Taguchi method–ANN integration for predictive model of intrinsic stress in hydrogenated amorphous silicon film deposited by plasma enhanced chemical vapour deposition T.B. Asafa n , N. Tabet, S.A.M. Said Center of Research Excellence in Renewable Energy, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia. article info Article history: Received 16 April 2012 Received in revised form 7 October 2012 Accepted 7 October 2012 Communicated by V. Palade Available online 19 November 2012 Keywords: Intrinsic stress ANN PECVD Hydrogenated amorphous silicon Taguchi method abstract An integration of Taguchi method and artificial neural network (ANN) technique for the prediction of intrinsic stresses induced during plasma enhanced chemical vapor deposition (PECVD) of hydrogenated amorphous silicon (a-Si:H) thin films is presented. Inputs to the ANN model are plasma power, hydrogen dilution ratio, chamber pressure and substrate temperature. Ninety-two data points were used for the network training, model validation and testing in a 2:1:1 relative proportion. An optimized model with a network architecture of 4-5-3-1, a Levenberg-Marquardt training algorithm and a learning rate of 0.1 was obtained from L 9 (3 4 ) orthogonal array based on Taguchi approach. By using the optimized network, parametric studies were conducted to show how the intrinsic stresses are influenced by the deposition parameters. Analysis of variance (ANOVA) of the ANN variables indicates that the first hidden layer is the most significant parameter contributing about 39% to the changes in the network mean square error (MSE) while the second hidden layer contributes about 15%. Accuracies of the predictive model are within 72.5% and 713% error bound for compressive and tensile stress regimes, respectively. Also, results of the parametric study show a clear trend between the deposition parameters and the resulting intrinsic stresses, and are found to agree with published data. The results are discussed in the light of physics of PECVD process. & 2012 Elsevier B.V. All rights reserved. 1. Introduction In the past few decades, hydrogenated amorphous silicon (a-Si:H) thin films have been extensively studied because of their several interesting properties and potentials [1–5]. They have large optical absorption in the visible region, making them potential absorber materials in solar cells and thin film transis- tors [1]. Their reduced bulk recombination and higher fill factors enhance their applications as active materials in solar cells [3]. Hydrogenated amorphous silicon thin films can be deposited on various substrates like float glass, metal foils, flexible and stretch- able polymers. These substrates are important for low-cost and large scale terrestrial applications [2]. a-Si:H films can be depos- ited by using chemical vapor deposition (CVD), plasma enhanced chemical vapor deposition (PECVD) and sputtering techniques [2–4]. Among these techniques, PECVD is commonly used because of its high throughput coupled with enhanced uniformity and conformity of the deposited films. Depending on the deposition technique, the structural, electronic, optical and mechanical properties of a-Si:H films are influenced by the deposition para- meters such as plasma pressure, frequency and power, substrate material, deposition temperature and hydrogen dilution ratio among others [4]. One of the most important mechanical proper- ties of these films is intrinsic or residual stress. Intrinsic stress significantly affects the carrier mobility as well as carrier lifetime [5] and consequently determines the efficiency of a-Si:H based solar cells. Excessive stress at either compressive or tensile regimes affects their fracture toughness [6,7] by inducing dela- mination and dislocations [7] which in turn promotes device failure. A significant number of a-Si:H films cracked at a tensile stress greater than 387 MPa or bucked at a compressive stress higher than 925 MPa [4]. Generally, stress in thin film can originate from thermal, intrinsic and extrinsic sources or a combination thereof. The presence of thermal stress is due to the difference in the thermal expansion coefficients of the film and the underlying substrate, and the temperature difference between the film deposition and the subsequent cooling states [8,9]. Depending on these differ- ences, contribution from thermal stress can be significant or negligible. Extrinsic stress is usually due to external effects where water molecules penetrate open voids and pores in partially dense films. The interactions between the polar species and the Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/neucom Neurocomputing 0925-2312/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.neucom.2012.10.019 n Corresponding author. Tel.: þ32470601889. E-mail address: g200902430@kfupm.edu.sa (T.B. Asafa). Neurocomputing 106 (2013) 86–94