In this study, 2^4 full factorial design of experiment was applied to an FEM model, which gives pattern dependant contact pressure distribution, to build a qualitative model of pad effects on within die non-uniformity (WIDNU) in CMP. Analysis of variance showed that every single effect and two-way interaction effects of hard layer stiffness and soft layer stiffness are significant compared to the round-off error from mesh change. Various regression models were built and residual analyses were done for each of them. Best model with best normality of residuals includes only the effect of hard layer stiffness, soft layer stiffness and hard layer thickness. Based on this model, basic qualitative design rule for a stacked CMP pad to minimize WIDNU were suggested.

A new chip scale model integrating pad height distribution and it’s interaction with topography on a patterned wafer was tested. Pad asperity height distribution was used to calculate mean contact pressure at a single asperity contact region. Material removal by a single asperity was evaluated from Hertzian elastic contact model and abrasive indentation model. Simulation on a test pattern predicted relatively higher removal rate and lower planarization efficiency with higher nominal down pressure. Oxide thickness variation over a test chip for a time period measured from specially designed test structure matched well with the model prediction.

Local material removal rate is inversely proportional to the local pattern density in ILDCMP. With the assumption that the local velocity and Preston coefficient are constant across a die, local MRR non-uniformity is attributed to the local pressure non-uniformity. Pressure distributions on two different test patterns consisting of five different pattern density sections were calculated with a finite element model. These pressure distributions were compared with semi-empirical pattern dependant oxide CMP model with three different weighting functions. Results showed that pressure distribution can be well approximated with the pattern density dependant oxide CMP model.

With increases in the use of multi-layered materials in the aerospace industry to reduce weight while still meeting strength requirements, the study of inter-layer burr formation in drilling of a multi-layered material becomes more important. Inter-layer gap formation due to material bending by drilling thrust force has significant effect on inter-layer burr formation. A finite element model for inter-layer gap formation in a multi-layered material was proposed. A gap formation was initiated by initial difference in elastic bending of layers and developed by plastic deformation of the first layer.

For an optimization of a drilling process to minimize burr formation, control chart or empirical model from design of experiment can be used. However, direct measurement of inter-layer burr is limited experimentally in the case of drilling through a multi-layered material, which is a common process in aerospace industry. A finite element model that can quantitatively predict the inter-layer burr formation from workpiece material properties and process conditions would significantly reduce the cost and time for building an empirical model. In this study, a finite element model using material properties of stainless steel 304L from previous work was applied to simulated burr formation process during drilling of a multi-layered material. Simulation showed inter-layer burr formation along with entrance burr formation. A quantitative prediction scheme of burr size using node displacement tracking for burr thickness and height was presented.

Acoustic emission (AE) is used as a technique to monitor and characterize surface interactions between wafer and pad in a chemical-mechanical planarization (CMP) process for copper. A significant variation in in-situ AE signal was observed during controlled variation of the slurry composition, demonstrating the sensitivity of AE to the unique surface reactions taking place. Two different levels of AE were observed; a high root-mean-square (RMS) signal when polishing H2O2 treated copper, and a low RMS signal when polishing bare copper. The rate of change in the AE signal was also tied to variations in species concentration and pH in slurry.

In this study, surface topographies of different pad samples from varying in-situ conditioning processes and differing pad life stages have been measured by white light interferometer. Copper removal rate and within wafer non-uniformity was monitored during pad degradation test. Pad surface height distribution, measured from white light interferometer, was more dependent on the plastic deformation and wear at the retaining ring and wafer contact region than on in-situ conditioning. Also, the overall removal rate in copper CMP was more dependent on the slurry flow than on pad surface height distribution.

Chemical mechanical planarization (CMP) of Cu interconnects is a critical bottleneck technology for semiconductor manufacturing at the 65 nm technology node and beyond. The pad condition in CMP is of utmost importance in maintaining acceptable production quality and throughput, as it directly affects many different parameters in CMP including material removal rate (MRR) and process uniformity. Acoustic emission (AE) is a proven technique for in-situ monitoring of a wide range of manufacturing processes, and has been demonstrated for the in-situ monitoring of various phenomena in the CMP process such as endpoint and MRR detection. A novel graphical mapping approach of sensor signal for monitoring pad condition during Cu CMP is proposed and tested, with AE demonstrating improved sensitivity over that of friction force.

CMP has been successfully used for shallow trench isolation (STI) process in front end semiconductor processing. In STI process, nitride covers active device area acting as the CMP stop layer. Amount of nitride erosion during CMP is directly related to the step height of oxide isolation structure after removing nitride, and affects device performance as well. As semiconductor technology node is decreasing beyond 90nm, variation of nitride erosion over a chip in CMP process should be controlled within tight specifications. A model based simulation tool for chip scale variation of nitride erosion would improve not only CMP compatible STI layout design but also help in optimization of high density plasma chemical vapor deposition (HDPCVD) process of oxide layer. In this paper, a chip scale modeling scheme for better prediction of nitride erosion map integrating HDPCVD oxide topography is presented. In addition, a simulation result for a specially designed test pattern resembling production level STI layout is presented.

With increases in the use of multi-layered material in the aerospace industry to reduce weight while still meeting strength requirements, studying inter-layer burr formation in drilling of a multi-layered material becomes more important. Inter-layer gap formation due to material bending by drilling thrust force has significant effect on inter-layer burr formation. A finite element model for inter-layer gap formation in a multi-layered material was proposed. A gap formation was initiated by initial difference in elastic bending of layers and developed by plastic deformation of the first layer. Influence of clamping location on gap size was also investigated.