Niall MacGearailt

Global and Local Virtual Metrology Models for a Plasma Etch Process

Virtual metrology (VM) is the estimation of metrology variables that may be expensive or difficult to measure using readily available process information. This paper investigates the application of global and local VM schemes to a data set recorded from an industrial plasma etch chamber. Windowed VM models are shown to be the most accurate local VM scheme, capable of producing useful estimates of plasma etch rates over multiple chamber maintenance events and many thousands of wafers. Partial least-squares regression, artificial neural networks, and Gaussian process regression are investigated as candidate modeling techniques, with windowed Gaussian process regression models providing the most accurate results for the data set investigated.

Improved fluid simulations of radio-frequency plasmas using energy dependent ion mobilities

Symmetric and asymmetric capacitively coupled radio-frequency plasmas in oxygen at 40 Pa, 300 V voltage amplitude and a discharge gap of 40 mm are investigated by means of one-dimensional numerical semi-kinetic fluid modeling on the basis of a simplified reaction scheme including the dominant positive and negative ions, background gas, and electrons. An improved treatment, by accounting for the dependence of ion mobilities on E/N, is compared to the standard approach, based on using zero-field mobility values only. The charged particle dynamics as a result of direct electron impact ionization of oxygen, secondary electron release from the electrodes, the spatial distribution of all involved particles as well as impact of geometry and model modification on ion energies is analyzed and compared to independent simulations and experiments.

Identifying key process characteristics and predicting etch rate from high-dimension datasets

In semiconductor manufacturing advanced process control (APC) refers to a range of techniques that can be used to improve process capability. As the dimensions of electronic devices have decreased, the application of APC has become more and more important for the critical stages of production processes. However, the economic disadvantage of employing APC is that it requires feedback information in the form of downstream metrology data, which is both time consuming and costly to obtain.

Weighted windowed PLS models for virtual metrology of an industrial plasma etch process

Virtual metrology is the prediction of metrology variables using easily accessible process variables and mathematical models. Because metrology variables in semiconductor manufacture can be expensive and time consuming to measure, virtual metrology is beneficial as it reduces cost and throughput time. This work proposes a virtual metrology scheme that uses sliding-window models to virtually measure etch rates in an industrial plasma etch process. The windowed models use partial least squares (PLS) regression and a sample weighting scheme to combat the effects of both process drifts due to machine conditioning and process shifts due to maintenance events. An industrial data set is examined and the weighted windowed PLS models outperform global models and non-weighted windowed models.

Phase-resolved optical emission spectroscopy for an electron cyclotron resonance etcher

Phase-resolved optical emission spectroscopy (PROES) is used for the measurement of plasma products in a typical industrial electron cyclotron resonance (ECR) plasma etcher. In this paper, the PROES of oxygen and argon atoms spectral lines are investigated over a wide range of process parameters. The PROES shows a discrimination between the plasma species from gas phase and those which come from the solid phase due to surface etching. The relationship between the micro-wave and radio-frequency generators for plasma creation in the ECR can be better understood by the use of PROES.

Real-time Virtual Metrology and Control of Plasma Electron Density in an Industrial Plasma Etch Chamber ?

Abstract Plasma etching is a semiconductor manufacturing process during which material is removed from the surface of silicon wafers using gases in plasma form. A host of chemical and electrical complexities make the etch process notoriously difficult to model and troublesome to control. This work demonstrates the use of a real-time model predictive control scheme to maintain a consistent plasma electron density in the presence of disturbances to the ground path of the chamber. The electron density is estimated in real time using a virtual metrology model based on plasma impedance measurements. Recursive least squares is used to update the controller model parameters in real time to achieve satisfactory control of electron density over a wide operating space.

Selecting signature optical emission spectroscopy variables using sparse principal component analysis

Principal component analysis (PCA) is a widely used technique in optical emission spectroscopy (OES) sensor data analysis for the low dimension representation of high dimensional datasets. While PCA produces a linear combination of all the variables in each loading, sparse principal component analysis (SPCA) focuses on using a subset of variables in each loading. Therefore, SPCA can be used as a key variable selection technique. This paper shows that, using SPCA to analyze 2046 variable OES data sets, the number of selected variables can be traded off against variance explained to identifying a subset of key wavelengths, with an acceptable level of variance explained. SPCA-related issues such as selection of the tuning parameter and the grouping effect are discussed.

Real-time virtual metrology and control of etch rate in an industrial plasma chamber

Plasma etch is a semiconductor manufacturing process during which material is removed from the surface of semiconducting wafers, typically made of silicon, using gases in plasma form. A host of chemical and electrical complexities make the etch process notoriously difficult to model and troublesome to control. This work demonstrates the use of a real-time model predictive control scheme to control plasma etch rate in the presence of disturbances to the ground path of the chamber, which are representative of maintenance events. Virtual metrology (VM) models, using plasma impedance measurements, are used to estimate the plasma etch rate in real time for control, with a view to eliminating the requirement for invasive measurements. The VM and control schemes exhibit fast set-point tracking and disturbance rejection capabilities. Etch rate can be controlled to within 1% of the desired value. Such control represents a significant improvement over open-loop operation of etch tools, where variances in etch rate of up to 5% can be observed during production processes due to disturbances in tool state and material properties.

Combination of modeling and simple real-time measurements to control plasma-surface interaction processes

Summary form only given. The increasing complexity in industrial plasma processing demands new strategies for process control and monitoring. The energy transport mechanisms in the interface region between non-thermal low-pressure plasma and surface are of particular importance. Measurements of the “in situ” surface condition, which strongly affects the plasma-surface interaction processes, are extremely challenging. The most promising approach for advanced process monitoring is the active coupling of semi-kinetic simulations and diagnostics.

Prediction of Vacuum Pump Degradation in Semiconductor Processing

Abstract This paper addresses the issue of vacuum pump degradation in semiconductor manufacturing. The ability to identify the current level of vacuum pump degradation and predict the Remaining-Useful-Life (RUL) of a dry vacuum pump would allow manufacturers to schedule pump swaps at convenient times, and reduce the instances of unexpected pump failures, which can incur significant costs. In this paper, artificial neural networks are used to model the current level of pump degradation using pump process data as inputs, and a double-exponential smoothing prediction method is employed to estimate the RUL of the pump.We also demonstrate the benefit of incorporating process data, from the upstream processing chamber, in the development of a solution.