학술논문

Modeling and Optimization of the Deposition of Shape Memory Polymers for Information Storage Applications
Document Type
Periodical
Source
IEEE Transactions on Semiconductor Manufacturing IEEE Trans. Semicond. Manufact. Semiconductor Manufacturing, IEEE Transactions on. 22(3):409-416 Aug, 2009
Subject
Components, Circuits, Devices and Systems
Engineered Materials, Dielectrics and Plasmas
Shape
Polymers
Nitrogen
Thickness measurement
Analysis of variance
Process design
Neural networks
Predictive models
Testing
Genetic algorithms
MEMS
nanotechnology
neural networks
optimization
shape memory polymers
Language
ISSN
0894-6507
1558-2345
Abstract
Shape memory polymers are of interest as high-capacity information storage media. This paper seeks to understand the effects of processing conditions on diethylene glycol dimethacrylate (DEGDMA) and bisphenol A ethoxylate dimethacrylate. Full factorial experiments are performed to characterize the impact of the following parameters: spin speed, spin time, and nitrogen flow rate. A total of ten experiments are conducted. The measured responses are film thickness, uniformity, hardness and modulus. Analysis of variance reveals the above input parameters are significant with respect to the output responses. The full factorial experiment is augmented by a central composite face centered (CCF) design to facilitate process modeling. Neural network models are developed to examine relationships. The average predictability of the models is better than 2% for training and less than 15% in testing. Genetic algorithms are used in optimizing recipes for the two materials.