**3.1 Simulation model**

Food simulant are compound developed to mimic the natural characteristics of food to be packaged. Food simulant are developed to have the exact physicochemical properties of the food under study. After the development, the simulant are then packaged into the packaging film kept under controlled/studied condition for a stated time to determine the interacting characteristics of the food. Example, vegetable oil simulant is used to measure migration into oily foods, 10% ethanol or 3% acetic acid are used for a water based drinks, 50% ethanol solution is used as simulant for butter and other amphiphillic foods. Using simulant for the estimation of migration gives only probable estimated values which are close to the actual values. The migration quantity is determined by evaporating the simulant and then calculating the weight of the remaining residue. Simulant procedure is limited especially on fatty foods due to their difficulty in vaporization. ASTM standard are put in place for the estimation of odor and taste transfer from a packaging material, the standard employed 0.9 M2 test material kept in a required environment for a minimum of 20 hours (**Table 3**) [49].


#### **Table 3.**

*List of some common simulant used for food-package interaction testing [3, 48].*

### **3.2 Mathematical and predictive models**

Mathematical models are predictive equation developed based on a simulation principle to predict the migration in food packaging. The models were developed to determine the diffusion and partition coefficient and explain the interacting substance concentration [50] since it is only when there is partition and when diffusion occurs that packaging materials are said to be loss. Permeation in packaging can also be determine in term of gaseous and moisture transfer rate. Since material permeability depends of the type and characteristics of the packaging film (such as thickness) as well as the condition of the surrounding environment. This is achieved by placing the packaging material in both low and high pressure environment followed by the measurement of pressure difference (both pressure and volume determined) usually expressed as the amount of gas permeated per unit time. Moisture permeation is determined by placing the package in between environment of different humidity rate and estimated by determining the vapor pressure difference in M2 day−1 [51].

#### **3.3 Analytical/chromatographic method**

Chromatographic techniques were employed to determine trace of metal compound that migrate to food surface. Though there is no standard or direct analytical method for the estimation of packaging migrant in a packaging system, different chromatographic method such as GC–MS and LC-UV. Begley [52] uses LC–MS method to determine interacting component in a packaging system. Although not all migrating compound in a packaging system can be detected using conventional method but the application of chromatographic method prove to be effective.

## **3.4 Stochastic/predictive models**

Stochastic models or predictive models are probabilistic mathematical function that provide a prediction of certain level of migration (food-packaging interaction) of packaging materials in a packaging system [53]. Stochastic models takes account the variability and uncertainty as well as the probability of the occurrence of food-packaging interaction. Latin Hypercube Sampling (LHS) and Monte Carlo Models (MCM) are example of stochastic models that gives numerical values of food migration based on numerical distribution data build through simulation method. Stochastic thermodynamics models measure the thermodynamic changes due to food packaging interaction such as temperature, pressure and moisture changes while molecular models of a stochastic methods measure the rate of molecular changes/interaction of gaseous or chemical molecules arises due to disintegration of either packaging or food components and subsequent molecular interaction [8, 54]. Stochastic models such as Life Cycle Assessment (CA) and High-Throughput Risk-based Screening (HTRS) tools are used to determine extent of migration in plastic packaging while empirical Weibull Model has been used to determine the chemical migration curves in paper packaging [55], Mechanistic Models are used to predict the migration of toxic metals into acidic food [56] while diffusion models are used for the determination of migration in ceramic packaging as well as packaging with low migration risk or plastic additives with low diffusion potentials.

#### *Food Preservation Packaging DOI: http://dx.doi.org/10.5772/intechopen.110043*

Predictive models are used to generate data and also been associated with statistic functions to develop software programs that simplify the determination methods for food-packaging interaction. The European Flavors, Additives and food Contact materials Exposure Task (FACET) developed a software for the prediction of probabilistic exposure to chemicals from food or packaging contact materials. The software measure the migration, permeation as well as sorption of flavors, moisture, food additives and packaging surface materials in a packaging system using existing probabilistic data in its database [8], the software consist of data that will be able to study over 6000 substances covering metal coatings, paper and paper boards, inks and adhesive, plastic and plastic coverings as well as components of food flavors and additives [8] with over 600 statistics functions all links to packaging use, composition, application, pack size, storage, environmental conditions and food composition/nature. Measurement is achieved through clustering based on chemical, physical properties (polarity and diffusion) [8] and physic-chemical parameters [57] other modeling software includes; MIGRATEST lite 2000/2001 [53], AKTS-SM by Advanced Kinetics Technology Solutions AG Switzerland, SMEWISE (Simulation of Migration Experiments with Swelling Effect), MULTITEMP, MULTIWISE, and SFPP3 by National Institute for Agricultural Research, and FMECAengine (Failure Mode Effect and Critically Analysis) [53, 57]. With the continues rise of artificial intelligence, the future of predictive models is bright and look at possibility of developing robotic system that will simplify the study of food-packaging interaction and its advent.
