Materials
Banana (Musa paradisica) fruits of good quality and ready for ripening was purchased from the local retail (Dinajpur, Bangladesh) and allowed to ripen under natural conditions. The yellowish colored banana was selected for conducting this research. The refined sugar (Fresh refined sugar, United Sugar Mills Ltd., Narayanganj, Bangladesh) was used for preparing the osmotic solutions.
Preparation of raw materials The osmotic solutions were prepared by dissolving required quantity of refined sugar in distilled water (w/v). The total soluble solids (oBrix) content of the solution was measured using a portable digital refractometer (Model-HI96801). Consequently, the banana was washed in clean water to remove the dirt particles and peeled manually. Then the samples were sliced into different thickness using a stainless steel knife and the thickness was measured by a vernier caliper. The initial moisture content of ripe banana slices were determined using a hot air oven at 105oC for 24 h (Sutar et al., 2012; Abano, 2016). The initial moisture content of the ripe banana samples varied from 75.2% to 77.4% (wet basis).
Experimental design The microwave assisted osmotic dehydration (MWOD) experiments were conducted using a face centered central composite design (FCCD). The four factors are sample thickness (A), sugar concentration (B), microwave power (C), and contact time (D) were taken at three level. The variable levels were chosen on the basis of previous experiments conducted by several authors (Chavan et al., 2010; Athmaselvi et al., 2012; Azarpazhooh and Ramaswamy, 2012). The experimental design included thirty experiments with sample thickness in the range of 5-10 mm, sugar concentration in the range 40-60oBrix, microwave power in the range 100-1000W, and contact time in the range 10-50 min. The sample to solution ratio of 1:30 was kept constant throughout all the experiments (Wray and Ramaswamy, 2013). The original values of all variables in terms of coded and actual units. All these variables were closely controlled and accurately measured during experimentation.
Experimental procedure Osmotic dehydration of banana slices was carried out in a batch system. The experiments were conducted at various combinations of slice thickness, sugar concentration, microwave power, and time as per the experimental design. The glass beaker containing the osmotic solution and banana slices was placed inside the microwave cavity in a sample to solution ratio of 1:30. During experimentation, agitation was not performed because it increases the cost of processing. After completion of total drying time as per experimental design, samples were withdrawn from the solution, quickly rinsed, gently blotted with paper towel to remove adhering solution, and then analyzed. Experiments were run in triplicate and the data are given an average of these results. The reproducibility of the experiments was within the range of ±1.63%. The water loss (WL) and solid gain (SG) were determined using a mass balance method as given by El-Aouar et al. (2006).
Drying efficiency: Drying efficiency is defined as the energy required to evaporate a unit mass of water (MJ/kg). Drying efficiency was calculated as described by (Yongsawatdigul and Gunasekaran, 1996).
Color difference Total color difference (ΔE) of fresh and osmotically dehydrated banana slices was measured with a Hunter Lab color meter (USA), Model 45/0-L. The instrument was calibrated using standard white tile before taking measurements of each sample. The color of fresh and dehydrated banana slices was assessed in terms of 'L', 'a', and 'b' after making a paste of the sample. For determining the color of the dehydrated sample, the paste was completely filled in Petri dish hence no light can pass during the measuring process. The obtained values were recorded and compared with the values of fresh banana sample. The color difference (ΔE) was determined by the equation given by (Giri and Prasad, 2007).
Statistical analysis Response surface methodology was used to determine the relative contributions of A, B, C and D to various responses under study such as water loss (WL), solid gain (SG), drying efficiency (DE), and color difference (ΔE) of banana slices. The second order polynomial response surface model was fitted to each of the response variables (Yk).
The experimental design was acquired from FCCD. The experimental design and data analysis was performed using a commercial statistical package, Design Expert, version 7.0 (Stat-Ease Inc., Minneapolis, MN). Analysis of variance (ANOVA) was conducted to fit the model represented by eq. (6) and to examine the statistical significance of the model terms. The adequacy of the model was assessed using model analysis, lack-of-fit tests, and R-squared values (R2 and Adjusted R2). The coefficient of variance (CV) was calculated to find the relative dispersion of the experimental points from the predicted values of the model. Response surfaces and numerical optimization were generated using the same software.