The orange rank 5^th in world ranking for production at global region in tropical fruit category (Patil et al., 2018). The global orange (Citrus reticulata Blanco) product processing industry generates a substantial amount of waste, primarily in the form of peels, seeds and membranes, which creates significant environmental and economic challenges if not effectively managed. Among citrus wastes, orange peels constitute a major proportion and are often underutilized despite being rich in bioactive compounds, dietary fiber, flavonoids, essential oils and pectin (Ajila et al., 2011; Tripodo et al., 2004).
       
Orange peel is a natural source of dietary fiber, particularly pectin, which improves gut health by promoting regular bowel movements and supporting healthy gut microbiota (Rafiq et al., 2018; Sharma et al., 2017; Sankalpa et al., 2017). Fiber also contributes to the textural and physicochemical properties of food products, aiding moisture retention and structure in formulations such as candy. In addition, orange peels contain substantial amounts of vitamin C (ascorbic acid), an essential antioxidant that supports immune function and contributes to the nutritional value of food products (Sonia et al., 2016). Transforming such by-products into value-added functional foods not only supports waste minimization but also aligns with the principles of environment, sustainability and the circular bioeconomy (Mirabella et al., 2014; Thakre et al., 2026).
       
Osmotic dehydration is an effective mild preservation technique widely used in developing fruit-based confectionery products, such as orange candy (Rastogi and Raghavarao, 2004). The process involves immersing fruit tissues in a hypertonic sugar solution, resulting in the simultaneous removal of water from the fruit (water loss) and the uptake of solutes, such as sugar (solid gain), due to osmotic pressure gradients (Shi and Le Maguer, 2002). This method enables partial dehydration while largely retaining the cellular structure, natural color and flavor of the fruit (Torreggiani and Bertolo, 2001). In recent years, the development of fruit-based candies has gained attention owing to the growing consumer interest in functional snacks with health-promoting properties (Kaur and Kumar, 2019).
       
Response surface methodology (RSM) provides a powerful statistical and mathematical technique for modeling and optimizing processes with reduced experimental runs. This enables the identification of optimal conditions through regression modeling and interaction analysis among variables (Myers et al., 2016). Several studies have employed RSM to optimize food processes, such as drying, extraction and confectionery product development, demonstrating its efficiency and predictive capabilities (Bas and Boyaci, 2007; Singh et al., 2020).
       
Several studies have reported the application of osmotic dehydration for fruits and fruit peels, mainly focusing on drying kinetics, product quality attributes, or bioactive compound retention. However, limited information is available on the systematic optimization of osmotic dehydration parameters specifically for citrus peel waste using response surface methodology (Prajapati et al., 2025). Most previous studies either investigate osmotic dehydration as a stand-alone process or emphasize final product characteristics, rather than focusing on mass transfer optimization as a valorization-oriented pre-treatment step. In this context, the present study differs from earlier work by applying box-behnken design to optimize water loss and sugar gain, which are critical indicators of osmotic mass transfer, with the aim of developing an efficient and controlled pre-treatment strategy for citrus peel valorization. This study aimed to model and optimize the formulation of orange peel candy using RSM, focusing on key parameters such as sugar syrup concentration, immersion time and solution temperature. The objective of the current study was to maximize product acceptability while enhancing the utilization of orange peel waste. Such valorization not only contributes to reducing agro-industrial waste but also offers a sustainable alternative to synthetic candies through the development of natural, fiber-rich and functional products.

Orange peels (Citrus reticulata Blanco) were collected from local juice vendors in Lonikalbhor, Pune, India. The orange peels were thoroughly washed with water and stored at 4°C. The research work was carried out at the School of Food Technology, MIT Art Design and Technology University in 2025.
 
Sample preparation and experimental procedure
 
The orange peel slices were cleaned with water to remove dust, dirt and other unwanted materials stuck to their surfaces. The slices were cut into 5±1 mm thick strips using a sharp stainless-steel knife. Osmotic dehydration was performed in sugar solutions with different concentrations of 50, 60 and 70 °Brix. The sample-to-immersion medium ratio was constant at 1:5 (w/w). The orange peel slices were first weighed and then submerged in sugar solution maintained at 40, 50 and 60°C. The samples were removed from the osmotic medium at fixed durations (30, 60 and 90 min.). The orange peel slices were drained after removal from the sugar solution. The excess solution on the surface of the slices was wiped away with tissue paper to measure the weight. The initial moisture content of raw and treated orange peel slices was determined using the oven method (AOAC, 2012). After osmotic dehydration, the optimized sample was dried in a cabinet dryer at temperatures of 50, 60 and 70°C, for 6-8 hours and at an air velocity of 1 m/s. The osmo-dried orange peel candy (Fig 1) exhibited a final moisture content of 12±0.4% (wet basis) after cabinet drying.


 
Mass transport parameters for osmotic dehydration
 
The mass transfer parameters, such as water loss (WL), mass reduction (MR) and sugar gain (SG), which are considered quality aspects of orange slices, were calculated using the equation provided by Kedarnath et al., (2014).

 

Where,
WL= Water loss (g per 100 g mass of orange slices),
SG= Solid gain (g per 100 g mass of orange slices).
MR= Mass reduction (g per 100 g mass of orange slices),
W_θ = Mass of orange slices after time θ, g.
W_i = Initial mass of orange slices (g). 
X_θ = Water content as a fraction of the mass of orange slices at time θ.
X_i = Water content as a fraction of the initial mass of the orange slices (fraction).
 
Experimental design and data analysis
 
Response surface methodology (RSM) using a box-behnken design was employed to evaluate the effects of osmotic dehydration on water loss and sugar gain in orange peel slices. The independent variables considered were sugar concentration (C), solution temperature (T) and immersion time (θ), as shown in Table 1. A three-factor, three-level box-behnken design comprising 17 experimental runs, including five central points, was generated using the design-expert software (version 11).



It is assumed that the mathematical function f_k (k = 1, 2, 3,.....n), exists for each response variable, Y_k in terms of the processing factors, ri (i =1, 2, 3,.....m), such as:
 

                                                     Y_k = f_k (r₁, r₂, r₃,...r_i)                                   ...(4)                                     

 
The exact mathematical representation of function (f) is either unknown or extremely complex. However, a second-order polynomial equation of the following form was assumed to relate the response, Y_k and the factors, r_i.

 
Where,
β_ko, β_ki, β_kii and β_kij= Constant coefficients and x_i.
       
The coded independent variables, are linearly related to T, C and θ. In practice, the levels of independent variables change from one application to another.
 
Sensory analysis
 
Sensory analysis of osmo-convective dried orange peel candy was carried out using 9^th point hedonic scale. Twenty-five semi-trained panel members (Male-13 and Female- 12 women) were selected for sensory evaluation from the School of Food Technology, Pune. The panelists were informed of the objective of the study before the sensory evaluation and written consent was obtained from each individual.

Effect of process variable on water loss of orange peel slices
 
The variation in the water loss of orange peel slices was studied by changing the syrup temperature, syrup concentration and immersion time in experimental studies, as shown in Table 2. A wide variation in water loss was observed, ranging from 16.02 to 40.13%. A second-order polynomial equation (Eq. 5) was developed to describe the experimental data. The quadratic model was statistically analyzed to determine the significance of the linear, quadratic and interaction effects on water loss and the results are summarized in Table 3. The coefficient of determination (R² = 0.997), calculated using the least squares method for water loss, indicates excellent agreement between the model predictions and experimental values. Furthermore, the high model F-value (480.31) and corresponding probability value (P<0.0001) indicate that the model is highly statistically significant. The lack-of-fit test was non-significant, demonstrating that the proposed model was adequate and reliable for predicting water loss during the osmotic dehydration of orange peel slices.




       
The regression equation representing the influence of the process variables on water loss, expressed in terms of their actual values, is presented in Eq. (6). Water loss data were analyzed using a stepwise regression approach, in which model terms with F-values less than one were excluded, in accordance with the criteria recommended by Snedecor and Cochran (1967).
 

  WL = 30.61 + 5.36A + 6.78B + 2.84C + 0.0250AB - 0.3900AC + 1.39BC - 1.18A2 - 1.38B² - 4.20C²      ...(6)

                                                               
The influence of individual process variables and their interactions on water loss (WL) was interpreted using the coefficients in Eq. (6). Among the variables studied, syrup temperature and immersion time had the most pronounced effect on WL, followed by syrup concentration. The inclusion of the quadratic terms in Eq. (6) confirms the curvilinear behavior of the response surface for the WL. Moreover, the negative coefficients associated with the quadratic terms indicate that beyond certain levels, further increases in the process variables led to a marginal or insignificant reduction in WL (%).
       
The response surface and contour plots for water loss were generated from the fitted model as a function of two variables, while the third variable was maintained at its central level Fig 1 (A-C). Water loss increased rapidly during the initial stages of osmosis and the rate of water loss gradually decreased over time.
       
The water loss increased with an increase in syrup concentration (Fig 2A and B) and immersion time (Fig 2B and C) over the entire osmotic dehydration process. Fig 2 (A and C) indicates that an increase in syrup temperature increases water loss. This may be due to the higher temperatures that seem to promote faster water loss through swelling and plasticizing of cell membranes (Patil et al., 2014). Similar findings have been reported for the osmosis of mangoes by Duduyemi et al. (2013).


 
Effect of process variable on sugar gain of orange peel slices
 
The sugar gain observed during the osmotic dehydration of orange slices ranged from 2.12 to 12.65% under different process conditions. Table 3 shows that the process variables exhibited significant effects on sugar gain at the linear, quadratic and interaction levels. The coefficient of determination (R²) for sugar gain was 0.992, indicating an excellent fit of the model to experimental data. The model F-value of 58.53 for sugar gain implies that the model was significant (P<0.0001) (Table 4). The F-value of the lack of fit was non-significant, indicating that the developed model was adequate for predicting sugar gain during the osmotic dehydration of orange slices.


       
A second-order polynomial model (Eq. 5) is fitted to the experimental data. The resulting regression equation, which describes the influence of the process variables on sugar gain using their actual values, is presented in Eq. (7). Sugar gain data were analyzed using a stepwise regression technique, in which terms with F-values less than one were eliminated, as recommended by Snedecor and Cochran (1967). 
 

      SG = 8.53 + 1.51A + 2.58B + 1.81C + 0.7950AB - 0.3950AC + 0.02300BC - 0.0850A² - 0.6350B² - 1.18C²      ...(7)

                                                                                                                                                                               
The coefficients of Eq. (7) were used to interpret the effects of individual process variables and their interactions on sugar gain (SG). Syrup temperature and immersion time had the most pronounced influence on SG, followed by syrup concentration. An excessive increase in the levels of these variables led to a significant increase in SG (%), as reflected by the negative coefficients associated with the quadratic and interaction terms.

The response surface and contour plots for water loss were generated for the fitted model as a function of two variables while keeping the third variable at its central value, as presented in Fig 3 (A-C). It was observed that sugar gain increased rapidly in the early stages of osmosis and the rate of sugar gain gradually slowed down with time.



Fig 3 (A-C) shows that sugar gain increased with an increase in syrup temperature, sugar concentration and immersion time. An increase in the concentration of sugar syrup also led to an increase in sugar gain (Fig 3A and B), which might be due to an increase in the osmotic pressure gradient and consequent loss of functionality of the cell plasmatic membrane that allows solutes to enter. Similar results were obtained for orange slices (Harati et al., 2011) and mango (Duduyemi et al., 2013).
 
Optimization of osmotic dehydration of orange peel candy
 
The optimization constraints are listed in Table 5. The process variables, temperature (T), concentration (C) and immersion time (θ), were set at their minimum feasible levels within the experimental range. The primary optimization objective was to achieve maximum water loss while targeting a sugar gain of 8.40%. The target sugar gain value of 8.40% was selected based on preliminary sensory trials, which indicated that this level of sugar uptake provided acceptable sweetness without imparting excessive sweetness or masking the characteristic citrus flavor (Ceballos et al., 2020). However, the experimentally obtained sugar gain under the optimized osmotic dehydration conditions was 7.38%, as predicted and validated through the box-behnken model. The slight deviation between the target and obtained values can be attributed to the fact that sensory-based targets represent subjective quality preferences, whereas the predicted value is derived from statistically modeled mass transfer behavior. Additionally, natural variability in raw material characteristics and diffusion-controlled mass transfer during osmotic dehydration may contribute to minor differences between targeted and observed responses (Kumar et al., 2021). Despite this deviation, the obtained sugar gain is close to the target value and remains within an acceptable range for sensory acceptability and process optimization.


 
Sensory evaluation of osmo-convective dried peel candy
 
Sensory analysis of the developed candy samples (CD₅₀, CD₆₀ and CD₇₀) was performed for five key quality parameters: flavor, color, taste, texture and overall acceptability. Table 6 presents the mean scores and their standard deviations. Among the three formulations, the CD₆₀ sample exhibited the highest scores across all sensory parameters, indicating better consumer preference. Specifically, CD₆₀ obtained the highest scores for flavor (8.1±0.4), color (8.2±0.11), taste (8.2±0.3), texture (8.1±0.4) and overall acceptability (8.2±0.3). This indicates that increasing the concentration to 60% positively contributes to the organoleptic characteristics of the product.


       
Sample CD₇₀ also showed good acceptability, with scores ranging from 7.8 to 8.0, but these values were slightly lower than those of CD₆₀. The CD₅₀ sample received comparatively lower scores than CD₆₀ and CD₇₀ for most attributes, indicating that a lower concentration may not sufficiently enhance the sensory quality of the product.

The authors would like to thank the School of Food Technology, MIT ADT University, for providing their laboratory facilities for the successful completion of this work.
 
Declarations
 
Funding
 
The authors have not disclosed any funding information.

The authors declare that there are no conflicts of interest.