Osmo-air drying of banana slices: multivariate analysis, process optimization and product quality characterization

Abstract

In the present work, osmotic dehydration (OD) was applied as a pretreatment to hot air drying of banana slices and the effect of OD parameters on mass transfer characteristics, color profile was analysed. Principal component analysis (PCA) of OD process revealed that solid gain, weight reduction, water loss and total soluble solids were positively correlated with each other but were found to be negatively correlated with moisture content (MC) of sample. Response surface methodology was used for optimizing the OD of banana slices and and the optimum conditions were 61.26°Brix sucrose concentration at 50 °C for 6 h, resulting in moisture reduction from 75 to 49.78%. PCA-biplot of osmo-air drying (OAD) process showed the association among response parameters, which further revealed a positive correlation of MC with bioactive components. Additionally, OAD samples were also studied for microstructure and Fourier Transform Infra-red analysis. Addition of calcium lactate to sucrose solution resulted in preserving the firmness and bioactive components during osmosis, in addition to fortifying the sample with calcium. The present study provides new possibilities for food industries in preserving the ripe banana and developing calcium fortified functional food products.

Introduction

India is the leading producer of banana in the world with an annual production of about 30.5 million tonnes (FAO 2019). Banana (Musa acuminata) is a high energy, low fat fruit with a plethora of vitamins, minerals, dietary fibres and anti-oxidants. However, it is a climacteric fruit and highly perishable which results in early spoilage (FAO 2019). Convective drying is a popular and simple technique for preservation of such perishable fruits, but it adversely influences the product quality due to the involvement of high temperature, and consequently affects consumer acceptability. The quality loss during drying can be countered with the help of various physical and chemical pre-treatments like blanching, dipping in sodium metabisulphite, citric acid, calcium chloride, ascorbic acid, osmotic solution etc. (Sharma et al. 2015).

Osmotic dehydration (OD) is the partial removal of moisture from a product by immersing it in a hypertonic solution which sets up an osmotic pressure across a semi-permeable membrane, as a result of which water flows out from the sample into the solution. OD has shown positive effect on colour, texture, flavour, product stability, nutrient retention and a significant water loss from banana slices was observed for the first 6 h of osmosis (Jalali et al. 2008). Study conducted by Kar et al. (2005) indicated that processing conditions of osmotic dehydration such as temperature, solution concentration and solution to sample ratio influence the driving force, i.e., osmotic pressure gradient for mass transfer. During the OD of pineapple slices in sucrose solution, water loss was more affected by osmotic temperature, while solution concentration had a more pronounced effect on the solid gain (Lombard et al. 2008). Moreover, sucrose was found to be a better agent for osmotic dehydration of carambola slices when compared to fructose and glucose solution (Ruiz-López et al. 2010).

Incorporation of calcium lactate in sucrose solution during the osmotic dehydration of pineapple slices displayed increased firmness of the samples (Silva et al. 2014). It was further reported that fruit treated with calcium lactate was found to contain about three times more calcium content than that of fruit osmosed with calcium gluconate. Chavan et al. (2010) prepared osmotically dehydrated ripe banana slices by soaking sulphur pretreated banana slices in 60°Brix sugar solution containing 0.1% KMS, 0.1% citrate and 0.2% ascorbic acid, followed by cabinet drying at 55 °C for 10 h to obtain final product having moisture of 18%. The dried products have good sensorial acceptability and microbial stability up to six months’ storage at ambient conditions. Shukla et al. (2018) investigated the mass transfer kinetics during osmotic dehydration of banana and revealed that Fructose being a monosaccharide caused higher water loss and solid gain as compared to sucrose and maltodextrin which are polysaccharides. Further, zero order kinetic model represented the moisture loss and solid gain more accurately than first-order kinetic model.

Jain et al. (2011) studied the effect of osmosis process temperature (30–50 °C), syrup concentration (50–70°Brix) and duration of osmosis (4–6 h) on water loss, sugar gain, and sensory properties of papaya, followed by the optimization of process using response surface methodology. It was observed that osmotic dehydration led to reduce the initial moisture of papaya from 87.5–88.5% to 67.6–81.1% (wb). OD is an effective process for moisture removal from a fruit, but in order to ensure preservation for safe consumption, the moisture content should be reduced to a lower level which can be achieved by successive air drying. Furthermore, food materials subjected to OD prior to air drying reported better quality retention as opposed to their fresh counterparts (Nabnean et al. 2017). One of the important quality indicators of the dried product is rehydration, which involves the restoration of raw material properties by immersing the dried products in water. Additionally, extensive study of rehydration process helps in meeting quality specifications and energy conservation (Rodríguez et al. 2015).

Osmotic dehydration involves simultaneous changes in water distribution, textural properties, quality attributes of the samples and their complex interdependencies, which are difficult to understand using the univariate analysis (Loredo et al. 2013). Similarly, food materials subjected to hot air-drying show changes in their bioactive profile, textural characteristics, color parameters in conjunction with the reduction of moisture content. Multivariable makes it easier to analyse the interaction among the parameters of the process. In this regard, the utilization of multivariate techniques like correlation and principal component analysis could provide a deep visualization of associations among the parameters. Principal component analysis has been successively employed for analysing interactions of volatile compounds with osmotic dehydration process of pineapple (Pino et al. 1998), association of drying methods with volatiles of banana powder (Wang et al. 2007), microwave dehydrated cherry tomatoes with volatile compounds (Heredia et al. 2012), banana fruit quality parameters with maturity (Rajkumar et al. 2012), and dried apple snacks with their shelf life (Saavedra et al. 2013).

It can be stated that osmotic dehydration followed by hot air drying is a potential technique for processing fresh perishable bananas in order to achieve longer shelf life with better quality attributes than conventional drying. Literature review shows that most studies on OD of banana are limited to process optimization (Atares et al. 2011; Da Silva 2017) and mass transfer analysis (Rastogi and Raghavarao 2004; Kar et al. 2005), while few have attempted air drying of the osmosed banana (Fernandes et al. 2006; Oliveira et al. 2006). Incorporation of calcium lactate to osmotic solution could be a simple and useful approach to develop the calcium fortified dried banana samples. However, the effect of calcium lactate on the quality attributes of osmosed and osmo-air dried banana slices is not very clear. Furthermore, very little information is available on interrelation among different processing parameters, moisture loss characteristics, quality attributes of food samples during osmotic and convective drying process. Based on these research gaps, the current study aims to (1) investigate the effect of temperature, osmotic solution concentration and immersion time on the OD of banana slices and optimize the osmotic dehydration process using response surface methodology, (2) to examine the effect of osmotic dehydration with calcium lactate and hot air drying on quality attributes of banana slices, and (3) investigate the interactions among different parameters during osmotic dehydration and hot air drying of banana slices using the multivariate approach.

Materials and methods

Sample procurement and preparation

Fresh ripe bananas (Musa acuminata) of Singapuri variety were procured from the local market of IIT Kharagpur. The initial moisture content of bananas was found to be in the range of 73–77% (wb) while total soluble solids were estimated as 26 ± 0.3°Brix. The bananas were peeled and cut into 0.52 ± 0.06 cm thick (L) slices with the help of a slicer. These slices were further cut with a stainless steel core cutter to maintain a diameter of 2.3 ± 0.14 cm. The weight of one fresh slice of banana was 2.735 ± 0.13 g.

Osmotic dehydration of banana slices

Process description

The OD of banana slices was carried out by immersing them in sucrose solutions of different concentrations, C (35, 50, 65°Brix) in polyethylene terephthalate jars maintaining a fruit:solution ratio of 1:5 (w/w). The jars were placed in a constant temperature water bath (SD instruments and Instruments, India) and water was maintained at different temperatures, T (30, 40, 50 °C) for varying immersion time, t (1, 2, 3, 4, 5, 6 h). After the osmotic treatment, the samples were lightly wiped with tissue paper to remove adhering solution and were weighed. To analyse the effect of calcium, the optimized sucrose solution was infused with calcium lactate (C₆H₁₀CaO₆) in varying concentration (1, 3, 5% w/w). At the end of osmosis, partially dehydrated banana slices with 49.78% (wb) moisture content were obtained which were subjected to hot air drying. All experiments were done in triplicate with each trial consisting of 3 banana slices.

Mass transfer parameters

The initial moisture content of banana slices was estimated by standard AOAC (1990) method by keeping the samples for 24 h at 70 °C in vacuum oven. The moisture ratio (MR), weight reduction (WR), water loss (WL) and solid gain (SG) were estimated using standard Eqs. 1–4 (Rani and Tripathy 2020).

$$MR=\frac{M_t-M_e}{M_o-M_e}$$ 1

$$WR,\%=\frac{w_i-w_f}{w_i}\times 100$$ 2

$$WL,\%=\frac{w_i{X}_i-w_f{X}_f}{w_i}\times 100$$ 3

$$SG,\%=\frac{w_f{X}_{\mathit{sf}}-w_i{X}_{\mathit{si}}}{w_i}\times 100$$ 4

where M_o, M_e and M_t are initial, equilibrium moisture content (% wb) and moisture content at any instant of time ‘t’ respectively; w_i and w_f are initial and instantaneous sample weight (g) at time ‘t’; X_i, and X_si are initial mass fractions of water and solids in the sample, respectively; X_f and X_sf are mass fractions of water and solids present in the sample at any instant of time ‘t’.

Experimental design and optimization

A three-factor full factorial design was employed to optimize the osmotic process of banana slices using Design-Expert 2019 software. The design setup consisted of 54 (3 × 3 × 6) experimental points. The three independent variables considered in the present work are osmotic process temperature (T) (°C), solution concentration (C) (°Brix) and immersion time (t) (h). The relationships between independent variables and responses are expressed by second order polynomial equation and the generalized form is given in Eq. 5.

$$Y={\beta }_0+\sum _{j=`1}^k{\beta }_j{X}_j+\sum _{j=1}^k{\beta }_{\mathit{jj}}{X}_j^2+{\sum }_i\sum _{<j=2}^k{\beta }_{\mathit{ij}}{X}_i{X}_j$$ 5

where Y is the response; X_i and X_j are variables (i and j range from 1 to k) and β₀ is the model intercept coefficient; β_j, β_jj, β_ij are interaction coefficients of linear, quadratic and the second order terms, respectively; k is the number of independent parameters (k = 3).

Response surface methodology (RSM) was used to optimize the OD process with the aim of achieving maximum values of WL and WR while targeting minimum values of MC, SG and total colour change. Some of these responses were competing, i.e., improving one response may have an opposite effect on another. Each response was transformed in to dimensionless desirability function (d_i) and the maximization of any response was denoted as desirability function (G), which was obtained by combining different individual desirability ‘d_i’ values of variables as shown in Eq. 6.

$$\text{G}={(d_1\times d_2\times \cdots \times d_n)}^{\frac{1}{n}}$$ 6

where d_i = d₁, d₂, d₃,… d_n denotes the desirability of the response and n is the number of responses. The developed model was validated by experimental trials at the optimum condition.

Air drying of osmosed banana slices

The OD of banana slices was followed by hot air drying in a convective tray drier (SD Instruments and Equipments, India) at different air temperatures (60, 65, 70 °C) and 1.5 m/s air velocity for subsequent removal of moisture till it reached 10% (wb). The weight of the samples was continuously measured at regular interval of 30 min throughout the drying period.

Rehydration of osmo-air dried banana slices

The rehydration process of the OAD samples was performed at room temperature (27 °C) and 50 °C maintained inside a constant temperature water bath (SD Instruments and Equipment, India). The samples were taken out from the bath at a regular interval of 15 min for the first two hours, after which, data was collected at an interval of 30 min. Samples were weighed after being blotted with a tissue paper to remove the excess water. Rehydration was stopped once a constant weight was attained.

Quality attributes

Colour

The colour of the samples was analysed in terms of L*, a* and b*, according to International Commission on Illumination measured using a colorimeter (Konica Minolta, Chroma meter CR-400, Japan). The total colour change (ΔE) was calculated as described by Rani and Tripathy (2019).

$$\mathrm{\Delta }E=\sqrt{{\left(L^{\ast }-L_o^{\ast }\right)}^2+{\left(a^{\ast }-a_o^{\ast }\right)}^2+{\left(b^{\ast }-b_o^{\ast }\right)}^2}$$ 7

where $L^{\ast }$, $a^{\ast }$ and $b^{\ast }$ stand for the lightness, redness or greenness, and blueness or yellowness of the treated sample, respectively, while $L_o^{\ast }$, $a_o^{\ast }$ and $b_o^{\ast }$ stand for the lightness, redness or greenness and blueness or yellowness of the fresh sample, respectively.

Texture

The hardness of fresh, osmosed and calcium osmosed banana slices (COBS) was analysed with the help of compression test in a texture profile analyser (SM-500N-168, Shimadzu, Japan). A compression plate probe of diameter 5 cm was used for penetration of 25% of sample thickness with a probe speed of 0.1 mm/s. The maximum force achieved in a force–deformation curve during the compression test provides the hardness of the sample (Rani and Tripathy 2019).

Bioactive properties

The total phenolic content (TPC) of the samples (mg gallic acid equivalents/100 g dry weight) was estimated by Folin–Ciocalteu method with some modifications (Singleton and Rossi 1965). The banana pulp was mixed with 80% ethanol (v/v) to prepare an ethanolic solution, which was further centrifuged and filtered to obtain an ethanolic extract supernatant. The obtained ethanolic extract was utilised for the estimation of phenolic compounds and antioxidants.

TPC was determined by treating the extract with Folin–Ciocalteu Reagent and 20% sodium carbonate solution (w/v). The samples were kept in dark for 90 min followed by absorbance measurement at 750 nm using UV–visible spectrophotometer (UV1700, Shimadzu, Japan). The TPC was expressed as milligrams of gallic acid equivalents (GAE) per g of dried sample using the standard gallic acid solution.

Antioxidant capacity of the samples was determined by mixing ethanolic extracts with DPPH (1,1-Diphenyl-2-picrylhydrazyl) and keeping them in the dark for 30 min followed by absorbance measurement at 517 nm using a spectrophotometer (Brand-Williams et al. 1995; Rani et al. 2018). A gallic acid standard curve was prepared to express the antioxidants of sample in terms of gallic acid equivalent (Kaushik et al. 2014; Chakraborty et al. 2015).

$$\text{GAEAC}=(\mathrm{\Delta }Ab{s}_{\mathit{sample}}/\mathrm{\Delta }Ab{s}_{\mathit{GA}})\times C_{\mathit{GA}}\times (V/W)\times 100$$ 8

where GAEAC: gallic acid (GA) equivalents antioxidant capacity per 100 g of dry weight, ΔAbs_sample: change of absorbance between blank and banana extract (Absorbance of blank-Absorbance of sample extract).

ΔAbs_GA:change of absorbance between blank and gallic acid solution (the volume of GA standard solution and fruit extract used for reaction with DPPH is same), C_GA: concentration of GA standard solution, V: volume of extract, W: weight of sample used for extraction, respectively.

Microstructural analysis

OAD banana slices were cut to get a specimen of 7 mm length and 2 mm thickness for each type of treatment- hot air dried, OAD, calcium osmo-air dried with varying calcium lactate content. The samples were coated with a fine layer of gold using a sputter gold coater (Q150R ES, Quorum Tech, United Kingdom). The samples were examined using scanning electron microscope (SEM) (ZEISS EVO-60, Carl ZEISS SMT, Germany) at a system with vacuum of 9.27 $\times$ 10⁻⁵ Torr and 500 × magnification.

FTIR spectroscopy

OAD banana slices were ground to powder form in a grinder and mixed with solid KBr to form a thin pellet using a hydraulic press (15 tons motorised pellet press, Kimaya engineers, India). The pellet was placed in the spectrometer (Nicolet 6700 FT-IR spectrometer, ThermoFisher Scientific, USA) where the absorbance values of the samples were measured over a wavelength spectrum ranging from 400 to 4000 cm⁻¹.

Statistical analysis

Statistical significance of the experimental results was examined by analysis of variance (ANOVA) at a confidence level of 95% using Design-Expert software (2019). All experiments were performed in triplicate and reported as mean ± standard deviation. Further, principal component analysis (PCA), a multivariate analysis method was performed on data set of mass transfer parameters and quality characteristics of banana slices during osmotic dehydration pre-treatment and hot air drying using the RStudio software. It is used to reduce the dimensions of data set without losing the useful information and helps to establish the relationship among different parameters. The number of principal components (dimensions) to be chosen from the data set was decided by obtaining the eigen values. As per the Kaiser-Guttman rule, the dimensions presenting the Eigen values > 1 should be retained (Njintang et al. 2007).

Results and discussion

Effect of osmotic dehydration on mass transfer parameters and process optimization

The effect of different process parameters on osmotic dehydration of banana slices is provided in Appendix (Appendix Table 2) where “p-value” < 0.05 indicated that the model terms are significant. It was observed that T, C and t had a significant effect on all the response parameters except SG, which was independent of the variation in temperature. C was found to have a significant effect on the water transport parameters (WL, SG and WR) and MC of samples. Higher concentration gradient set up at extreme levels of sucrose concentration produces greater osmotic pressure across the semi-permeable membrane which, in this case, was the complex cellular structure of the fruit that causes higher moisture removal from the sample and increased the solid gain. Responses followed a quadratic model with an R² value ranging from 0.75 to 0.95.

Table 2.

ANOVA of moisture content, water loss, solid gain, weight reduction, total colour change of banana slices during osmosis

Source	p-value
	Moisture content	Water loss (%)	Solid gain (%)	Weight reduction (%)	Total colour change
Model	< 0.0001*	< 0.0001*	< 0.0001*	< 0.0001*	< 0.0001*
T- Temperature	< 0.0001*	< 0.0001*	0.0563	< 0.0001*	< 0.0001*
C- Concentration	< 0.0001*	< 0.0001*	< 0.0001*	< 0.0001*	< 0.0001*
t- Immersion Time	< 0.0001*	< 0.0001*	< 0.0001*	0.0119*	< 0.0001*
TC	< 0.0001*	0.0301*	0.0530	0.6279	0.2184
Tt	< 0.0001*	0.0002*	0.1614	0.0575	0.2045
Ct	< 0.0001*	< 0.0001*	0.2456	0.0164*	0.0072*
T2	< 0.0001*	< 0.0001*	0.5535	0.0044*	< 0.0001*
C2	0.0406*	< 0.0001*	< 0.0001*	< 0.0001*	0.4475
t2	< 0.0001*	0.1136	0.3102	0.5912	0.1469
R2	0.95	0.93	0.79	0.75	0.84

*p-value (Prob > F) less than 0.05 indicated that the model terms are significant

Figure 1a represents the effect of C and T on MC after 6 h of osmosis. It was observed that increase of temperature and concentration led to significant reduction of MC. Simultaneous increase of temperature and concentration showed interaction effect as indicated by the lowest values of MC at the highest temperature (50 °C) and sucrose concentration (65°Brix). Furthermore, there is change in WL from the sample due to variation of T and concentration as shown in Fig. 1b. It was found that maximum WL of 47% occurred at extreme higher levels of all process variables, i.e., 65°Brix sucrose solution, 50 °C temperature and 6 h of immersion time. The increased value of WL at higher temperature can be ascribed to the structural alterations in cell such as swelling and plasticization of cell membrane rendering increased permeability to water and solids transfer at higher temperature (Campos et al. 2012). Moreover, at higher temperature, water molecules gained kinetic energy and the increased molecular motion contributing to higher water loss from sample. Furthermore, the viscosity of the osmotic solution also decreased at higher temperature resulting in reduced resistance to mass transfer. Chavan et al. (2010) also witnessed WL of banana slices in range of 48.7–49.3%, during osmotic dehydration in 60°Brix sugar solution and also observed a significant effect of contact time on weight reduction, sugar gain and water loss.

Fig. 1.

Effect of osmotic process variables on a moisture content (% wb) after 6 h of osmosis, b water loss (%) after 6 h of osmosis, c solid gain at 50 °C d weight reduction (%) after 6 h of osmosis, e total colour change after 6 h of osmosis

The effect of C and t on SG (%) at 50 °C temperature is shown in Fig. 1c. It was observed that increase of sucrose concentration and time increased the SG, and this effect was intensified with their simultaneous increase. During the first three hours, the rate of solid gain was high and it slows down gradually. This could be because of the reduced solid uptake potential with time as solid gain will lead to development of a high solid layer acting as a barrier to further solid uptake. Osmotic dehydration of banana with the application of high hydrostatic pressure, performed by Verma et al. (2013) also demonstrated reduced water loss and solid gain after 4 h of immersion. Shukla et al. (2018) also observed that increased concentration and temperature during banana osmotic dehydration develops high osmotic pressure gradient and reduced resistance to movement of solid and water, hence leading to high solid gain. The maximum solid gain by banana samples in sucrose solution was 15.39%, which was observed to be higher as compared to 6.69% obtained by Shukla et al. (2018). This is due to the longer duration of osmotic dehydration process in the present work. Significant impact of sugar concentration and duration on solid gain was also observed by Jain et al. (2011) during the process of papaya osmosis. Solid gained by the sample was found to be independent of temperature.

Figure 1d shows the variation in overall WR due to the change in sucrose concentration and temperature for immersion time of 6 h. Overall WR is considered to be the combined effect of WL from the sample and SG by the sample. It was found to be increased consistently with osmotic duration. It was noticed that for all temperatures of solution, with an increase of C, WR of samples increased up to 60°Brix, followed by 3% drop at 65°Brix. This can be attributed to the high values of SG that counters the reduction in weight due to WL. Figure 1e shows the effect of increase of temperature and concentration on total colour change (ΔE) of banana slices after 6 h of osmosis. It was found that ΔE was increased with increase of t that indicated more degradation of colour when samples were immersed for longer duration. Furthermore, lowest value of ΔE in the sample was observed at higher temperature (50 °C) and concentration (65°Brix), when immersed for 6 h. The lightness of the banana slices, (L*) decreased from an initial value of 61.86 ± 6.44 to 48.08 ± 2.57 due to OD after 6 h. Verma et al. (2013) also observed lower values of L* after osmotic dehydration of high pressure pre-treated banana slices because of enhanced sample opacity after osmosis. Interestingly, higher SC caused less reduction in L* value which may be due to the prevention in browning. No appreciable change was observed in a* and b* values during osmosis. The regression equations obtained for MC, WL, SG, WR, and ΔE are given as Eqs. 9–13, respectively, as provided below.

$$\begin{array}{cc}& MC=1.66+2.79\times T+0.89\times C+3.28\times T\times C-0.11\times t-0.09\times t\times C-0.03\times T^2-5\times {10}^{-3}\times C^2+0.5\times t^2\hfill \\ \hfill \end{array}$$ 9

$$\begin{array}{cc}& WL=40.9-3.96\times T+1.62\times C-5.95\times t+9.3\times {10}^{-3}\times T\times C+0.12\hfill \\ \hfill & \times T\times t+0.09\times t\times C+0.05\times T^2-0.02\times C^2-0.27\times t^2\hfill \\ \hfill \end{array}$$ 10

$$\begin{array}{cc}& SG=45.5-0.65\times T-1.57\times C-0.34\times t+6.1\times {10}^{-3}\times T\times C+0.03\hfill \\ \hfill & \times T\times t+0.02\times t\times C+3.9\times {10}^{-3}\times T^2+0.02\times C^2-0.13\times t^2\hfill \\ \hfill \end{array}$$ 11

$$\begin{array}{cc}& WR=-4.63-3.3\times T+3.2\times C-5.6\times t+3.2\times {10}^{-3}\times T\times C+0.09\hfill \\ \hfill & \times T\times t+0.08\times t\times C+0.04\times T^2-0.03\times C^2-0.14\times t^2\hfill \\ \hfill \end{array}$$ 12

$$\begin{array}{cc}& \mathrm{\Delta }E=58.9-2.5\times T-0.37\times C+8.8\times t+3.8\times {10}^{-3}\times T\times C-0.04\hfill \\ \hfill & \times T\times t-0.05\times t\times C+0.03\times T^2+2.2\times {10}^{-3}\times C^2-0.3\times t^2\hfill \\ \hfill \end{array}$$ 13

where T, C and t denote temperature (°C), sucrose concentration (°Brix) and immersion time (h), respectively.

The process parameters during OD were optimised by RSM to find the optimal operating conditions to achieve maximum water removal (higher values of WL, WR and lower value of MC) and better quality (lower values of TSS and SG). The optimum processing condition for OD of banana slices were found to be 61.26°Brix sucrose concentration at 50 °C temperature and 6 h immersion duration with a desirability value of 0.85 (Appendix Table 3). Experimentation was carried out at optimized conditions and it was observed that values of dependent parameters obtained from experiments were close to the values predicted by the RSM at optimsed conditions.

Table 3.

Optimization of process variables for osmostic dehydration of banana slices

Input	Constraints				Optimal solution
	Goal	Lower limit	Upper limit	Importance
Temperature,°C	In range	30	50	3	50.00
Sucrose concentration, °Brix	In range	35	65	3	61.26
Immersion time, h	In range	1	6	3	6.00
Water loss, %	Maximize	3	47.09	5	46.56
Solid gain, %	Minimize	0	15.98	1	11.72
Weight reduction, %	Maximize	0	38.18	5	34.82
Moisture content, % wet basis	Minimize	42	68	5	49.78
Total colour change	Minimize	3.25	31.09	1	13.8

Principal component analysis of osmotic dehydration process

PCA was carried out on the data set of MC, MR, WR, SG, WL, TSS, L*, a*, b* and ∆E obtained at different immersion time of banana slices in osmotic solution at varied solution temperature. The first two principal components (PCs) i.e., dimensions (dims.) generated from the dataset were found to have eigen values (dim.1 = 4.49, dim.2 = 3.43) greater than 1 and hence were retained for further analysis and plotting. The dim.1 and dim.2 explained the variance of 44.9% and 34.4%, respectively, together explaining a total variance of 79.3%. Figure 2a plots the contribution of factors to all 10 dimensions and it can be observed that all the factors except SG showed maximum contribution to PC1 and PC2. Moreover, WR, WL, MC, MR, L* and TSS were found to be mainly related with dim. 1; on the other hand, a*, b* were mainly influenced by dim.2, whereas ∆E was influenced by both the dims. The magnitude of correlation among different pairs of factors and the PCA biplot illustrating different factors along with distribution of loading points are shown in Fig. 2b–e. The score points are grouped according to immersion time, osmotic solution temperature (OT) and osmotic solution concentration in individual Fig. 2c–e, respectively.

Fig. 2.

Principal component analysis of osmotic dehydration process of banana a contribution of factors to dimensions b correlation matrix among different factors c PCA-biplot grouping data set based on time d PCA-biplot grouping data set based on solution temperature e PCA-biplot grouping data set based on solution concentration (SC)

The mass transport properties SG, WR, WL and TSS were positively correlated with each of them and were positioned close to each other on biplot indicating that these parameters showed similar type of behaviour. However, these responses were negatively correlated with MC, MR and positioned far away from each other on biplot. It is because, as the osmotic dehydration proceeds, the value of MC and MR reduces; on the other hand, there was increase in solid gain, water loss and TSS of the sample. The L* and b* presented positive correlation with each other, but negatively correlated with a* and ∆E. As the MR is reduced with osmosis, there was increase in ∆E and hence both were negatively correlated. Distribution of data points along with factors permits the visualization of relationship among them. The data points presenting 0 h of osmosis were near to MR, MC, L* indicating higher values of these parameters in these experimental runs. In a similar manner, data sets of osmosis for 6 h were found to lie near to SG, WR, WL, TSS and ∆E, hence indicating higher moisture loss and higher ∆E. In the same region, the dataset of experimental runs carried out at maximum concentration and higher temperature also lie down, indicating their positive effect on moisture loss.

Hot air-drying kinetics

The drying time to achieve 10% (wb) MC at 70, 65 and 60 °C was 4.5, 5.5 and 6 h, respectively, indicating increased moisture transfer at higher drying temperature. The equilibrium moisture content of the fresh banana slices after drying was 8 ± 0.45% (wb), while the equilibrium moisture content of the osmo-air dried banana slices was 8.35 ± 0.23% (wb). The drying kinetics of fresh and osmosed banana slices at 70 °C are presented in Fig. 3a. It can be observed that the MC of fresh samples and osmosed samples decreased continously. During the initial hours, moisture removal from osmosed samples was faster than fresh samples, however after 2.5 h of drying, moisture removal from osmosed samples became slower than fresh samples. This decline of moisture removal after 2 h in osmosed samples is due to high resistance to mass transfer offered by sucrose molecules gained during osmosis. Thus, osmotic pre-treatment was found to have a negative effect on moisture removal rate. This might be ascribed to the interaction of sugar with water molecules that can hinder the internal channels for water diffusion of material, reducing the transfer of water (Mandala et al. 2005). It must also be noticed that the addition of calcium lactate had no prominent influence on the drying kinetics of osmosed banana slices.

Fig. 3.

Drying and rehydration kinetics a Drying kinetics of fresh and osmosed banana slices at 70 °C, b Rehydration kinetics at 50 °C and c Rehydration kinetics at 27 °C (room temperature)

Rehydration kinetics

Many times, dried fruits and vegetables are rehydrated prior to use in different food products to reconstitute their raw properties. Rehydrated banana can find its application as an ingredient in different food products like ice-cream, cakes, flakes, custard, granules etc. Further, rehydration is one of the quality indicators of dried food product and also provides the information about internal damage in the product occurred during drying (Tripathy and Kumar 2009). Therefore, rehydration of dried product is important characteristic to be studied to represent the ability of the dried banana samples to reconstitute and gain original properties.

The rehydration behaviour of OAD banana slices at 50 °C and 27 °C (room temperature) are shown in Fig. 3b and c. The rehydration period of air-dried samples at 50 °C was 33% lesser than that at room temperature, which was found to be similar for all samples regardless of the drying temperature or pre-treatment. Furthermore, the samples were visually similar and the rehydration capacity was also not considerably affected by the change in rehydration temperature (RT). The banana slices showed high water absorption initially because the capillary imbibition is important at early stages, which leads to almost instantaneous uptake of water. As the soaking time increased, the absorption rate reduced which was reported in a previous rehydration study by Ajala et al. (2015). It was also observed that equilibrium moisture content was highest for un-osmosed samples and lowest for the osmosed samples containing higher percentage of calcium lactate. Furthermore, osmosed samples contain about 11–13% solutes which leached into water during the rehydration process.

Quality of osmosed and osmo-air dried banana slices

Colour

The osmosed banana slices reported a reduced value of lightness, L* with respect to fresh banana slices resulting in an overall colour change of 12.52 ± 3.99. The addition of calcium lactate had no effect on the total colour change of the osmosed banana slices as shown in Table 1. However, when subjected to drying the osmosed banana slices exhibited excellent colour retention properties (ΔE = 9.08 ± 4.89) while fresh bananas experienced extensive browning (ΔE = 37.97 ± 6.91). This stability of the brightness (L*) of samples may be explained by the presence of sucrose which reduces the water activity of the samples and inactivate enzymes to a certain extent, thereby preventing non-enzymatic and enzymatic browning, respectively (Krokida et al. 2000). The increase of drying temperature from 60 °C to 70 °C, led to reduction in the total color change during the drying of omosed banana slices, indicating that better colour retention was achieved at high drying temperature. Vega-Gálvez et al. (2012) also witnessed a lower colour change at higher drying temperature during the drying of apple slices. The change in the L* of the banana slices was also least at high temperature drying conditions. This is due to the fact that at higher temperature, the samples are dried faster and hence they were exposed to higher temperature for lesser time as compared to 65 °C and 60 °C.

Table 1.

Effect of drying temperature and osmotic pretreatment on the quality of dried banana slices

Sample	Drying Temperature (°C)	L*		ΔE	TPC (mg GAE/g d.w)		Antioxidant (mg GAE/100 g d.w)
		Before drying	After drying		Before drying	After drying	Before drying	After drying
Drying without osmotic pretreatment
Control	60	66.09 ± 1.2	31.6 ± 1.3	41.38 ± 2.4	1.67 ± 0.22	1.03 ± 0.30	25.12 ± 3.57	3.77 ± 0.59
	65		41.61 ± 3.1	30.02 ± 1.1		1.03 ± 0.42		2.37 ± 0.74
	70		30.96 ± 1.2	42.52 ± 2.1		1.12 ± 0.25		3.17 ± 1.03
Drying with osmotic pretreatment
Sucrose + 0% Calcium lactate	60	54.32 ± 2.1	37.15 ± 2.1	23.09 ± 3.4	1.46 ± 0.17	1.00 ± 0.29	11.52 ± 6.74	13.27 ± 2.02
	65		47.26 ± 2.2	9.73 ± 0.9		1.06 ± 0.37		12.42 ± 1.31
	70		42.99 ± 3.2	7.80 ± 2.3		1.18 ± 0.33		10.79 ± 2.89
Sucrose + 1% Calcium lactate	60	53.38 ± 5.1	52.41 ± 1.9	12.64 ± 1.6	1.58 ± 0.61	0.76 ± 0.14	17.22 ± 2.21	11.26 ± 1.34
	65		50.44 ± 1.6	7.04 ± 2.6		1.05 ± 0.45		14.00 ± 0.79
	70		51.78 ± 3.7	7.05 ± 1.7		1.21 ± 0.41		13.69 ± 4.28
Sucrose+3% Calcium lactate	60	56.77 ± 3.4	53.32 ± 2.1	6.43 ± 1.4	1.44 ± 0.22	0.99 ± 0.20	17.95 ± 1.12	15.49 ± 0.92
	65		54.44 ± 1.4	7.50 ± 1.3		1.06 ± 0.33		12.82 ± 2.08
	70		54.47 ± 3.1	4.42 ± 1.8		1.22 ± 0.35		11.77 ± 0.59
Sucrose + 5% Calcium lactate	60	53.58 ± 2.1	47.46 ± 3.4	9.77 ± 1.2	1.42 ± 0.25	1.07 ± 0.26	9.23 ± 5.47	13.76 ± 2.23
	65		49.8 ± 2.1	5.83 ± 2.2		1.00 ± 0.15		9.63 ± 0.60
	70		51.17 ± 1.1	7.65 ± 1.3		1.23 ± 0.36		8.32 ± 4.23

Texture

The hardness of fresh sample was 11.7 ± 0.7 N, which reduced to 2.08 ± 0.21, 1.65 ± 0.32 N, and 0.75 ± 0.08 N during OD process at 30, 40 and 60 °C, respectively. Such a decline in the hardness value during the OD is expected due to the loss of turgid pressure and cell integrity (Deng and Zhao 2008) and this reduction was more pronounced at higher temperature because of the increased rate of water loss. The reduction in hardness of the samples was prominent in the first hour, after which there was no significant change (p > 0.05). The samples osmosed in sucrose solution added with 1%, 3% and 5% calcium lactate exhibited 3.77 ± 0.5 N, 3.88 ± 0.1 N, 9.63 ± 3 N hardness indicating the preservation of firmness of the samples during osmosis. This is justified as calcium binds with the pectin present in the fruit, and provides a more rigid structure to the middle lamella of the cell wall (Lovera et al. 2014). It was also found that the hardness value of the osmosed slices was directly proportional to the calcium lactate concentration in the solution.

Bioactive contents

The bioactive content of banana slices reduced during osmosis because of the leaching of water-soluble compounds like phenolics into the osmotic solution. The TPC of fresh banana slices was reduced from 1.67 ± 0.22 mg GAE/g d.w to 1.47 ± 0.08 mg GAE/g d.w while the AOX reduced from 26.39 ± 3.57 mg GAE/ 100 g d.w to 13.98 ± 4.28 mg GAE/100 g d.w after osmosis process. Study by Almeida et al. (2015) reported higher phenol retention in banana slices during osmotic dehydration at lower temperature, i.e., at 30 °C as compared to 40 °C and 50 °C. Calcium lactate did not impart any appreciable effect on the TPC and AOX. Drying of fresh and osmosed samples led to significant reduction in bioactive content of banana due to their exposure to oxygen and heat for a longer period as shown in Table 1. Further, loss of antioxidants during air drying was higher in un-osmosed banana slices (88.23%) when compared to osmosed banana slices (23.02%). Similar observations were also reported in dried goji berry (Dermesonlouoglou et al. 2018), apricots (Sakooei-Vayghan et al. 2020) and papaya (Islam et al. 2019), where osmotic pretreatment resulted in higher retention of polyphenols and antioxidant activity during drying.

Microstructure

The fractured surface micrographs indicated by SEM of fresh and pre-treated banana slices are shown in Fig. 4a–e. The microstructural surface of banana slices dried without osmosis appeared rough, little porous while the surface of calcium osmosed dried samples exhibited a compact structure. The higher roughness observed in untreated dried banana slices indicates higher structural changes occurred during drying. This facilitated the higher migration of moisture from the sample to surrounding during drying of fresh samples as compared to calcium osmosed samples. During the banana slices osmosis using calcium-sucrose solution, calcium binds with the pectin and protects cell integrity leading to slow drying process as indicated by compact and continuous microstructure shown in Fig. 4c–e. It can be noted that osmosed banana slices of 5% calcium lactate solution presented more binding and compactness in the structure. Study carried out by Quiles et al. (2004) on osmotic dehydration of apple slices also observed that cells in the case of CaCl₂ treated osmotic dehydrated apples were not that collapsed as in osmosed samples without CaCl₂ treatment. They concluded that stabilized cellular integrity of CaCl₂ osmosed apples slices was because of the microstructure preserving, cellular wall strengthening effect of calcium.

Fig. 4.

Fractured surface SEM images of a fresh, and osmosed banana slices with b 0%, c 1%, d 3% and e 5% calcium lactate at 500 × , f Fourier Transform Infrared (FTIR) spectroscopy image of fresh and osmosed banana samples

FTIR spectroscopy

The FTIR spectra of fresh and osmosed banana slices with varying concentration of calcium lactate is shown in Fig. 4f. The spectrum showed peaks around 3.4 × 10³ cm⁻¹ indicate the O–H vibrations and presence of primary amines (N–H stretching). The peaks around 2.9 × 10³ cm⁻¹ and 1.64 × 10³ cm⁻¹, are due to the alkanes (C–H stretching) and aromatic groups of proteins (C=C stretching), respectively (Wu et al. 2020). The peaks near 1.27 × 10³ cm⁻¹ represents the C-O stretching, while peak around 1.06 × 10³ cm⁻¹ is indicative of OCH₃ vibrations in polysaccharides (Dehsheikh and Dinani 2020). All the samples exhibited similar trend but the peaks observed in the pretreated samples had higher absorbance values. The FTIR peaks at 6 × 10² cm⁻¹ to 5 × 10² cm⁻¹ represents the Aliphatic iodo compounds (C–I). A weak peak was around 1.7 × 10³ cm⁻¹ which eludes the presence of aromatic compounds (C–H bending) (Coates 2006). The absorbance values for this peak were higher for calcium osmosed compounds which could be attributed to higher phenol retention in calcium osmosed samples during drying. The sharp peaks observed at 1.64 × 10³ cm⁻¹ and 2.9 × 10³ cm⁻¹ are characteristic features of calcium lactate infrared spectrum (NIST 1997). This explains the significantly higher absorbance values for the calcium in the banana slices which were osmosed with sugar and calcium lactate. It is interesting to note that the absorbance values at these wavelengths for 5% calcium lactate osmosed samples were not the highest despite having maximum calcium content in the osmotic solution. This may be due to inhibition of calcium absorption due to surface accumulation.

Principal component analysis of hot air-drying process

The PCA analysis on final moisture (M_f), drying time (DT), total phenolic content (TP), antioxidants (AOX), color profile of fresh, osmotic dehydrated and osmo air dried samples at different temperature was carried out. Figure 5a shows the contribution of the factors for different dimensions (principal components) and first two PCs explained a total variance of 81%. The M_f, TPC, AOX, L*, a*, b*, ∆E were primarily affected by dimension 1, however, dimension 2 also had some effect on TPC, L*, a*, b*. Figure 5b and c shows the correlation among different parameters and PCA biplot with two dimensions showing the relationship among datasets and factors, respectively. Drying of banana slices reduces the moisture content with increase in drying time, hence M_f and DT were positioned far way and showed a slight negative correlation. With a decrease in moisture content, TPC and AOX also reduces, therefore showing a good positive correlation with final moisture content. It can be further observed from the biplot that, osmotic dehydrated samples were distributed around TPC and M_f, indicating that osmotic dehydrated samples had higher moisture content along with total phenolic content. For hot air-dried samples at different temperature, TPC and AOX were reduced along with increase of a* parameter and total color change.

Fig. 5.

Principal component analysis of hot air drying of osmosed banana a contribution of factors to dimensions b correlation matrix among different factors c PCA-biplot showing response factors and data point distribution

This study explored the prospect of developing calcium fortified banana dried product, which can be utilized as dried banana chips/snacks, dried banana powder. Dried banana products can be further rehydrated to reconstitute the properties of raw banana for utilization in different food products like ice-cream, cakes, flakes, granules, breakfast cereals etc. Osmotic pretreatment also exhibited improvement in the quality of the dried banana slices in terms of higher bioactive components, colour stability. Since both drying and osmotic dehydration are simple processes, osmo-air drying can be easily adopted by small scale industries without major capital investment. In addition to it, information obtained about interrelationships among different mass transfer parameters during osmotic dehydration and associations among drying characteristics and quality attributes during hot air drying will be helpful for the researchers and engineers in optimizing these processes.

Conclusion

In the present research work, OD of banana slices was followed by hot air drying and rehydration process. During the osmosis, solution temperature, concentration and immersion time were found to have a significant effect on the mass flux parameters of banana slices. The optimum osmotic conditions as predicted by RSM was found to be sucrose concentration as 61.26°Brix, temperature as 50 °C and immersion time for 6 h, which resulted in moisture reduction from initial moisture content of 75% to 49.78% (wb). PCA of OD process summarized the relationships among the response variables and demonstrated a positive correlation among SG, WR, WL and TSS with each other, however negative correlation with MC of banana slices was observed. The addition of calcium lactate in the osmotic solution preserved the firmness and bioactive properties of banana slices. Osmosed banana slices subjected to hot air drying at 70 °C reduced the drying time by 22% and 33% with respect to drying time at 65 °C and 60 °C, respectively. PCA-biplot of OAD process presenting the associations among dataset and response factors, indicated a positive correlation of bioactive component with MC and therefore, reduced bioactive component retention was observed after hot air drying. Osmo-air drying along with addition of calcium lactate was found to be useful in preserving the ripe banana, resulting in superior quality with no adverse effect on the overall process. Moreover, the dried banana slices can also serve as a source of calcium for individuals with low calcium diet. However, further studies must be conducted to understand the in vivo effect of the developed product and its potential in acting as an effective calcium delivery system. The effect of different osmotic conditions on TPC and antioxidant activity can be further studied by researchers and academicians for quality attributes accounted optimization of osmotic process.

Abbreviations

OD

Osmotic dehydration

RSM

Response surface methodology

OAD

Osmo-air dried

wb

Wet basis

MC

Moisture content (%)

MR

Moisture ratio

WR

Weight reduction (%)

WL

Water loss (%)

SG

Solid gain (%)

OBS

Osmosed banana samples

COBS

Calcium osmosed banana samples

TPC

Total phenol content (mg GAE/ g d.w)

AOX

Antioxidant capacity (mg GAE/ 100 g d.w)

GAE

Gallic acid equivalent

DPPH

1,1-Diphenyl-2-picrylhydrazyl

SEM

Scanning electron microscope

FTIR

Fourier Transfer Infra-red

ANOVA

Analysis of Variance

RT

Rehydration temperature (°C)

PCA

Principal component analysis

List of symbols

d

Diameter (cm)

L

Thickness (cm)

T

Osmotic temperature (°C)

C

Solution concentration (°Brix)

t

Immersion time (h)

L*

Brightness

a*

Redness or greenness

b*

Yellowness or blueness

ΔE

Total colour change

Appendix

See Tables 2 and 3.

Authors' contributions

RR conceived the idea, carried out the experiments, analysed data, and wrote the manuscript; PR supervised the work, analysed data, wrote the manuscript; PPT conceived the idea, provided the resources, supervised the work, and edited the manuscript.

Funding

The authors declare that this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Availability of data and material

The datasets generated and/or analysed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Declaration

Conflict of interest

The authors declare that they have no conflict of interest.