission Electron Microscopy (HR-TEM) (JEM-2000, JEOL, Japan) image of line pattern was obtained with an accelerating voltage of 200 kV. Figure 1 shows the HR-TEM of the colloidal silica and ZrO2 particle used in the study. We used commercially available nano-sized colloidal silica (PL-7, Fuso, Japan), and ZrO2(99.9%, SS-Nano, USA) nano-particles as the abrasive consisted of crystal forms of monoclinic. Figure 2 shows the polishing rate of W and oxide films in colloidal silica and ZrO2 slurries as a function of the Fe(NO3)3 concentration as catalyst. The polishing rate of W film was slightly increased from 56 to 1100 A/min in comparison with ZrO2 slurry, while the polishing rate of oxide film was not changed to about 650 A/min according to Fe(NO3)3 concentration in the colloidal silica slurry. However, the polishing rate of W film was continually increased from 29 to 2390 A/min, while the polishing rate of oxide film was decreased from 372 to 85 A/min according to Fe(NO3)3 concentration in the ZrO2slurry. Figure 3 shows the TEM images of the line-pattern before and after polishing in the colloidal silica and ZrO2 slurries. The average dishing and erosion was 6.8 and 10.7 nm in the colloidal silica slurry, respectively. In the ZrO2slurry, however, the average dishing and erosion reduced in comparison with colloidal silica, which was 0.9 and 5.8 nm, respectively.The ZrO2 slurry in comparison with colloidal silica slurry indicated good performances such as high polishing rate, selectivity, low dishing and erosion, etc. However, we need more investigation to understand the CMP mechanism of the ZrO2 slurry. We just guess that the good performances in comparison with colloidal silica are due to a small particles size and a crystal structure of the ZrO2 abrasive in the slurry.

Characterization of non-amine-based post-copper chemical mechanical planarization cleaning solution

Copper surfaces can become contaminated by slurry particles and organic residues during chemical mechanical planarization (CMP). Silica particles are widely used as abrasives, while benzotriazole (BTA) is widely used as corrosion inhibitor, in copper CMP slurries. The contaminated copper surface needs to be cleaned by using an effective cleaning solution. These material contaminate the copper surface during CMP and need to be removed by using an effective cleaning solution. The objectives of this work were to develop a non-amine-based alkaline cleaning solution and characterize the solution based on the benzotriazole (organic residue) removal and particle removal efficiency. Cesium hydroxide and potassium hydroxide were used as cleaning agents and ethylene glycol was used as a corrosion inhibitor. Ethylene glycol acts as a chelating agent as well in the cleaning composition. BTA removal was characterized using contact angle measurements, X-ray photoelectron spectroscopy and electrochemical impedance spectroscopy (EIS) techniques. The corrosion protection ability of the cleaning solutions was quantified by potentiodynamic polarization studies. Both potassium hydroxide- and cesium hydroxide-based solutions exhibited high BTA and silica particle removal. When compared to a potassium hydroxide based cleaning solution, cesium hydroxide based cleaning solutions were found to be more effective in terms of low surface roughness and low etch rate.

A new weakly alkaline slurry for copper planarization at a reduced down pressure

This study reports a new weakly alkaline slurry for copper chemical mechanical planarization (CMP), it can achieve a high planarization efficiency at a reduced down pressure of 1.0 psi. The slurry is studied through the polish rate, planarization, copper surface roughness and stability. The copper polishing experiment result shows that the polish rate can reach 10032 Å/min. From the multi-layers copper CMP test, a good result is obtained, that is a big step height (10870 Å) that can be eliminated in just 35 s, and the copper root mean square surface roughness (sq) is very low (< 1 nm). Apart from this, compared with the alkaline slurry researched before, it has a good progress on stability of copper polishing rate, stable for 12 h at least. All the results presented here are relevant for further developments in the area of copper CMP.

Effects of chemical mechanical planarization slurry additives on the agglomeration of alumina nanoparticles II: Aggregation rate analysis

The aggregation rate and mechanism of 150nm alumina particles in 1mM KNO3 with various additives used in chemical mechanical planarization of copper were investigated. The pH of each suspension was ∼8 such that the aggregation rate was slow enough to be measured and analyzed over ∼120min. In general, an initial exponential growth was observed for most suspensions indicating reaction-limited aggregation. After aggregate sizes increase to >500nm, the rate followed a power law suggesting diffusion-limited aggregation. Stability ratios and fractal dimension numbers were also calculated to further elucidate the aggregation mechanism.

Germany’s Merck Bets On Chips

In a bid to broaden its offering to the electronics industry, the German drug and chemical maker Merck KGaA has agreed to purchase AZ Electronic Materials for $2.6 billion. AZ is the former electronic chemicals business of Clariant, which sold the unit to the private equity firm Carlyle Group in 2004. An initial public offering of AZ stock in 2010 valued the company at about $2 billion. More than two-thirds of AZ’s nearly $800 million in 2012 sales came from materials that are used to fabricate computer chips. They include dielectrics, chemical mechanical planarization slurries, and photoresists. Most of the rest of its business is in materials for displays and lighting. The firm generated about $260 million in pretax profits last year. Merck is already a player in electronic chemicals through its business in materials for liquid-crystal displays. Its electronics sales last year were about $1.6 billion. The company has ...

Abrasive and additive interactions in high selectivity STI CMP slurries

Ceria based slurries with additives are widely used in shallow trench isolation (STI) chemical mechanical planarization (CMP) process to obtain high selective removal of silicon dioxide over silicon nitride. In this study ceria from different sources were used as abrasives and l-proline and l-glutamic acid were used as additives, with a focus on identifying the interactions between abrasives and additives and their effect of selectivity. Ceria particles were characterized using X-ray diffraction (XRD), energy dispersive X-ray (EDX) spectroscopy, transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) while the additive abrasive interactions were evaluated by UV–Visible spectroscopy and inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis. While slurries with l-proline yielded high selectivity only with certain type of ceria, slurries with l-glutamic acid were found to be less sensitive to the ceria source than those with l-proline, and yielded high selectivity regardless of the source of ceria used. The purity of the abrasive and its crystal structure appear to play a significant role in determining the selectivity.

Preparation of Composite Spherical Alumina Nanoparticles and Study on their CMP Performance

The properties of abrasives have great influences on the surface qualities of glass substrate, hard disk, etc[1], during the chemical mechanical planarization (CMP) process. α-Al2O3 particles, as one of the most widely used abrasives in CMP slurries, often cause great surface defects because of its high hardness and unfeasible control of agglomeration during the synthesizing process [2,3]. In this paper, spherical nanocomposite alumina powders with particle sizes of 10-50 nm were prepared, via a homogeneous precipitation method. The structures and surface morphologies of the alumina nanopowders were characterized by means of X-ray diffraction (XRD), scanning electron microscope (SEM) and atomic force microscope (AFM). The results showed that the nanopowders composed of both alpha and theta phase alumina, which had a small particle size and narrow size distributions. The CMP investigations performed on glass substrate showed that the nanocomposite alumina abrasives exhibited a better surface planarization performance than the pure single alpha phase Al2O3 nanopowders, with lower surface roughnesses and less scratches after CMP process. The results indicated that the current processing for the synthesis of nanocomposite Al2O3 powders were of great potentials for practical applications of Al2O3 based abrasives.

(Invited) Surface Adsorption of CMP Slurry Additives on Abrasive Particles

Fluorescence correlation spectroscopy (FCS) is used to study additive-abrasive particle interactions in chemical-mechanical planarization (CMP) slurries. FCS provides quantitative determinations of the binding between additives and abrasive particles by characterizing the competitive adsorption of the additive and a fluorescent probe molecule, Alexa fluor 546 (A546) at the surface of a silica abrasive particle. Analysis of the adsorption of benzotriazole (BTA) on colloidal silica confirms the displacement of A546 by added BTA. However, glycine enhances the adsorption of A546 by colloidal silica. FCS is also shown to be a sensitive technique for detecting differences in the surface chemistry between different types of colloidal silica abrasive particles used in CMP processes.

BEOL Cu CMP Process Evaluation for Advanced Technology Nodes

The back end of the line (BEOL) copper (Cu) interconnect line width and spacing is decreasing with each advancement in technology node. It is important to understand how the decreasing dimensions will impact BEOL chemical mechanical planarization (CMP) process. This study is focused on evaluation of Cu CMP using three categories of slurry formulations (i.e. alumina abrasive, silica abrasive and abrasive-less) by utilizing test structures with dimensions relevant to sub-20 nm technology. The results presented in this study show the key factors driving the performance of slurry formulation (e.g. liner selectivity, defects) are strongly dependent on the copper (Cu) interconnect line width and spacing. The results show the blanket rate selectivity does not translate to equivalent selectivity on patterned structures. In addition, the slurry formulations lost selectivity to liner as interconnect width/spacing is reduced from 40 nm to 32 nm highlighting the need to account for pattern density when predicting selectivity. Furthermore, evaluation of Cu CMP slurry formulations with cobalt (Co) liner indicates that existing formulations exhibit no Cu to Co selectivity. Overall, the results on multiple interconnect dimensions from this study are very important for researchers focused on slurry formulation development and CMP modeling.