Partial diagnostic data to plasma etch modeling using neural network

Abstract-This paper presents work carried out with data from an industrial plasma etch process. Etch tool parameters, available during wafer processing time, are used to predict wafer etch rate. These parameters include variables such as power, pressure, temperature, and RF measurement. A number of variable selection techniques are examined, and a novel piecewise modelling effort is discussed. The achievable accuracy and complexity trade-offs of plasma etch modelling are discussed in detail.

show abstract

“…For the training set, the output of the network is calculated. The sum squared error E is defined as [14] …”

Section: B Artificial Neural Networkmentioning

confidence: 99%

Virtual metrology for plasma etch using tool variables

Lynn

Ringwood

Ragnoli

et al. 2009

2009 IEEE/SEMI Advanced Semiconductor Manufacturing Conference

View full text Add to dashboard Cite

show abstract

“…However, anisotropy was shown to have a high dependency on RF power. Kim and Kim [73] achieved similar levels of accuracy for etch profile angle, using a reduced set of OES lines, in preference to manipulated inputs, with PCA-OES models also performing poorly. Kim et al [74] successfully produced an ANN model that predicted the discrepancy in sidewall bottom etch rate compared to center etch rate, using genetic algorithms to optimize the spread values.…”

Section: ) Statistical Analysismentioning

confidence: 86%

“…The compression ANN returned seven features while PCA returned five, with comparable results for both, giving prediction errors as low as 0.2%. Kim and Kim [73], however, reported a drastic performance improvement with partial OES models (110 wavelengths) compared to conventional PCA-OES reduction.…”

Section: ) Statistical Analysismentioning

confidence: 95%

“…Lee and Spanos [35] reported little success in modelling uniformity using manipulated variables (gas flows, power, pressure, and gap) as inputs, while Kim and Kim [73] found that the addition of dc bias (as an input) only served to reduce the accuracy and increase the complexity, of the model. However, Kim et al [63] reported a prediction error of just 0.4% using pressure, gas flows, and power as inputs.…”

Section: ) Statistical Analysismentioning

confidence: 99%

See 1 more Smart Citation

Estimation and Control in Semiconductor Etch: Practice and Possibilities

Ringwood

Lynn

Bacelli

et al. 2010

IEEE Trans. Semicond. Manufact.

View full text Add to dashboard Cite

Abstract-Semiconductor wafer etching is, to a large extent, an open-loop process with little direct feedback control. Most silicon chip manufacturers rely on the rigorous adherence to a "recipe" for the various etch processes, which have been built up based on considerable historical experience. However, residue buildup and difficulties in achieving consistent preventative maintenance operations lead to drifts and step changes in process characteristics. This paper examines the particular technical difficulties encountered in achieving consistency in the etching of semiconductor wafers and documents the range of estimation and control techniques currently available to address these difficulties. An important feature of such an assessment is the range of measurement options available if closed-loop control is to be achieved.Index Terms-Closed-loop control, plasma etch, run-to-run control, semiconductor wafer, state estimation, virtual metrology.

show abstract

“…MLPs are trained by finding the optimal set of network weights and biases that minimizes the sum squared error (SSE), defined [23] as…”

Section: B Artificial Neural Network (Anns)mentioning

confidence: 99%

Global and Local Virtual Metrology Models for a Plasma Etch Process

Lynn

Ringwood

MacGearailt

2012

IEEE Trans. Semicond. Manufact.

View full text Add to dashboard Cite

Abstract-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.Index Terms-Gaussian process regression, local modeling, neural network applications, plasma etch, virtual metrology (VM).

show abstract

Partial diagnostic data to plasma etch modeling using neural network

Cited by 11 publications

References 15 publications

Virtual metrology for plasma etch using tool variables

Virtual metrology for plasma etch using tool variables

Estimation and Control in Semiconductor Etch: Practice and Possibilities

Global and Local Virtual Metrology Models for a Plasma Etch Process

Contact Info

Product

Resources

About