Chemical Mechanical Planarization Model Research Articles

Noise Reduction of Atomic Force Microscopy Measurement Data for Fitting Verification of Chemical Mechanical Planarization Model

In advanced integrated circuit manufacturing processes, the quality of chemical mechanical flattening is a key factor affecting chip performance and yield. Therefore, it has become increasingly important to develop an accurate predictive model for the chip surface topography after chemical mechanical flattening. In the modeling process, the noise problem of atomic force microscopy measurement data is relatively serious. To solve this problem, the noise characteristics of atomic force microscope measurement data for chip surface topography in this field are studied and discussed in this paper. It is found that the noise present in such problems is mainly triggered by the vibration and tilt of the probe. Two types of noise, low-frequency and high-frequency, are presented in the time domain. In order to solve the noise problem in this modeling data, this paper analyzes the spectral characteristics of the measurement data using Fourier transform, and a wavelet-Fourier transform composite noise reduction process is proposed. The algorithm is applied to the noise reduction of the chip surface data of 32 nm copper interconnect process. The noise reduction results were compared with scanning electron microscope photographs to verify the effectiveness of the noise reduction.

A Physics-Based Chip-Scale Surface Profile Model for Tungsten Chemical Mechanical Planarization

In this work, a new physics-based chip-scale surface profile model is proposed to focus on investigating the influence of the design pattern effects on the tungsten surface topography in the chemical mechanical planarization (CMP) process. Due to its significance of the contact pressure on CMP planarity simulation, a two-scale contact pressure computation method is constructed to obtain an accurate pressure distribution between the wafer surface and the polishing pad. First, chip-scale contact mechanics-based global pressure has been introduced to capture the long range height variation of W CMP caused by deposition and polishing processes. Then feature-scale pattern dependent effect is considered to accurately calculate the local contact pressure in constructing the final removal rate formula and achieving chip surface profile simulation. The calculated local contact pressure is further integrated with the fundamental of steady-state oxidation reaction to construct a new material removal rate model, which systematically captures the effects of mechanical abrasion and concentration of chemical reagent on the polishing rate. The model prediction results are consistent with the collected experimental data in predicting the dishing effect at different slurry conditions. The simulated surface topography and polishing removal rate characteristics post tungsten CMP indicate prominent design pattern-dependent effects. Therefore, the present W CMP model can be utilized to assist in analyzing the influence of the design pattern effects on the wafer surface topography and performing sensitivity analysis of design parameters on the surface planarity. It can also be readily incorporated into a design for manufacturability flow to form a chip-scale planarity simulator to detect the hotspots of the entire design layout.

A Material Removal Rate Model for Tungsten Chemical Mechanical Planarization

In this work, a new tungsten removal rate model is proposed in the chemical mechanical planarization (CMP) process to investigate the removal mechanism with consideration of the synergistic effect of chemical reaction and mechanical abrasion. Based on the fundamentals of steady-state chemical reaction and mechanical abrasion, a chemical reaction kinetics CMP model is first built up to relate the removal rate to the chemical reagent and mechanical rate parameters. Then the Greenwood-Williamson (GW) contact theory is introduced into the chemical reaction kinetics model to construct a closed-form equation of removal rate, which captures the synergistic coupling effect of chemical and mechanical interactions. Furthermore, the present model is verified by the collected experimental data and utilized to investigate the impact of the design pattern effects on the removal rate. The consistency of the model prediction and the experimental data as well as the removal characteristics of design pattern structures indicate that the new proposed tungsten CMP model can be adopted to elucidate the synergistic effect of chemical and mechanical interactions and perform the sensitivity analysis of the design pattern dependency on removal rate of tungsten films.

Comparative Analysis of Chemical Mechanical Planarization Performance on Low-K Films

Mirroring the pursuit for ever-shrinking feature size in memory architectures, the need for materials capable of preventing crosstalk between increasingly closer conducting lines has led to the replacement of silicon dioxide with alternative low-k materials. Within this subset of materials exist films with notably low dielectric constants; however, these materials present unique challenges for chemical mechanical planarization (CMP) in comparison to more traditional dielectrics, the least of which include undesirable scratching and erosion (e.g. dishing) of low-lying areas. Fortunately, some degree of tuning of film properties (e.g. modulus and hardness) is possible through various deposition methods and post treatment conditions, among others. In this study, two slurries containing common Ceria abrasives were utilized to examine a selection of films of varying properties and preparations across several CMP process conditions. A structural test vehicle having a series of feature sizes designed to feed CMP modeling was used for analysis. Static etch rate and material removal rate were measured using optical metrology. Dishing was quantified via atomic force microscopy (AFM), and scratching performance was analyzed via optical methods. Using the collected data, a mechanism for removal was proposed for each slurry. Figure 1

A Neural Network-Based Approach to Material Removal Rate Prediction for Copper Chemical Mechanical Planarization

In modern integrated circuits, chemical mechanical planarization (CMP) has emerged as one of the most important solutions for surface global planarization. The surface uniformity and quality control are greatly dependent on the material removal rate (MRR) of the polished wafers. The construction of accurate physical CMP models to predict MRR is a great challenge due to the complexity of the coupling interplay of mechanical, chemical and design pattern effects in the CMP process. In this work, CMP experiments are designed and performed under different process conditions to obtain the removal rates and a data-driven neural network-based approach is developed to predict the MRR and reveal the relationship between the removal rate and polishing parameters for copper CMP. It is shown that the predicted results of the removal rates are consistent with the experimental data with an optimized network structure. The investigation of the neural networks (NNs) to model the MRR indicates that the NN-based method can provide a general way of capturing the removal rate profiles regardless of the complexity of the polishing mechanism. Therefore, the present CMP model has good potentials of assisting in achieving the surface uniformity control of the copper wafer in semiconductor manufacturing.

A Wafer-Scale Material Removal Rate Model for Chemical Mechanical Planarization

In this work, a new wafer-scale material removal rate (MRR) model is proposed to investigate the underlying removal mechanism of wafer surface in the chemical mechanical planarization (CMP) process. Based on the governing equation of the plate theory, chemical reaction kinetics and wear theory, an analytical CMP model has been constructed to capture the influence of mechanical abrasion and chemical reactions on the removal rate with high efficiency. The effect of the elastic modulus of the pad, platen rotational speed, polishing temperature and reactant concentrations on the removal mechanism is investigated in detail to elucidate the intrinsic removal behavior. It is found that the material removal rate is sensitive to the elastic modulus of pad, temperature distribution, and chelator concentration, more sensitive to the oxidizer concentration, and most sensitive to the platen rotational speed. The magnitude of the removal rate can be adjusted significantly by properly selecting a good configuration of these influencing factors. The model predictions of the MRR profiles are consistent with the published experimental data at different polishing conditions. Therefore, the present CMP model can give insights into the wear mechanism of polishing process and may be utilized as a simulation tool for further optimization of removal rate and surface uniform control.

Application of Neural Network-Based Oxide Deposition Models to CMP Modeling

Numerous deposition processes are used in modern semiconductor manufacturing, including high-density plasma chemical vapor deposition (HDP-CVD), spin-on dielectric (SOD), flowable CVD (FCVD), and enhanced high aspect ratio processes (eHARP). Generation of high quality post-deposition surface profiles is crucial for chemical-mechanical planarization (CMP) model building, due to the complex nature of the CMP process and long-range effects in CMP. Measurements show complicated post-deposition surface profile height dependence on the underlying pattern for these deposition processes. While high-quality compact models exist for HDP-CVD and SOD processes, building compact models for FCVD and eHARP is a challenge, since they include several deposition and annealing steps, and show complicated behavior width respect to the underlying pattern. In this paper, we present the application of neural networks (NNs) to post-deposition surface profiles modeling for the above-mentioned deposition processes. Experiments showed that NNs should have at least two hidden layers and 6–10 neurons per hidden layer to capture the complexity of deposited surface profiles. The application of NNs to modeling surface profiles of deposition processes has shown that NNs provide a general approach for modeling surface profiles of deposition processes without long-range effects for CMP modeling, irrespective of the complexity of the deposition process.

A Material Removal Rate Model for Tungsten Chemical Mechanical Planarization

In this work, a new material removal rate (MRR) model for tungsten chemical mechanical planarization (CMP) is proposed to provide some fundamental insights of slurry components and contact behavior on the removal mechanism. The synergistic effect of mechanical abrasion and chemical reactions on MRR is systematically investigated through using concepts of the steady-state equilibrium, chemical growth kinetics and contact mechanics in W CMP process. A closed-form equation of MRR is formulated to capture the influence of the slurry chemistry and operating parameters on the surface removal mechanism. Then the tungsten CMP model is simplified and verified through comparing the simulated results of removal rates with experimental data. It is found that the MRR has complicated variation behaviors depending on the specific model parameters. The consistency of the simulation results and experiments indicates that the present model can be adopted to capture the influences of chemicals, mechanical abrasion and pad properties on the polishing rate and assist in optimizing the surface uniformity of tungsten wafers.

An Oxide Chemical Mechanical Planarization Model for HKMG Structures

In order to predict the surface topography, a chip-scale oxide chemical mechanical planarization (CMP) model is proposed in poly opening polishing processing of high-k metal gate (HKMG) structures. A geometrical chemical vapor deposition model is first constructed to describe the influence of design pattern effects on the initial surface profiles of dielectric oxides. Based on contact mechanics, the global pressure distribution is then computed to incorporate the long-range features of design structures and polishing pad deformation. The global pressure is further combined with local pattern effects to construct a feature-scale material removal rate model, which can be utilized for surface topography simulation of design chips. The predicted results of the proposed CMP model are consistent with the experimental data collected from the test patterns of HKMG CMP process. Consequently, the oxide CMP model has good potentials of predicting the surface planarity of polishing wafers and optimizing some design rules for HKMG design layouts.

A Chemical Mechanical Planarization Model Including Global Pressure Distribution and Feature Size Effects

In this paper, we propose a chip-scale Chemical mechanical polishing (CMP) model based on the Greenwood–Williamson theory combining the effects of both the pad bulk deformation and the feature size, which are mainly responsible for global pressure distribution and feature-scale removal rate variation, respectively. A reference camber is first constructed to simulate the pad bulk deformation, which can be directly introduced into the global pressure distribution computation for incorporating the long-range characteristics of chip profiles. Then, both the effective curvature and the modified distribution function of pad asperity height are involved in the material removal rate. The model is first verified through some experimental data from CMP test structures and found to be in good agreement with the test data. For further validating the rationality and accuracy of the present model, some simulations are compared with our experimental data and discussed in detail.

A chip-scale chemical mechanical planarization model for copper interconnect structures

In this work, a new chip-scale chemical mechanical planarization (CMP) model for copper interconnect structures is proposed to capture the design pattern effects, abrasive removal mechanism, and evolution of surface profiles and assist in CMP simulation for planarity detection and optimization. To physically describe the abrasive mechanism and surface thickness variation of copper, the global pressure distribution induced by morphologies of pattern structures and the coupled removal mechanism of chemical kinetics and mechanical abrasion are integrated into a full model framework to predict the surface topography of the design structures. Two sets of testkeys including different types of line width/space arrays covering the design rules of 45nm and 40nm technology nodes are designed to verify the accuracy of the present model. Compared with the experimental data, the model prediction results show that the current method can capture well the experimental behaviors and be utilized for analyzing the influential factors on the copper removal rate and predicting post-CMP copper surface topography. It also indicates that our physically based chip-scale Cu-CMP model can be readily incorporated into design for manufacturability flow to compose of a CMP simulation tool to optimize some basic design rules for evolution control of surface profiles and hotspots detection.

A chemical mechanical planarization model for aluminum gate structures

In this work, we propose a new aluminum gate chemical mechanical planarization (CMP) model to describe the metal gate height variation in 32nm high-k metal gate (HKMG) process based on chemical kinetics and contact mechanics. The effects of mechanical abrasion, concentration of different types of chemical reagents and hydrogen ions, pattern geometry and pad elastic properties on surface profile are physically captured. In the process of constructing the model, the discrete convolution and fast Fourier transform (DC–FFT) technique is integrated with conjugated gradient method (CGM) for calculating the contact pressure between the wafer surface and the polishing pad. Then the computed pressure distribution is introduced into the new constructed chemical kinetics formula to determine the local removal rate of the underlying patterns and predict the evolution of the wafer surface topography. The detailed relationship between the metal gate dishing post-Al-CMP and the design pattern structures are systematically investigated. The model agrees reasonably well with the experimental data measured from the HKMG test structures. Therefore, it can be utilized for quantifying the effect of pattern geometry on dishing, predicting the wafer surface height evolution and optimizing some design rules for manufacturability to improve the surface planarization of HKMG structures.

A modifying algorithm of the topological VLSI layer by dummy filling features based on modeling the chemical-mechanical planarization

The approach to modeling chemical-mechanical planarization (CMP) based on the notion of the effective filling density, which makes it possible to calculate the thickness distribution of the interlayer dielectric (ILD) on the surface of the VLSI crystal, is considered. The polynomial CMP model previously proposed in [1] is compared to model [2]. The algorithm of modifying the pattern of the VLSI topological layer by the dummy filling features, which was proposed in [3], is described using the proposed polynomial model. The results of model investigations of applying the developed modification algorithm of topological layers by examples of various VLSI types are presented.

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.

A method for characterizing the pad surface texture and modeling its impact on the planarization in CMP

Chemical-mechanical planarization (CMP) is one of the most demanding process steps in interconnect integration. Therefore, with respect to the pad roughness, we systematically characterize and model the planarization of special CMP test chips, which emulate integrated circuit (IC) layouts. Therefore, a novel pad roughness characterization methodology is developed and used for the extraction of important pad surface parameters like the mean asperities radius of curvature and the asperities size distribution. The obtained pad surface data is used for the derivation of a novel chip scale CMP model based on the Greenwood-Williamson theory. It is validated by experimental data from CMP test structures containing variations of both pattern-density and pattern-size and describes the wafer topology evolution with high accuracy throughout the planarization process, indicating a strong impact of the asperities size distribution on the planarization in test chip areas having trench widths smaller than the mean asperities radius of curvature.

Electrochemical investigation of copper passivation kinetics and its application to low-pressure CMP modeling

During the process of chemical mechanical planarization (CMP) of copper interconnection in ultra large scale integration (ULSI), copper passivation plays a critical role in material removal. The kinetics of copper passivation in glycine solutions containing BTA was studied by the chronoamperometry technique. The results showed that the current density transients followed with a double-exponential decay, including both non-faradaic double layer charging and faradaic reaction effects. Furthermore, a model based on the passivation kinetics was proposed for low-pressure CMP. Combining the model and the experimental results, the material removal mechanism was analyzed. The mechanical effect dominated the material removal at pH 4, while it was a chemical dominant process at pH 10.

Chemical Mechanical Planarization from Macro-Scale to Molecular-Scale

Chemical mechanical planarization (CMP) has been an essential method to fabricate wafer surfaces in IC industry. This article provides a relatively comprehensive review on the state-of-the-art and recent progress in the modeling of CMP and material removal mechanism, and addresses the limitations and further research directions of the modeling work and material removal mechanism.

A CMP Model Including Global Distribution of Pressure

In this paper, we consider the chemical-mechanical planarization (CMP) of adjacent line-space structures with different pattern density. Such structures are used on test-chips for process characterization and provide similar features like typical DRAM layouts. The predictions of previous chip-scale models based on a local balance of forces for this case are scrutinized and it is found that the distribution of the applied pressure between the different density regions is not treated sufficiently accurate in these models. We propose a simple way to include the global force balance and show that the resulting model is able to capture behavior which is more in line with general expectations for the time-evolution of typical CMP processes. The model is calibrated with data obtained for a set of test-structures with varying density and found to give a better match compared to the models described above. Especially, the prediction of the total indicated range, which is a measure for the global planarity after CMP is improved. Additionally, a parametric study shows that qualitatively different removal rate diagrams that confirm engineering experience can be obtained with the new model.

Study on Stiffness and Conditioning Effects of CMP Pad Based on Physical Die-Level CMP Model

Chemical mechanical planarization (CMP) pad stiffness and conditioning effects were evaluated based on a physical die-level CMP model, with pad effective modulus and asperity height as parameters. In one study, patterned dielectric wafers were polished using polymeric pads of different stiffnesses. In a second study, wafers were polished by standard pads using different conditioning disks. Polishing experimental data (dielectric thickness and step height) were fitted by the physical model, enabling the extraction of the pad effective modulus and asperity height model parameters. A higher pad stiffness gives better within-die uniformity, and the conditioning disk with blocky diamonds achieves up-area only polishing for longer times. Polishing simulations using the physical model reflect a clear pattern density dependence.

A three-step copper chemical mechanical planarization model including the dissolution effects of a commercial slurry

The etching behavior of copper oxide by diluted oxidant-free slurry as a function of temperature was characterized and a three-step copper removal rate model was proposed. Pre-oxidized wafers were exposed to the diluted slurry for times up to 500 s for three temperatures and mass loss was monitored. For the highest temperature, 60 °C, all of the oxide available for reaction was etched in 90 s, however the 25 and 40 °C results did not reach saturation. Measurements up to saturation were used for modeling. A one-dimensional model was proposed where the diffusion of the complexant through a byproduct film found to exist on the wafer surface after etching controlled the process. The model fits the data well with two parameters, E a and A, which were found to be 86.9 kJ mol − 1 and 4.12 × 10 − 2 mol cm − 1 s − 1 , respectively. Similar to a previous copper oxidation study, the rate of copper consumption from dissolution was found to be a function of a characteristic reaction byproduct film thickness. However, the dissolution rates demonstrated a much weaker function of film thickness than the oxidation profiles. For this slurry, the etching process controls the combined oxidation and etching system and static oxidation–dissolution experiments agreed well with the dissolution model. The methodology and modeling developed in this work can be directly applied to other commercial or experimental slurry formulations to quantify the dissolution characteristics of the formulation. Once the static dissolution characteristics of a slurry formulation are quantified, use of the proposed three-step removal rate model will provide a more accurate depiction of the relative chemical and mechanical contributions of a given consumable set.