Tungsten Chemical Mechanical Polishing Research Articles

Assessing the wear degradation of tungsten in acidic solutions using a lubricated tribocorrosion model

The tribocorrosion property of tungsten plays a great role in the chemical mechanical polishing of tungsten plugs used in integrated circuits. This study applied a lubricated tribocorrosion model to tungsten, verified and calibrated the model using results from laboratory tribocorrosion experiments, and predicted the effect of the key parameters on the mechanical and chemical wear behavior of tungsten. The results showed that the model could be successfully applied to the tribocorrosion of tungsten considering that the prevailing wear mechanism and the electrochemical corrosion property of tungsten fulfills the model's basic concepts. The model was able to predict the influence of the material, mechanical, electrochemical and rheological parameters on the mechanical wear and chemical wear behavior of tungsten. The model could clearly distinguish the contribution of the mechanical wear and chemical wear to the total wear degradation of tungsten, which provides useful instructions on how to control the related parameters in order to obtain tailored removal rate and surface quality in tungsten's chemical mechanical polishing process.

Origins of wear-induced tungsten corrosion defects in semiconductor manufacturing during tungsten chemical mechanical polishing

During chemical mechanical polishing (CMP) process optimization, a unique corrosion defect was identified in the tungsten (W) contact plugs, which were investigated to develop practical solutions to improve CMP mass productivity. The corrosion defect occurred in the absence of external incident energy and was observed to occur only at the W plugs connected to the P+/n-well close to the wafer edge. To examine the root cause of the corrosion, a series of experiments was carried out at each W CMP step, using in-line defect inspection. Additional experiments were conducted using on-line inductively coupled plasma mass spectrometry (ICP-MS) to understand pH- and potential-dependent electrochemical dissolution behaviors of W. Experimentally, W corrosion was confirmed to be exacerbated if the surface oxide of the W plugs is removed by the deionized water (DIW) rinsing step under wafer pressurization, then exposed to DIW and alkaline cleaning solution. Although the DIW rinsing step is commonly performed to clean polishing pads right after the main polishing, a small amount of applied pressure (0.5 psi) without slurry supply can lead to corrosion defects. This wear-induced corrosion of the W plugs can be prevented by removing the applied pressure during the post DIW rinsing step of W CMP.

Origins of Wear-Induced Tungsten Corrosion Defects in Semiconductor Manufacturing During Tungsten Chemical Mechanical Polishing

Origins of Wear-Induced Tungsten Corrosion Defects in Semiconductor Manufacturing During Tungsten Chemical Mechanical Polishing

Study on Abnormally Accelerated Oxidation of Tungsten Plugs during Tungsten Chemical Mechanical Polishing in Semiconductor Manufacturing

The root cause of the abnormally accelerated oxidation behavior occurring in tungsten (W) chemical mechanical polishing (CMP) has been performed to provide guidance to those developing practical solutions. Abnormal oxidation behavior refers to the case where the surrounding W contact plugs are normal, but oxidation occurs rapidly and abruptly only at a specific contact, causing device failure. For this study, the failure location was determined through a wafer-level probe test of fully processed DRAM device wafers. With this specific location information, a top-down view inspection was performed on the location immediately after CMP to assess the characteristics, and then a wafer-level probe test was conducted to confirm the failure. In this way, the top-down view plug images immediately after CMP and their potential failure were classified. Based on these results, the CMP process step involving abnormal oxidation was determined, as were the conditions needed to cause abnormal oxidation. Potential solutions to abnormal oxidation are presented based on our results.

Investigation of particle agglomeration with in-situ generation of oxygen bubble during the tungsten chemical mechanical polishing (CMP) process

This research investigates abrasive particles agglomeration via interaction between O2 bubbles and slurry abrasives during the tungsten chemical mechanical polishing (W CMP) process. The abrasive particles in slurry were highly agglomerated due to higher volumes of O2 bubbles produced in the reaction between the catalyst Fe(NO3)3 and the oxidizer H2O2. Results obtained from a gas pressure sensor confirmed the generation of higher O2 volume via the decomposition of H2O2 at a high catalyst concentration and an increase in reaction temperature. The decomposed O2 volume rate at 80 °C was reported at the maximum value of 2.0 × 10−2 L/s at 120 ppm as compared to the moderate and minimum rates of 3.5 × 10−3 and 3.2 × 10−4 L/s for catalyst concentrations of 60 and 30 ppm, respectively. Images of O2 bubbles, captured using a high-speed camera, exhibited subsequent enhancement in average O2 bubble diameters of 91, 427, and 503 μm at 25, 60, and 80 °C, respectively. Analysis of surface scans confirmed large abrasive particles contamination on the TEOS wafer with an increase in the O2 bubble flow rate and bubbling time. Also, large abrasive particles agglomeration was observed in the presence of O2 bubbles as compared to no bubbles, as measured by dynamic light scattering DLS. It is believed that higher hydrophilicity of abrasive particles with O2 bubbles increased the adhesive force between the abrasive particles and the in-situ generated O2 bubbles. The high drag force generated during the collapse of O2 bubbles is essentially attributed a strong attractive force between the abrasive particles and the TEOS wafer which strongly binds with the abrasive particles and intensifies the defect level as particle agglomeration.

A New Approach to the Formation Mechanism of Tungsten Void Defect in Chemical Mechanical Polishing

The advanced node semiconductor device fabrication needs tungsten chemical mechanical polishing as a critical process to enable transistor formation as well as metal interconnection. Accordingly, CMP induced defect concerns have drawn attention to the process integration. Tungsten voids by CMP are one of the major defects to detract device yield. Their formation mechanism has been known as electrochemical corrosion or an optical induced corrosion phenomenon associated with polishing slurry chemistry. However, this paper proposes a new mechanism of tungsten void formation. Different from known corrosion mechanisms, it is observed that tungsten void frequency has strong correlation with CMP in-situ cleaner brush lifetime. At early brush life, tungsten void forms more frequently. Moreover, this paper reveals that CMP brush cleaning play a critical role in the formation of tungsten voids. Tungsten etch rate dependence on brush condition, tungsten void defect density correlation to brush life, and surface chemical analysis of tungsten blanket wafers show that tungsten voids are induced by combined effect of friction force and oxidation in brush cleaner module, which is a new approach to the formation mechanism of tungsten void defect.

Study on the Mechanism of Nano-Flake Defect during Tungsten Contact Chemical Mechanical Polishing

Tungsten chemical mechanical polishing (CMP) has been used for metal interconnect formation in semiconductor manufacturing for a long time. Its application continues to the advanced device manufacturing with transistor scaling. However, the defect concerns of the tungsten CMP process have become critical to manufacturing such as metal flake defects that can cause electrical shorts between contacts. Although there are many published studies on tungsten CMP fundamentals, process applications and CMP-induced defects, the mechanism of thin metal flake formation in nanoscale (nano-flake defect) has not been clearly known to industry. In this study, nano-flake defects are identified after tungsten contact CMP, and mechanism of its formation is explored. Significant CMP dishing at wide contact area removes all tungsten and open barrier metal. Friction force by 3-body interaction among wafer, abrasive particle and pad asperity breaks the adhesion of barrier material from intermetallic dielectric material. And it allows re-deposition of barrier metal to the wafer surface. Insufficient CMP in-situ cleaning results in residual nano-flake defects at the end of CMP process. Analysis of device macros, thickness of flake, cross sectional analysis by transmission electron microscope (TEM) and topographical analysis support this suggested mechanism.

Investigation of the pad-conditioning performance deterioration in the chemical mechanical polishing process

The surface roughness of a chemical mechanical polishing (CMP) pad provides microscopic asperities where the actual polishing takes place. Thus, the roughness should be maintained in a proper regimen to obtain stable CMP process quality. A conditioner is a key consumable in the CMP process because it maintains the roughness of the CMP pad surface against surface glazing or flattening during polishing. Because the conditioner mechanically breaks the pad surface to maintain the roughness, previous research has expressed its ability index in terms of the pad wear rate (PWR). In the present study, we investigate the reason for PWR change of CMP pads having vigorously changing moduli with respect to temperature changes during the mass production. The pad property changed steeply in the process temperature range from 60 to 80°C. The tensile strength is a key property change that leads to sudden PWR drop. We found that the shape and configuration of the diamond tip are the major parameters affecting the ability to maintain surface roughness of the pads. Thus, we suggested a proper conditioner design for tungsten (W) CMP based on the experimental data to sustain the process quality and increase the consumable lifetime. As a result, the peak surface roughness has improved from 29.4 μm to 32.6µm and the consumable lifetime was also increased.

Control of Tungsten Protrusion with Surface Active Agent during Tungsten Chemical Mechanical Polishing

With shrinkage of the minimum feature size to sub-14 nm, protrusion/dishing issues in chemical mechanical planarization (CMP) processes have become increasingly important to address. In this study, we propose an advanced slurry formulation with a surface active agent to prevent W protrusion during the W CMP process. In the presence of surface active agent, blanket removal rates on both W and SiO2 films showed non-Prestonian behavior at a low threshold pressure, which is explained by adsorption characteristics of the surface active agent on the films. To study the effect of the surface active agent on W protrusion, W-patterned wafers were polished at the threshold pressure. As the concentration of the surface active agent increased from 0 to 12 mM, W protrusion from patterns with line/space widths of 0.5 μm/0.5 μm decreased significantly from 186 to 30 Å. The mechanism on the significant decrease in W protrusion behavior in the presence of the surface active agent is discussed.

Impact of Film Morphology on Chemical Mechanical Polishing of Tungsten

This study investigated the effects of tungsten (W) film morphology on the chemical mechanical polishing (CMP) of W. Chemical vapor deposited (CVD) W films with two distinctively different grain sizes were used for comparison. During polishing, W film thickness and optical reflectance, friction, and pad temperature were monitored in-situ. It was found that larger-grained W film took longer to pass the initial low removal rate stage. By correlating four different sensor signals, comparing friction dependency on film morphology, in slurry vs. DIW, it was concluded that W CMP comprises three main stages. First is the low rate initiation stage: grain is being partially planarized, reflectance increases, friction decreases. Second is the transition stage: rate is ramping, grain becomes fully planarized, optical reflectance reaches maximum, and friction becomes minimal followed by a significant rise caused by formation of tungsten oxide passivation layer on the planarized W surface. Third is the high and constant rate stage: passivation and removal occur in a repetitive cycle, friction is high and stable, optical reflectance changes as polishing reaches different film depths. In all three stages, pad temperature increases continuously as friction-induced heat dissipates, with the rate of temperature increase following that of friction magnitude.

Effects of Pad Temperature on the Chemical Mechanical Polishing of Tungsten

In this study, the effects of pad temperature on the chemical mechanical polishing (CMP) of tungsten were investigated. During polishing, the pad temperature was monitored and the removal rate behavior with respect to the pad temperature evolution was explored. The pad temperature was increased during tungsten CMP, and the tungsten removal rate increased accordingly. Further, it was found that the two had a strong correlation. Although tetraethyl orthosilicate (TEOS) polishing experiment showed that the pad temperature increased similarly under various experimental conditions, the TEOS removal rate did not change for all conditions. A novel pad-cooling nozzle effectively reduced the pad temperature during tungsten polishing and influenced the tungsten removal rate drop. Table motor current monitoring revealed that friction forces were almost identical under all experimental conditions for tungsten polishing regardless of the application of pad cooling. On the basis of pad temperature and table motor current data, it was clear that the tungsten removal rate was determined by temperature, indicating that the tungsten removal mechanism was mostly driven by a chemical reaction. Further, the pad cooling function can be used to control the tungsten removal rate effectively, resulting in a stable tungsten CMP process.

Recycling a slurry for reuse in chemical mechanical planarization of tungsten wafer: Effect of chemical adjustments and comparison between static and dynamic experiments

Recycling abrasive slurry that has been used in chemical mechanical polishing (CMP) is one of the options for reducing the cost of manufacturing microchip processors. We use ultrafiltration which is a method of choice to recycle silica (SiO2)-based slurry. Taking into account that the chemical composition of abrasive slurry plays an important role in tungsten CMP (W-CMP), chemical adjustments have to be made so that the concentrated after used slurry can be reused. In this study, we investigate the effects of chemical additives (iron catalyst, oxalic acid as complexing agent and surfactants as stabilizers) in slurry that has been retreated by ultrafiltration. Experiments are conducted both under static and dynamic conditions and results are compared to better understand the effect of chemical adjustments on the main performances of W-CMP. An optimal chemical adjustment is proposed through a design of experiments evaluation to obtain a concentrated after used and chemically adjust slurry comparable to the operational point of use slurry.

Diamond Conditioner Microwear Effect on Pad Surface Height Distribution in Tungsten Chemical Mechanical Polishing

In this study, the surface topographies of chemical mechanical polishing (CMP) pad samples for varying levels of diamond microwear of a conditioner have been measured using a confocal microscope and an X-ray computer tomography (CT) scanner. The experimental results showed that the increase in the pad debris on the pad surface reduced the pad height ratio of the asperity called the “top surface area (TSA) ratio”. In addition, the overall removal rate in tungsten CMP was more dependent on the TSA ratio after polishing than on that after conditioning because the pad surface condition became worse with deformed asperities and micropores due to the insufficient conditioning.

Diamond Conditioner Microwear Effect on Pad Surface Height Distribution in Tungsten Chemical Mechanical Polishing

Diamond Conditioner Microwear Effect on Pad Surface Height Distribution in Tungsten Chemical Mechanical Polishing

Static dissolution rate of tungsten film versus chemical adjustments of a reused slurry for chemical mechanical polishing

Tungsten is widely used as deposited layer for the multi-level interconnection structures of wafers. The chemical composition of abrasive slurry plays an important role in chemical mechanical polishing (CMP) process. Removal of tungsten is driven by complex oxidation mechanisms between slurry components. The slurry for tungsten CMP generally contains oxidizer, iron catalyst, complexing agents and stabilizers in a pH adjusted solution of abrasive particles. Interaction between iron complex and H 2O 2 in the slurry is the main factor governing the chemical mode of material removal, oxidation potencies and kinetics. In this study, we investigate the effects of chemical additives in silica (SiO 2)-based slurry on the removal rate of the tungsten film. Experiments were carried out in static batch as a preliminary study to understand and optimize chemical mechanisms in CMP–Tungsten process. Experiment designs were conducted to understand the influence of the chemical additives on the main performances of W–CMP. Used slurry, concentrated and retreated with chemical adjustments, is compared to the original slurry as a reference.

タングステンＣＭＰスラリーのリサイクルに関する基礎的研究

Chemical Mechanical Polishing (CMP) system is an essential technology to produce the recent high performance electronic devices. One of the problematic point of CMP processes is its high cost. Another is its heavy load to the environment. The principal cause of these two points is that this system needs to be poured and flushed by slurry continuously. One method to solve this problem is to recycle and reuse waste slurry used in this system. This research is targeted on slurry for “Tungsten CMP” which is particularly problematic for its high cost. Untreated slurry from “Tungsten CMP” contains water and metal. These contents will degrade the processing characteristics of slurry. Examinations were conducted to ascertain if recycling waste slurry is possible by eliminating contained water and metals. It was concluded that slurry can be recycled by abstracting water by ceramic filter and metal by ion-exchange. It was also verified that this recycled slurry reserved equivalent processing characteristic as the new slurry.

Novel Approach for W Loss Defect Prevention on WCMP Re-Clean Process

With the device dimension rapidly shrinking, Tungsten Chemical Mechanical Polishing (WCMP) becomes one of the critical manufacturing processes for multi-layer interconnection of ULSI (ultra large scale integrated) circuits. The effectiveness of Post-CMP cleaning is one of critical step for particles removal and product Cp yield improvement. WCMP re-clean process is a general way to further reduce particles on high-defect-density impacted wafers in IC manufacturing. However, W loss defect could be generated frequently due to the extra re-clean process introduced in post WCMP step for sub-0.13um devices. In this paper, the mechanism of W loss defect after re-clean process is studied. Also one novel approach is proposed to prevent W loss and improve product CP yield.

Experimental Analysis of Pad Wear Response Effect on Removal Rate Variations in Tungsten Chemical Mechanical Polishing

We investigated the effects of the pad surface response during polishing and conditioning on the removal rate variation in a tungsten chemical mechanical polishing (CMP) process. We examined the coefficient of friction (COF) to identify the tribology of the pad conditioning. The variation in the removal rate was dependent on the COF during conditioning. We also evaluated the effects of diamond edge degradation of the pad conditioner disk on its cutting ability from the viewpoint of mechanical wear. The results showed that the decline in pad wear rate was caused by a decrease in the cutting ability as the diamond sharp edges rounded off, while the diamond contact area with the pad surface depended little on the diamond abrasive degradation. Furthermore, we demonstrated the usefulness of measuring the height ratio of the asperity called “top surface area (TSA) ratio” as an appropriate topographical parameter to study the dependence of COF during conditioning on the decay of the removal rate. We present a clear correlation between the pad wear rate degradation and the TSA ratio after CMP to elucidate an increase in the conditioning COF related to truncation of the asperity tip by wear and plastic deformation.

Mobile-Ion-Induced Charge Loss Failure in Silicon–Oxide–Nitride–Oxide–Silicon Two-Bit Storage Flash Memory

In silicon–oxide–nitride–oxide–silicon (SONOS) 2-bit storage flash memory, we discovered deterioration of data retention (DR) in the form of charge loss, which is dependent on the distance between contact windows and word lines (WLs) and also on the thermal treatment performed after the formation of contact windows. We hypothesized that the unique structure of the SONOS flash memory leads to susceptibility to mobile ion contamination. We concluded that Na mobile ion contamination originates in the tungsten chemical–mechanical polishing (W-CMP) process, and that the ions diffuse through the boundaries of the boron phosphosilicate glass (BPSG) and stacked oxide–nitride–oxide (ONO) films into the cell area. We successfully reduced the charge loss by cleaning of the contamination source and by the stable control of phosphorus concentration at the bottom of the BPSG. As a permanent countermeasure, we proposed the complete isolation of contact windows from the adjacent ONO layer, and we were able to demonstrate the effectiveness of this proposal.

Local-Topography-Induced Defects during Tungsten Chemical Mechanical Polishing and Their Impact on Back End of Line

As very large scale integrated fabrication technology continuously scales down, the requirements for contact tungsten chemical mechanical polishing (CMP) become more and more stringent. In this paper, we reported newly observed surface particle defects which were confirmed to be slurry particles or residues. These particles were trapped among W plugs and were difficult to be removed with existing cleaning methods and chemicals. Consequently, these particles may give rise to shorts between metal 1 (M1) and contact plugs or cause severe reliability problems, which have been directly revealed by failure analysis. A mechanism of defect-induced short was proposed. In addition, defect counts were found to have a strong correlation with tungsten plug protrusion. With the reduction of plug protrusion, defect counts decrease significantly. Oxide buff is found to be able to effectively control the local plug protrusion. However, due to the inversely proportional relationship between oxide buff and erosion, erosion may get worse if the oxide buff amount is reduced for the purpose of diminishing plug protrusion. An optimized topography before the oxide buff plays an important role in widening the process margin of the oxide buff. A correlation study of erosion and local protrusion is envisioned to be applicable to other CMP processes.