Mass transfer parameters and quality characteristics of aonla slices under refractance window drying

Abstract This study investigates the effects of refractance window (RW) drying process parameters on mass transfer and quality characteristics of dried aonla slices. RW drying of aonla slices was carried out at three levels of water temperature (75, 82 and 90 °C) and slice thickness (2, 4 and 6 mm). In terms of quality characterization, the total phenolic content, ascorbic acid content, and browning index were determined. Higher retention of ascorbic acid (64.49 ± 0.34%) and phenolic content (37.84 ± 0.08 mg GAE/g dry matter) was found at 90 and 82 °C water temperature, respectively with inconsequential variation in browning index. Mass transfer parameters such as Biot number, effective moisture diffusivity and mass transfer coefficient were estimated using Dincer and Dost model. Effective moisture diffusivity during drying varied from 4.27 × 10−10 to 1.09 × 10−09 m2s−1 and Biot number was observed to be in the range of 0.268 to 8.666 for different drying conditions. XRD pattern suggested that RW dried aonla slices had semi-crystalline structure. The changes in crystalline structure to amorphous was pronounced more at high water temperature. It was also revealed by the disintegration of cell wall/membrane and connecting surfaces, which resulted into smooth and flaky microstructures with sharp edges. The gamut of crystallite size was obtained as 49.12 ± 0.26 to 99.12 ± 1.19 nm. The presence of bound water after completion of drying was represented by FTIR peaks centered around 3577 cm−1. The infrared spectroscopy of RW dried aonla postulated that peak intensity of absorption bands negligibly changed with varying processing conditions, but minor peak shift was observed. This study elucidates the suitability of RW drying for retention of heat-sensitive compounds in food produce such as aonla with better quality retention and morphological characteristics.


Introduction
Drying is the most common method used for preserving food and related products.Its primary purpose is eliminating free water, leaving only the essential bound water in the product.The reduced water content protects the food against microbial attacks and preserves the heat-labile bioactive compounds present in the food matrix, such as phenolic compounds, ascorbic acid, antioxidants, etc.These bioactive compounds are indicators of successful food drying.If ascorbic acid is well-preserved during drying, other nutrients are also retained well. [1]Therefore, ascorbic acid is considered a marker for nutrient retention during drying.Freeze drying (FD) is the most suitable technique for preventing ascorbic acid degradation; however, it is not energy-efficient.Refractance window (RW) drying is a promising alternative to FD, as it is highly effective in preventing nutrient degradation and comparable to FD in preserving ascorbic acid.
Aonla (Emblica officinalis) is a rich source of functional and nutraceutical components with good commercial value.It is a good source of antioxidants, minerals, and polyphenols and is particularly rich in vitamin C. [2] Compared with apples and oranges, aonla has 160 and 20 times more vitamin C concentration, respectively. [3]Additionally, aonla contains a good amount of minerals and antioxidants.Despite bearing such nutritional qualities, aonla is not preferred to be consumed in its fresh form due to its high acidity and astringent taste.Therefore, most aonla fruits are processed into value-added products such as candy, jam, murabba, juice, dried slices, and aonla powders.6][7][8] Among these methods, hot air drying (HAD) is predominately used with different pretreatments, such as blanching with varying concentrations of salt, sugar solution and KMS. [4,5]However, HAD exhibits poor energy efficiency and direct exposure of hot air to products, which can compromise its quality attributes.The refractance window drying method has emerged as a novel drying technique complementary to HAD in recent decades.The heart of the RW drying technique is a thin polymeric film known as a mylar film, which bridges the food product and the heating source (hot water reservoir).The food products are spread over the mylar film, and the bottom side of the film should be in contact with hot water for a continuous heat supply. [9]The emitted infrared radiation by hot water and received by the product through the mylar film is responsible for moisture removal.RW drying greatly depends on the water temperature and thickness of the product, which regulate the drying process, with the limitation of penetration depth of infrared radiation. [10]Understanding the drying phenomena involved in RW drying, including knowledge of heat and moisture transfer, is crucial.It is also a significant concern in the drying process for the reproducibility of quality-controlled products.Accurate estimation of mass transfer parameters, such as Biot number, effective moisture diffusivity, and mass transfer coefficient, can determine the mass transfer phenomena and optimize energy utilization. [11]Several diffusion models, such as Fick's, Anomalous, Dincer and Dost's models, have been developed to understand the mass transfer phenomena and calculate mass transfer parameters.Fick's diffusion model is the widely used model based on one-dimensional diffusion equations, but it does not consider cellular-level structural changes. [12]However, in the past few decades, microstructural change during drying process had been included in mathematical modeling which also improved the effectiveness of the model for drying process.In addition to it, Dincer and Dost proposed a mathematical diffusion model for infinite slab, infinite cylinder and sphere which includes various parameters i.e., Biot number, effective moisture diffusivity and mass transfer coefficient.This model, based on criteria of Biot number, represents external and internal resistance to moisture transfer during drying process. [13]Several studies were carried out for diffusion modeling based on Dincer and Dost model for biological material in various drying methods such as hot air drying, [14,15] microwave drying, [16] infrared drying, [17] refractance window drying, [12,15,18,19] heat pump drying, [20] pulsed vacuum drying. [21]uring the drying process, the microstructure of materials undergoes changes that can be demonstrated through instrumentation analyses, such as scanning electron microscopy (SEM), X-ray diffraction (XRD), and Fourier Transform infrared (FTIR) spectroscopy.However, there is a need for more literature on the mass transfer parameters of RW drying for aonla slices, as well as the changes in the structural and morphological properties of dried aonla.Therefore, this study aims to investigate the mass transfer phenomenon and characterize the quality and microstructures of aonla under the RW drying process.

Raw material
Freshly harvested matured pale green aonla fruits (Var.-Banarasi) were obtained from the farm section of ICAR-Central Institute of Agricultural Engineering, Bhopal, India, in the winter of 2021.The matured aonla fruits had total soluble solids content of 10 ± 1 Brix (measured with a digital refractometer, Pal-1, ATAGO, Japan).The fruits were thoroughly washed under running tap water.The adhered water from surface of the fruits was soaked using blotting paper before slicing the fruits for drying in the RW dryer.The aonla fruit was divided into halves by cutting through its ridges with stainless steel knives.The kernels were then removed manually before proceeding with the slicing using a vegetable slicer (Make: Ponnmani Industries, Coimbatore, India) having 2, 4 and 6 mm blade thickness.

Drying setup
The drying experiments were carried out in a laboratory-scale continuous RW dryer (Figure 1) developed at ICAR-Central Institute of Agricultural Engineering, Bhopal.The dryer consists of a hot water reservoir (200 L), drying chamber and separate water tank of 200 L capacity having heating element, horizontal multistage centrifugal pump (CNP Pumps India Pvt., Ltd., India), J-type thermocouples and a mylar sheet of 0.25 mm thickness.The food grade mylar sheet was placed over the hot water reservoir on which monolayer of the aonla slices were spread.The steel hood with an exhaust fan (Havells Ventil Air DX 200 mm and air delivery rate-0.14m 3 /s) was provided for removing excess vapor around the drying chamber.
Uniform water temperature was maintained by continuous circulation of hot water from the water tank to the water reservoir in controlled manner.Water temperature was continuously regulated by PID controller, acting directly on K-type thermocouple dipped in water tank with automatic cutoff system to control heating element.

Drying procedure
With the combination of water temperature (75-, 82-and 90 ± 1 C) and slice thickness (2-, 4-and 6 ± 0.5 mm), total nine experiments were performed with three replications.The selection of water temperature and slice thickness was based on a thorough review of the existing literature pertaining to the study and requirements for aonla drying.The literature revealed that previous studies had employed a diverse range of water temperatures, ranging from 60 to 95 C, and varying product thicknesses of up to 6 mm.About, 20 g of aonla slices were placed on the Mylar film in a monolayer.The change in weight of the sample's during drying was measured with a weighing balance (model-TE214S, make-Sartorius, Goettingen, Germany) at every 10 min interval.The drying process was continued until a constant weight was obtained.The average of three replications was reported for calculation of dimensionless moisture ratio.
The dimensionless moisture ratio (MR) was calculated using following equation: where M o , M t , M e are the moisture content at initial of drying, at time t and equilibrium moisture content (kg water/kg dry matter), respectively.The values of M e are relatively small as compared with M or M 0 for long drying time.Thus, MR can be simplified to MR ¼ M t /M 0 .

Determination of mass transfer parameters
Dincer and Dost diffusion model was used for determination of Biot number, mass transfer coefficient and effective moisture diffusivity.The aonla slice was considered as infinite slab geometry.The assumptions of this model were (1) during drying process solid objects have constant thermophysical properties, (2) there is negligible effect of heat transfer on mass transfer phenomenon and (3) moisture diffusion is occurred in the direction of thickness. [22]imilar to unsteady heat transfer process, the Dincer and Dost model analyzes the unsteady state mass transfer for different conditions.In first condition, Bi 0.1 refers to negligible internal resistance of the material to moisture diffusivity, and this condition is not common in drying of food material.Similarly, Bi > 100 refers to negligible surface resistance in food material are most common situations.However, 0.1 Bi 100 implies finite internal and external resistance which usually occurs in practical drying applications. [22]or infinite slab, Dincer and Dost diffusion model is represented by following equation: where A n and B n are defined as: In above equations, l n is a dimensionless number representing the nth root of the transcendental characteristic equation.The Fourier number (F 0 ) and Bio number (Bi) for the slab of characteristic thickness Y are given as: When the Fourier number (F 0 ) is very small (l 2 1 F ᴼ <1.2), Eq. ( 2) deduced to the first term, i.e., MR ¼ A 1 B 1 (7)   where, 1:31þBi for 0:1 Bi 100 (8) J represents a dimensionless lag factor in Eq. ( 8); m 1 in Eq. ( 9) is characteristics root depend upon geometry of solid object and for infinite slab, and can be calculated from the following equation: Since, drying of food material follows an exponentially decreasing trend of change in weight, thus, Dincer and Dost proposed following equation by introducing Lag factor (J) and drying coefficient (S): By calculating moisture ratio from the experimental data, the lag factor and mass transfer coefficient were determined.Afterward, equating the value of lag factor in Eq. ( 8), the Biot number was estimated.Further, effective moisture diffusivity (D eff ) and mass transfer coefficient (k) were estimated by using the value of characteristics root (m 1 ) and drying coefficient (S) by the following equations: where Y is characteristic length i.e., half of the slice thickness (m).
The procedure described by Torki-Harchegani et al. [17] was followed to validate the Dincer and Dost model for all the experimental drying conditions.The values of correlation coefficient (R 2 ) and root mean square error (RMSE) were also determined.

Ascorbic acid
The spectrophotometric method [23] was used to determine the ascorbic acid content, following the extraction process described by Rajoriya et al. [12] Initially, 10 mL of the extracted sample was mixed with 2 mL of tri-chloro acetic acid (10%) and incubated for 5 min in an ice bath.Then, 2 mL of Foilin-Ciocalteu (10-fold diluted) reagent was added to the mixture, and the absorbance was measured at 760 nm using a spectrophotometer (UV-1800, Shimadzu, Japan).The ascorbic acid content was calculated using a standard curve of ascorbic acid (Sigma-Aldrich, India) and expressed as the percentage retained with respect to the fresh aonla fruit.

Total phenolic content
The total phenolic content was estimated by the method adopted from Singleton et al. [24] with slight modification.Firstly, 0.5 g of ground aonla slices were mixed with an 80% methanolic solution and centrifuged (C 24, Remi Group Laboratory Instruments, India) at 553.41 G for 20 min.The supernatant was collected by filtering the methanolic extract using qualitative filter paper (100125 R, AXIVA SICHEM BIOTECH, India).Then, 0.2 mL of the methanolic extract was mixed with 10 mL of Foilin-Ciocalteu (10fold diluted) reagent, followed by adding 8 mL of sodium carbonate (7.5%) to the solution.The mixture was incubated in dark conditions for 90 min, after which the absorbance was recorded at 750 nm and expressed in mg GAE/g dry matter.

Browning index
Browning of the aonla slices with different refractance window drying conditions were assessed by image processing method as described by Tepe et al. [25] The image of dried aonla slices was captured by digital camera in the JPEG format under uniform illumination conditions and L Ã , a Ã and b Ã color values were obtained, using MATLAB R2016a (Mathworks, Natick, MA) software.After getting L Ã , a Ã and b Ã color values following formula was used to calculate BI: The surface morphology of the samples was executed by SEM (Carl Zeiss, Ultra Plus), worked in the secondary electron mode. [26]Diffraction pattern of dried aonla slice was identified by a X-beam diffractometer (Bruker, D8 Advance, Germany) at room temperature utilizing Cu K a (1.5406 Å and step size ¼ 0.02 ) radiation in the scattering range of 10-60 in 2h scale.Crystallite sizes and crystallinity [27] were calculated by following equations: where D is the crystallite size (nm), K is shape factor (0.9), k is the wavelength of the X-ray radiation (Cu-K a ¼ 0.1541 nm), b (radian) is the full width at half maximum (FWHM) of the intense and broad peaks and h is diffraction angle, C rys is crystallinity (%), I crys and I am intensity of crystal and amorphous region, respectively.

Statistical analysis
The model parameters J and S were estimated using Solver tool of Microsoft Excel 2016 (Microsoft Corporation.Redmond, Washington, USA).The test of significance (P < 0.5) of drying process variables on mass transfer parameters and quality parameters of dried aonla were analyzed by ANOVA using JMP PRO 10 software (SAS Institute, 2005).

Mass transfer parameters
The least square curve fitting method was used for regression analysis between dimensionless moisture ratio (MR) and drying time (t) in the exponential form of Eq. ( 11) (Figure 2).The model parameters i.e., lag factor (J) and drying coefficient (S) were obtained and presented in Table 1.The lag factor J is the indicator of the magnitude of the both internal and external resistance of the product to mass transfer during the drying process.Irrespective of water temperature and slice thickness, obtained values of J were greater than 1 which implied the presence of internal resistance to moisture diffusion.The parameter S denotes the drying capability of the product per unit time.From Table 1, value of lag factor varied between 1.044 and 1.241, and drying coefficient was in the range of 4 Â 10 À4 to 1 Â 10 À3 s À1 .The values of J and S reported in other studies during convective, [28] microwave drying [16] and refractance window drying [18] were in the range of 0.873-1.134,1.045-1.398and 1.02-1.06;and 1.09-1.60Â 10 À4 , 1-3 Â 10 À4 and 1.13 Â 10 À3 -4.5 Â 10 À4 s À1 , respectively.The values of J decreased and S increased in response to a rise in water temperature and decrease in slice thickness, except at 90 C water temperature where 4 mm slice thickness had higher S value than 2 mm slice.This can be explained as, increase in water temperature and a decrease in slice thickness has tense the mass and heat transfer between the heating medium and wet food product kept on the mylar film, which enhances the drying capability.
The mass transfer parameters were estimated by Dincer and Dost diffusivity model and are presented in Table 2. ANOVA was used to test the significance of RW drying process variables on the mass transfer parameters.Biot number is the most important mass transfer parameter for Dincer and Dost diffusivity model, and the criteria of the model were (1) 0.1 Bi 100 and (2) Bi > 100.RW Drying of aonla slice fell under the first criteria, commonly occurring in most heat mass transfer operations related to food.The value of the Biot number obtained was between 0.268 and 8.66, indicating the presence of both internal and external resistances to moisture transfer, which can be pertained to the results obtained from drying poria cubes [21] and lactose powder. [16]The lowest value of Biot number was obtained for 2 mm slice thickness and at 90 C (Table 2), experiencing more internal moisture resistance than surface resistance.A similar result has been reported by Rajoriya et al. [19] Further, increase in slice thickness and reducing water temperature resulted in a higher Biot number, and both the variables had a significant effect (P 0.05) on Biot number.Apart from drying conditions, the type of dryer and product also influences the Biot number. [11]Generally, the Biot number is less in the convective drying process and hikes in microwave drying.Whereas, in refractance window drying the value of Biot number were less than that of convective drying.This can be confirmed by the results reported by McMinn [16] who compared microwave and convective drying for lactose powder, and Rajoriya et al. [12] who compared refractance window drying with convective drying.
Mass transfer during the drying process of biological materials is governed by moisture diffusivity, which is a fundamental property based on molecular interactions and the physical structure of the material. [10]The effective moisture diffusivity was calculated as per Eq. ( 12), a function of drying time, slice thickness and Biot number.The D eff for aonla slices was found to vary from 4.27 Â 10 À10 to 8.64 Â 10 À09 m 2 s À1 .These values were found to be higher than the diffusivity values of 5 mm thick aonla shreds (1.02-1.73Â 10 À10 m 2 s À1 ) obtained in thin layer convective hot air drying process. [29]It was observed that D eff values were lower at 75 and 82 C water temperature compared to 90 C (Table 2).ANOVA indicated that the interaction term of water temperature and slice thickness had significant effect on effective moisture diffusivity (P 0.05).In terms of individual parameters, water temperature had more effect on diffusivity as compared to product thickness.Elevating water temperature and reducing slice thickness led to faster moisture removal in general and resulted in high  value effective moisture diffusivity.This attributed that higher drying temperature affects the surrounding vapor pressure, and consequently increasing the moisture diffusivity during drying.Similar findings were also observed during the drying of kiwi slices, [20] blueberry [15] and saffron stigmas. [17]he mass transfer coefficient (k) can be described by the interfacial mass transfer through liquid or solid to gases which are affected by the thermal environment created around the wet product due to the synergic compliance of heat source (water temperature), properties of mylar film and biological material, and thicknesses of the thermal and concentration boundary layers.In the present study, the moisture transfer coefficient values ranged from 1.60 Â 10 À6 to 9.84 Â 10 À6 ms À1 .The values of the moisture transfer coefficient reported by earlier studies [11,16,17] are comparable to those found in the present study.As shown in Table 2, the mass transfer coefficient was found to be highest for 6 mm thick slices at all the water temperatures.Also, increasing water temperature contributed to an average increase in k value by about 27%.The variation of the mass transfer coefficient with a combination of water temperature and slice thickness results in varying drying time (Figure 2) and signifies that it possesses mass transfer characteristics.Thus, k can be referred to as the rate of absorption of water vapor by a drying medium.

Validation of the model
The Dincer and Dost model was validated for all the drying conditions of refractance window drying process for aonla slices.The variation of dimensional less moisture ratio with drying time, for model and experimental values are shown in Figure 2. The estimated value of R 2 was more than 0.95 indicating the closer agreement between model and experimental values and the accuracy of model is denoted by the lower values of error (Table 1).This shows that Dincer and Dost model is very suitable for estimating mass transfer parameters of aonla slices during refractance window drying.

Quality parameters
Aonla is highly regarded for its rich antioxidant properties, mainly due to its high content of ascorbic acid and phenolic compounds.The retention of ascorbic acid and total phenolic content was calculated to assess the quality of the dried product obtained through refractance window drying.The impact of water temperature and slice thickness on the retention of ascorbic acid (AAR) and total phenolic content (TPC) is presented in Table 3.
Higher water temperature led to less degradation of ascorbic acid because it reduced the exposure time of the product to heat, thus preventing the oxidation of ascorbic acid.Increasing the water temperature resulted in an average increase in ascorbic acid retention of 7-15%.However, the effect of slice thickness on ascorbic acid retention showed a slightly different trend.The maximum retention was observed for 4 mm thick slices, while degradation was higher for 2 mm and 6 mm thick slices.The possible reason is the larger cut-surface area per unit volume exposed to the ambient environment in 2 mm slices, leading to an accelerated loss per unit weight.On the other hand, the longer drying time required for 6 mm slices allowed for more interaction of heat-sensitive compounds with the thermal environment, resulting in ascorbic acid degradation.Similar findings have been reported in red pepper [30] and sapota. [31]The individual and synergic effects of both the drying parameters had a significant effect on the ascorbic acid retention (P < 0.05).
Regarding total phenolic content, RW dried aonla slices had concentrations ranging from 20.81 ± 0.14 to 37.84 ± 0.08 mg GAE/g dry matter.The highest phenolic content was obtained for 4 mm thick slices dried at 82 C water temperature, while the lowest concentration was observed at 90 C water temperature and 6 mm slice thickness (Table 3).The difference in total phenolic content under different drying conditions was approximately 17 mg GAE/g dry matter.Slice thickness significantly affected phenolic content, with 4 mm slices having a higher concentration.However, increasing water temperature had a negative impact on phenolic content specially in thin slices (2 mm thickness).It can be explained by the fact that lower drying temperatures result in less inactivation of oxidative enzymes and may initiate the Maillard reaction, which liberates phenolic compounds. [32]Therefore, the combined effect of both drying parameters minimized the oxidation of the phenolics compound present in aonla.It may be owing to the formation of building blocks of phenolic compounds by the inactivation of oxidative enzymes due to the thermal treatment. [33]he browning index of aonla slices varied from 53.39 ± 0.61 to 67.94 ± 0.68% under different drying conditions (Table 3).Non-enzymatic browning processes such as Maillard reaction, ascorbic acid degradation, phenolic compounds' oxidation, and caramelization are typically more pronounced at higher drying temperatures. [34]However, in the present study, the browning of aonla slices was less at higher water temperatures.It can be attributed to the unique nature of the refractance window drying technique, which creates a thermal "window" on the wet product that gradually closes with moisture removal, protecting the product from overheating and preventing excessive browning. [10]The lowest browning index value was observed at 90 C water temperature and 4 mm slice thickness, which coincided with the maximum ascorbic acid retention.When the water temperature was kept constant, and the slice thickness varied, the browning index ranged between 8 and 9%.
The highest value of the browning index was observed for 6 mm slices, which can be attributed to the longer drying process, consistent with Figure 2. Similar observations have been reported by Li et al. [33] and Shewale et al. [35]

Instrumental characterization of refractance window dried aonla
Instrumental characterization was performed to analyze the change in structural behavior of aonla slices, in three different aspects; morphological behavior (SEM), crystallinity (XRD) and bending or stretching of molecules (FTIR).The drying process alters the morphologic behavior of the product at microscopic level which eases the mass transfer phenomenon.The influence of water temperature and slice thickness on the microstructure is presented in SEM micrographs (Figure 3).The microstructure of RW dried aonla slices was observed to be smooth and flaky with sharp edges.The cellular integrity of biological material can elaborate by how tightly or loosely cell/cell membrane or cell walls are intact to the surface. [36]It can be postulated from the SEM micrographs that the intactness of the cells is more pronounced in thick (i.e., 6 mm) aonla slices, represented by the presence of a more significant number of adjacent and complex cell walls.It can be attributed to the transport phenomenon of liquid water, which is responsible for the dragging out of soluble solids such as sugars, minerals, vitamins and nonvolatile compounds with water to the outer surface, due to which cell membrane/tissue forms clusters. [37]Also, a prolonged drying process and less heat penetration occurred in thick slices, which caused a low-temperature gradient between the surface and interior of the product, leading to slow moisture migration.As a result, the disintegration of cell membrane/walls occurred at a microscopic scale (Figure 3c).With the increase in water temperature, the adjacent cells start to disintegrate, resulting in sharp edges.Moreover, the rate of moisture migration was accelerated in thin slices with the combined effect of cellular expansion pressure raised due to the heat source (i.e., water temperature), which impacted the cell/tissue membrane.It causes rapid tissue dehydration, cell wall rapture and the deformation in cellular texture. [38]It can be visible by the increasing intercellular space, breakdown of adjacent cell walls and appearance of cracks in Figures 3a,b.The disruption of microstructures at high water temperatures may be caused by escaping water vapor from the interior part of the aonla slices during the drying process. [39]On a similar note, the high magnitude of diffusivity was obtained at 90 C water temperature (Table 2), which may hasten the diffusion of gases and liquids.Escaping of vapor in the RW drying process emerges from tissue, which could become a more porous structure.
The XRD analysis also revealed that RW-dried aonla slices had semicrystalline nature.The X-ray diffraction pattern provides information about the crystallinity and amorphous nature of a food product and reflects the product's stability under changing process conditions.X-ray diffraction patterns are used to infer the structural behavior based on the diffraction angle, the intensity of the peaks, half bandwidth of the diffraction peak, crystalline percentage, and crystalline size.All these parameters of aonla slices obtained from refractance window drying process are presented in Table 4.The X-ray diffraction pattern is represented by the graphical plot between diffraction angle (2h) and intensity (Figure 4); the sharp peaks indicate crystallinity, while broad peaks indicate amorphousness.The  prominent peak corresponding to crystallinity was around 19.62-20.32, assigned to 10 lattice planes of cellulose component; this was also reported in Moringa oleifera leaf and sugar beet powders. [40,41]On a similar note, some other peaks were also observed at around 34 and 37 , such diffraction patterns were also observed in food powder of sapota and guava, which reveals the A-type diffraction pattern of semicrystalline structure. [41]he peaks corresponding to an amorphous region of RW dried aonla slices were observed in the range of 16.13-19.01(2h diffraction angle) under varying drying conditions.The diffraction angle increased with slice thickness, attributed to the halo of amorphous substances in aonla slices.It may be caused due to comparatively slow moisture migration in thick slices and non-crystalline substances forming hydrogen bonds at the surface of the crystalline region. [27]The crystallinity of aonla slices ranged from 22.11 to 29.32%, similar to Moringa (leaf, flower, seed) and sugar beet powders. [41,42]Crystallinity of aonla slices was enhanced by increasing slice thickness at varying water temperatures, probably due to agglomeration of polysaccharide and crystallization of non-crystalline substances, which can be depicted by the slight variation in half bandwidth of diffraction peak [27,42] with drying conditions.Table 4 shows that water temperature had a negative correlation with crystal size, which was between 49.11 ± 0.26 and 99.12 ± 1.19 nm.High water temperature transformed the crystalline region to amorphous, which can also be seen in Figure 4 that more significant peaks in the amorphous region were found at higher water temperature.Raaf et al. [43] also found similar observations for the hot air drying of aonla slices.Thus, the XRD pattern suggests that RW dried aonla slices had a crystalline-amorphous or semicrystalline structure described by a succession of sharp peaks functioning as a crystal and a diverse array of peak patterns functioning as an amorphous.
FTIR analyses the absorbance of mid-IR radiation of dipole bonds inside functional groups of molecules, causing chemical bonds to bend or stretch.All the peaks of FTIR obtained in the present study are presented in Table S1 (provided in the supplement file).In the present study, the entire range of IR spectroscopy is divided into three regions: 4000-2900 cm À1 (region A), 1700-1500 cm À1 (region B) and 1500-500 cm À1 (region C).No significant peaks were observed in the region between 2900 and 1700 cm À1 .Region C belongs to the characteristic fingerprint of the molecule, which gives useful information for identifying functional groups mainly governed by the bending of molecules due to the vibration.Two prominent peaks were found in Region A. The first peak in this region is broad and ranges from 3245 to 3577 cm À1 , corresponding to the absorption due to stretching of the -OH bonds of bound water, [44] observed in all the drying conditions.The broadness of the peak decreased at high-water temperatures and  for thin slices due to the evaporation of water.
Figure 5 shows that -OH bonds stretched more in 6 mm slice thickness at 75 C water temperature, while minimum stretching was observed in 2 mm slice thickness at 90 C water temperature.It may be caused due to high moisture content present in slice thickness of 6 mm at 75 C water temperature, while lowest in 2 mm slice thickness at 90 C water temperature.The combined effect of low water temperature and thick slices attributed to the stretching of -OH bonds in the region of 3245-3577 cm À1 .The second peak, centered around wavenumber 2930 cm -1 indicates the C-H stretching of carboxylic acid [45] that did not have any significant difference with varying drying conditions, as can be observed from Figure 5 that peak intensity is similar for all the conditions.It can be postulated that the C-H stretching of carboxylic acid was not affected significantly by refractance window drying process.In region B, three major peaks were centered around 1725, 1624 and 1570 cm À1 .Peaks around the wavenumber 1725 cm À1 are attributed to carbonyl groups of hydrolyzable tannins. [46]Comparatively, low-intensity peaks were observed at 2 mm slices while considering water temperature at 90 C water temperature.It indicates that minimum hydrolyzable tannins were found in a sample obtained at 2 mm slice thickness and 90 C water temperature.The peak centered around 1624 cm À1 denotes the stretching vibrations of C ¼ O from phenolic compounds and the N-H bending vibration of amine or amide groups. [47]The most extended peak was found in a 4 mm slice followed by 2-and 6mm slice thickness, while the variation with water temperature longer peak length was found in 82 C water temperature followed by 75 and 90 C. It indicates that aonla slices obtained at 4 mm slice thickness and 82 C water temperature possess the highest amount of phenolic compounds, supported by the quality analysis (Section 3.4).The presence of ascorbic acid is characterized by the O-H stretching vibration, the presence of the ester group, and the C ¼ O stretching vibration corresponding to wavenumber 3577, 1724 and 1624 cm À1 , respectively. [47,48]From Figure 5, the peak intensity corresponding to these wavenumbers were high in a sample obtained at 4 mm slice thickness of aonla having water temperature 90 C, confirming the maximum ascorbic acid content.
Region C is considered a fingerprint region in IR spectroscopy, and more peaks were obtained.The significant peaks of refractance window dried aonla slices obtained in this region are centered at around 1450, 1350, 1225, 1060, 950 and 860, 765 cm À1 .The peaks corresponding to wavenumber 1450, 860, and 765 cm À1 are attributed to the strong bending of the C-H bond of methyl group; wavenumbers 1350 and 1060 cm -1 indicate the stretching of C ¼ O while 1225 and 950 indicate the stretching of C-O and C ¼ C bonds respectively. [49]A significant peak shift was observed below 1060 cm -1 .For a slice thickness of 6 mm, a peak shift from 915 to 954 cm À1 was observed at a water temperature of 82 ᴼC.Considering 4 mm slice thickness, the peak shifted from 951 to 918 cm À1 ; and 817 to 862 cm À1 were observed at 75 C water temperature.However, 2 mm slices with varying water temperatures had negligible peak shifts.Overall, absorption of IR radiation in the spectrum region of 950-750 cm À1 increased gradually from 2 to 6 mm slice thickness, which can be attributed to the significant increase in signals from C-O-C, C-O, C-C and C-O-H vibrations. [50]hus, shifting of peaks, broadness of peaks, and intensified absorption of IR radiations by molecules correspond to all three regions indicating the synergic effect of water temperature and slice thickness on the structural variation of molecules present in refractance window dried aonla slices.

Conclusions
Refractance window drying is a thin polymeric film based drying technique and is equipped with simple components.In this study the RW drying process parameters such as the water temperature and slice thickness controlled the product quality.The ascorbic acid and phenolic content of aonla were preserved well at higher water temperature.However, thick aonla slices required more drying time resulting in thermal degradation.The study of heat or/and mass transfer parameters helps in reproducibility of quality-controlled products.Dincer and Dost diffusion model accurately estimated the mass transfer parameters (Bi, D eff and k), which were in the range with other food material dried under various methods.The validation of the model resulted in close agreement with experimental values (R 2 > 0.95) having minimum value of error (RMSE).The Biot number of more than 1 observed in this study signified finite internal and external resistance which was closely controlled by water temperature and slice thickness.The mass transfer phenomenon occurred during drying process altered the microstructure of the product, which revealed that in thick aonla slices, the cell walls/membrane and adjacent surface were tightly intact to each other by agglomeration of polysaccharide and crystallization of noncrystalline substances, as the result of its diffusion was comparatively difficult in thick slices.However, intactness in microstructures were loosen at high water temperature which was also observed in XRD pattern that high water temperature transforming the crystalline region to amorphous region.Small variation was observed in bending or stretching of molecules by the absorption of IR radiations of aonla slices.

Figure 2 .
Figure 2. Experimental and predicted moisture ratio for (a) 90 C, (b) 82 C and (c) 75 C water temperature.

Table 1 .
Model parameters for refractance window drying of aonla slice.

Table 2 .
Mass transfer parameters for refractance window dried aonla slice as estimated by Dincer and Dost model.

Table 3 .
Quality parameters of RW dried aonla slice.

Table 4 .
Crystallinity (C rys ) and size of crystal (D) of RW dried aonla slices.