Prediction of WVTR with General Regression Models Kimmo Lahtinen and Jurkka Kuusipalo Tampere University of Technology, Institute of Paper Converting ABSTRACT In this study, general regression models are used to predict moisture barrier properties of extrusion coated papers in different atmospheric conditions. Basically, water vapour transmission rates (WVTR) of different polymers are affected by three factors: coating weight (or squared mass) of a studied polymer, temperature and moisture concentration of surroundings. Regression models find mathematical connections between WVTR and these variables making WVTR analysis even more effective in determining moisture barrier properties of studied polymers. As a result of the study, a practical computer program is established where WVTR of extrusion coated paper is predicted as a function of user-defined temperature, relative humidity and coating weight. INTRODUCTION Moisture barrier properties of extrusion coated papers are mainly controlled by the used coating polymers. The general selection of moisture barrier polymers includes low and high density polyethylenes (PE-LD and PE- HD), polypropylene (PP) and polyethylene terephthalate (PET) [1]. Also, some new polymers have recently been introduced for the same purpose. These include cyclo-olefin copolymers (COC), liquid crystal polymers (LCP), nanocomposites etc. [2-5]. Common to the coatings is that they give a specific water vapour barrier for the material depending on the atmospheric conditions and the used coating weights. Water vapour barrier of extrusion coated paper can be measured with different standard methods [6-9] and is specified as water vapour transmission rate (WVTR). Concerning laboratories, the evaluation of WVTR can be a time consuming procedure. Aiming towards lighter load of testing the target of this study was to establish a mathematical model that is employed as a practical, fast and easy-to-use tool to predict WVTR. Many studies have recently been performed to find prediction models and mathematical expressions for permeability of polymers [10-14]. Also, in terms of food deterioration, many shelf life prediction models have been developed based on the water vapour barrier properties of the package [15-18]. This study represents another expression for WVTR based on regression analysis and computer simulation made for experimental studies. Regression modelling is commonly used in technology when more deterministic models are not efficient due to complexity and disturbances of the problem. Regression analysis establishes sort of a “black-box type” model where the input and output values are not bound to each other with physical laws or scientific facts but the correlation is found via experimental testing and statistical treatment of the results [19]. Mathematical treatment of WVTR starts with Fick’s first law. It states that dx dc D J − = (1) where J = diffusion flow through unit area of film, D = diffusion coefficient, c = concentration of a penetrant and x = distance of the point from the film surface. Concerning WVTR test, the target is to achieve a steady state phase where the diffusion flow of water vapour doesn’t change over time. In addition, the product D*S is called the coefficient of permeation (P) and Henry’s law state that c = S*p, where S = solubility coefficient and p = partial pressure of the penetrant. As a conclusion, we find a simple mathematical expression for WVTR (equation 3) [20-21].