Thickness-dependent non-Fickian moisture absorption in epoxy molding compounds K.J. Wong a, ⁎, K.O. Low b , H.A. Israr a , M.N. Tamin a a Centre for Composites, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor Bahru, Malaysia b Centre for Advanced Materials and Green Technology, Faculty of Engineering and Technology, Multimedia University, Jalan Ayer Keroh Lama,75450 Melaka, Malaysia abstract article info Article history: Received 10 June 2016 Received in revised form 27 July 2016 Accepted 21 August 2016 Available online 27 August 2016 The objective of this research is to characterize the relationship between the moisture uptake behavior and the thickness in epoxy-based molding compounds (EMCs). Experimental results from the literature were adopted for this purpose. A thickness-dependent moisture uptake model was proposed to describe the moisture uptake behavior. In order to apply the model, a methodology to develop the fictitious Fickian curve was suggested. Sub- sequently, the relationships between the non-Fickian parameters and the thickness were correlated and com- pared. Results showed that the apparent diffusivity of the fictitious curve was sensitive to the environmental conditions but not the thickness. In addition, when combining all data, it was found that each normalized non- Fickian parameter could be described by a single equation with respect to the normalized thickness. Based on the thickness-dependent model, the moisture concentration across the thickness was further characterized. In conclusion, the model proposed in this study allows the prediction of moisture uptake behavior at various thick- nesses of EMCs. This could greatly reduce the time and cost of extensive experimental works. © 2016 Elsevier Ltd. All rights reserved. Keywords: Thickness-dependent Non-Fickian Moisture Absorption Molding compound 1. Introduction Moisture has been recognized as one of the major factors that con- tribute to the microelectronics packages failure [1]. Moisture could plas- ticize and reduce the glass transition temperature, T g of the epoxy molding compounds (EMCs) [2]. In addition, hygroscopic swelling due to moisture absorption could lead to cracking of the material [3]. It is thus important to quantify the moisture absorption and its related ef- fects on the performance of microelectronics packages to ensure their reliability. Moisture characterization for materials used in microelectronic packaging has been carried out by many researchers [2–25]. In most of the cases, non-Fickian behavior was observed. Consequently, Fick's law [26] was not capable in fitting those experimental data and different approaches have been proposed to describe the non-Fickian behavior. These included dual-stage Fickian [3,5,6,8–11], Langmuir [5,6], power variation diffusivity [16,17], concentration dependent diffusivity [14, 19] and temperature-concentration dependent diffusivity [15]. Re- searchers have shown good fits between their experimental results and proposed models. When non-Fickian behavior was observed, the specimens needed more time to reach saturation compared to the Fickian ones. Conse- quently, additional time and cost were needed for moisture absorption tests. For example, weight measurement for a 2 mm-thick EMC at 60 °C/ 60%RH could take up to more than 100 days but saturation was still not attained [5,6]. Hence, there is a need to devise a solution to reduce the measurement period. Common approaches to accelerate moisture absorption test include increasing the temperature and relative humidity. However, the recom- mendation by JEDEC standard for temperature and relative humidity are 30 °C–85 °C and 60%RH–85%RH, respectively [27]. Even at the upper limit 85 °C/85%RH, saturation is not easily obtained. Using EMCs at 1.4 mm and 3 mm thickness as an example, saturation was not attained even after conditioned for almost one month [3,8]. Further- more, if the activation energy for moisture diffusion and solubility are to be measured, it is essential to carry out the moisture absorption test at a minimum of three temperatures, with interval of 20 °C–30 °C [27]. In this situation, the duration needed for the moisture absorption test would be much longer. Another approach could be by reducing the thickness of the speci- mens. JEDEC recommends 0.3 mm–1 mm to shorten the test period [27]. However, thickness-independent moisture absorption behavior is only applicable to materials that exhibit Fickian diffusion curve [28]. For materials that show non-Fickian moisture absorption behavior, the thickness of the specimens has to follow the same thickness in their intended applications. In view of this, there is a need to develop a meth- odology to characterize the moisture uptake behavior in materials that exhibit non-Fickian behavior at reduced duration. In this study, the moisture uptake behavior of EMCs were character- ized using a thickness-dependent moisture absorption model. The data used was based on the experimental results published by Fan et al. [2] and Placette et al. [3]. In general, the major components in a convention- al EMC are epoxy as the resin, hardener, filler and flame retardant, Microelectronics Reliability 65 (2016) 160–166 ⁎ Corresponding author. E-mail address: kjwong@mail.fkm.utm.my (K.J. Wong). http://dx.doi.org/10.1016/j.microrel.2016.08.014 0026-2714/© 2016 Elsevier Ltd. All rights reserved. Contents lists available at ScienceDirect Microelectronics Reliability journal homepage: www.elsevier.com/locate/microrel