Sensitivity analysis and optimization of excimer laser ablation for microvia formation using neural networks and genetic algorithms

Higher levels of integration on the chip and package with increasing wiring density require highly repeatable and reliable microvias. To meet future needs of local via densities >5000 vias per cm/sup 2/ and via diameters of <50/spl mu/m, a scanning projection excimer laser ablation process is being developed for build-up and flex substrates. Vias with diameters of 30, 40, and 50 /spl mu/m are ablated in 257mu;m thick DuPont Kapton /spl reg/ E polyimide using an ANVIK HexScan/spl trade/ 2150 SXE excimer laser. A 25-1 fractional factorial experiment is conducted to determine the significance of laser fluence, shot frequency, number of pulses, and vertical and horizontal positions of the debris removal system in the laser tool in affecting ablated dielectric thickness, top via diameter, via wall angle, and via resistance. The complex nonlinear interactions between process set points and responses are empirically modelled using the feed-forward neural networks (NNs) employing the error back-propagation training algorithm. Neural networks encode the functional relationship between the process conditions and responses, and are then used to perform sensitivity analysis to quantify the variation in responses for incremental changes in particular process conditions. In addition, genetic algorithms (GAs) are used to identify optimized recipes for neural network response models. Experimental verification of the optimized recipes is performed to achieve completely open microvias, specific microvia diameters and wall angles, and low via resistance.

[1]  Leonid V. Zhigilei,et al.  A Microscopic View of Laser Ablation , 1998 .

[2]  Seung-Soo Han,et al.  Using neural network process models to perform PECVD silicon dioxide recipe synthesis via genetic algorithms , 1997 .

[3]  Kanti Jain,et al.  Large-area high-resolution lithography and photoablation systems for microelectronics and optoelectronics fabrication , 2002, Proc. IEEE.

[4]  J. Brannon,et al.  Excimer-laser ablation and etching , 1990, IEEE Circuits and Devices Magazine.

[5]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[6]  Optimization of nanocomposite integral capacitor fabrication using neural networks and genetic algorithms , 2002, 27th Annual IEEE/SEMI International Electronics Manufacturing Technology Symposium.

[7]  Keshav Prasad,et al.  Multilevel thin film packaging: applications and processes for high performance systems , 1994 .

[8]  Gary S. May,et al.  Advantages of plasma etch modeling using neural networks over statistical techniques , 1993 .

[9]  J. S. Hunter,et al.  Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building. , 1979 .

[10]  Gary S. May,et al.  An Object-Oriented Neural Network Simulator for Semiconductor Manufacturing Applications , 2006 .

[11]  G.S. May,et al.  Manufacturing ICs the neural way , 1994, IEEE Spectrum.

[12]  D. E. Goldberg,et al.  Genetic Algorithms in Search , 1989 .

[13]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[14]  G. S. May,et al.  Neural network modeling of the resistance of metallized vias formed by laser ablation in polymer dielectrics , 2003, Proceedings of the 15th Biennial University/Government/ Industry Microelectronics Symposium (Cat. No.03CH37488).

[15]  Gary S. May,et al.  Modeling the Laser Ablation Process Using Neural Networks , 2003 .

[16]  J.F. Frenzel,et al.  Genetic algorithms , 1993, IEEE Potentials.