Abstract-- The purpose of this study was to optimize ohmic heating (OH) conditions using response surface methodology (RSM) for maximum recovery of polyphenols from industrial chestnut peel waste. Box-Behnken design (BBD) was applied to investigate the effects of three independent variables, namely electric field (V/cm), heating time (sec) and salt concentration (%). From RSM, optimum pretreatment conditions were obtained as 20 V/cm, 100 sec and salt concentration of 0.32%. The peels of three chestnut varieties (Sarıaşı, Düzce and Kiraz) were pretreated by these optimal conditions and their polyphenols were extracted using different solvents. Polyphenol contents and antioxidant capacities of peel extracts were dependent on the variety and extraction solvents used. HPLC analysis showed that ellagic acid was the most abundant phenolic compound in all varieties. OH pretreatment had a positive effect on extraction. Additionally, alcoholic extraction had a better recovery effect compared to water. The peel waste of Kiraz variety was found to be a good source for polyphenols.
Keywords-- Chestnut, ohmic heating, optimization, polyphenols, response surface methodology.
I. INTRODUCTION
In the food industry, synthetic antioxidants are commonly used to increase the stability of foods by preventing lipid peroxidation and other oxidation reactions (Babbar et al., 2014). However, they have potential health hazards due to their toxic and carcinogenic effects (Amado et al., 2014). In recent years, increased attention has been paid to alternative natural antioxidants. Therefore, antioxidant potential of many plants and by-products from different natural sources has been investigated (Franco et al., 2018). The agro-food industries generate huge quantities of wastes and consist of organic residues of the processed raw materials. The polyphenols extracted from fruit or vegetable wastes have been used as natural food preservatives as they extend the shelf life of the final product and increase antioxidant capacity.
The chestnut (Castanea sativa) is one of the most remarkable fruit crops due to its economic importance (Aires et al., 2016). The chestnut processing produces large amount of waste as outer and inner shells around 10-15% by weight of the whole chestnut (Cerulli et al., 2020; Vella et al., 2018). The chestnut peels are well-known to be rich in phenolic compounds which show anticancer and antimicrobial effects as well as antioxidant activity (Cacciola et al., 2019; Lee et al., 2016; Sorice et al., 2016). Therefore, the peels might be an economic source of polyphenols, which have great potential in industrial applications as food additives/ingredients or supplements of high nutritional value.
Appropriate and effective extraction is very important to recover maximum amount of polyphenols from natural sources and affected by several factors such as time, solvent type, solid to solvent ratio and pretreatment of sample (Ćujić et al., 2016). Therefore, it is necessary to optimize extraction conditions to obtain the highest extract yield. RSM is a mostly used statistical tool to evaluate the effects of various factors and interactions between them (Liu et al., 2018). Recently, there has been growing interest in the use of novel electrotechnologies such as the application of pulsed electric fields (PEF) or ohmic heating (OH) in order to enhance the extraction of valuable bioactive compounds from different plant materials (El Darra et al., 2013). OH is based on alternating electric current through the food causing the generation of internal heat (Salari and Jafari, 2020). There have been some studies which reported that OH resulted in increasing the extraction yields of bioactive compounds such as carotenoids from tomato by-product (Coelho et al., 2019); inulin from artichoke (Termrittikul et al., 2018) and polyphenols from red grape pomace (El Darra et al., 2013). However, no published information is available in the literature related to application of OH for polyphenol extraction from chestnut peel and its optimization for the maximum polyphenols recovery using the response surface methodology (RSM).
In this study, polyphenols of the peels from three Turkish chestnut varieties (Sarıaşı, Düzce and Kiraz) with high economic interest for industrial production were extracted using different extraction conditions after the peels were pretreated by OH. Pretreatment parameters were optimized using the Box-Behnken design of RSM. Total polyphenol content (TPC) and antioxidant capacity (AC) of the extracts obtained using three solvents (aqueous ethanol, methanol and water) were analyzed. Individual phenolic compounds of the extracts were also identified and quantified by HPLC.
II. METHODS
Chestnut fruit (C. sativa) peels (a mixture of the outer brown peel and the inner skin) were supplied in December of 2019 by chestnut processing plant in Turkey. The peels were generated from three chestnut varieties (Sarıaşı, Düzce and Kiraz). The latitude and longitude of the regions where these chestnuts were collected were Ketendere (Aydın) 38° 0' 35.0172'' N, 28° 21' 34.5204'' E; Umurlu (İzmir) 38° 7' 18.4908'' N, 28° 26' 9.6576'' E and Yeniyer (Düzce) 40° 59' 35.7468'' N 31° 21' 43.5132'' E in Turkey, respectively.
The peels were filled in the polyethylene bags and kept at 4 ºC before the experiments were conducted. All chemicals and reagents used in the study were either HPLC or analytical grade.
A. Pretreatment of chestnut peel by OH
The pretreatment of the peels (Düzce) was carried out in OH chamber which consists of a rectangular Cast Polyamide/PA6G (15 × 6.6 × 8 cm) and two planar AISI 304 stainless steel electrodes (14.5 × 8 cm). The chamber had a capacity of 500 mL. Temperature was measured with type-K thermocouples coated with teflon to prevent interference from the electrical field which were inserted into the center of the sample. The electrodes of OH were connected to a variac (50 Hz, 0-600V, 25A) (Artsan Energy and Test Instruments, Turkey). All the output data (current, voltage and temperature etc.) were recorded at 1s intervals on a data logger with special software and monitored in the PC (İncedayi, 2020).
Peel samples were placed between two stainless steel electrodes inside the treatment chamber. The sample to liquid ratio in the treatment chamber was 1:20 (w/w). The distance between electrodes was adjustable and fixed according to electric field strength defined in Table 1. Table salt solution at a known concentration was added to the chamber to assure a better contact between electrodes and sample. The electric field strength within the treatment chamber was calculated as follows:
Electrical field strength (E) = (Output voltage (V) )⁄
(Distance between the electrodes (cm) ) (1)
After heating (treatment) time period, samples were drained off rapidly, rinsed, cooled in running water and air-dried at room temperature until they reached constant moisture content for approximately 4-5 hours. Peel samples were milled, passed through a 2 mm sieve and stored at 4 °C before experiments.
B. Pretreatment of chestnut peel by OH
In this study, Box–Behnken design (BBD) was employed to determine pretreatment parameters for polyphenols from waste of chestnut peel. The current design comprised 30 experimental runs with three levels, −1 (lower limit), 0 (central point) and +1 (upper limit), for each factor. A (time), B (S/S ratio), and C (temperature) were chosen as the independent variables whose selection and range were based on previous studies while response variable was TPC. Table 1 shows the experimental design (coded and actual values of the factors) for each run. The experiments were performed in duplicate and average of the duplicate has been taken as a response. The experimental data were fitted to the following second-order polynomial model:
(2)
here Y is the predicted response; 
is the model
intercept coefficient; 
are the
regression coefficients for the linear effect terms; 
and 
are the square
effect terms; and 
and 
are the
interaction effect terms, respectively. A, B and C are the independent
variables (Table 1). To test the predicted model on the response variable, an
analysis of variance (ANOVA) with 95% confidence level was performed to
evaluate the effect of each factor. Besides, the regression coefficient (R2)
and the p-value of the lack of fit (LOF) were employed to evaluate the fitness
of the regression model. The relationship between the independent variables and
the response variables was demonstrated by the response surface plots and the
optimum conditions were determined. In order to find out the accuracy and
suitability of the optimized conditions, additional experiment was performed
under optimal conditions. For data analysis, MINITAB 17 software (State
College, PA) was used.
C. Extraction of polyphenols
Ground peel pretreated was extracted with distilled water and organic solvents to investigate the effect of different extracting solvents on TPC and AC of peel. For highest yield of polyphenols, extraction conditions applied were optimized using response surface methodology (data not shown). For water extraction, the sample was extracted with distilled water in a falcon tube in a controlled water bath. Extraction was performed at solid to solvent ratio of 1/39.70 (w/v) and temperature of 60 °C. For organic solvent extraction, the sample was extracted with 60% of ethanol and 60% of methanol in a falcon tube on a mechanical shaker. Solid to solvent ratio was 1/10 (w/v) for both solvents. The falcon tube was wrapped with aluminum foil to prevent degradation during extraction. After extraction time of 22.02, 82.41 and 116.97 min for water, ethanol and methanol, respectively, the mixture was filtered through Whatman No.1 filter paper. Water extract was also rapidly cooled under tap water. All clear extracts were stored at –18 °C until analyzed. Each extraction was carried out in triplicate.
D. Extract analyses
Total polyphenol content (TPC)
TPC of the extract was determined using the Folin-Ciocalteu method (ISO 14502-1, 2005). A calibration curve of gallic acid (5-50 µg/mL) was prepared and the results determined from regression equation of the calibration curve (R2=0.99) were expressed as mg gallic acid equivalents (GAE) per gram of dry matter.
Antioxidant capacity (AC)
AC was determined by the 2,2,diphenyl-2-picryl- hydrazyl (DPPH) method of Türkmen Erol et al. (2009). It was calculated as percentage inhibition (I, %) of the DPPH radical by the following equation:
(3)
where 
is the
absorbance of the DPPH solution without sample and 
is the
absorbance of the test sample. I (%) of samples was converted to
ascorbic acid equivalent (AAE) defined as mmol of ascorbic acid equivalents per
100 g of DM.
E. Determination of phenolic profile of the peel extracts
The identification and quantification of individual polyphenols in the extracts were performed on a HPLC system including LC-20 AD Shimadzu pumps, a CTO-10 ASVP column oven and SPD-M20A photo diode array (PDA) detector, a Shimadzu DGU-20A5R degasser and SLC-10 A VP system controller. A computer-controlled system with LC solution software was employed for data analysis. The column used was a C18 reversed phase Nova Select (250 × 4.6 mm ID, 5μm) and was operated at 25 °C. UV spectra were recorded from 190-370 nm and peak areas were measured at 270 nm. The two mobile phases used for gradient HPLC elution were (A) 0.1% orthophosphoric acid in water (w/v) and (B) acetonitrile. The gradient elution profile was as follows: from 0 to 6 min, 7% B; from 6 to 82 min, 7-38% B; from 82 to 90 min, 38-60% B. The column was re-equilibrated with the initial conditions for 5 min before the next injection. The flow rate was 1.0 mL/min. The injection volume was 20 μL.
Chromatographic peaks in the samples were identified by comparing their retention times and UV spectra with those of their reference standards and by co-chromatography with added standards. Quantification was performed from the peak area of each component and its corresponding calibration curve.
F. Statistical analysis
Experimental results were expressed as means ± standard deviation of triplicate measurements and analyzed by SPSS software (SPSS statistics 23, IBM.2015). Analyses of variance were performed by one-way and two-way ANOVA procedures. Means were compared by using Tukey multiple comparison test. Values of p < 0.05 were considered as significantly different (α=0.05).
III. RESULTS AND DISCUSSION
A. Analysis of the model
The responses of TPC and AC of the peels pretreated are presented in Table 1. On the other hand, those of TPC and AC of untreated peel sample (control) were 23.02 mg GAE/g DM and 18710.62 mmol AAC/100 g DM, which were lower than treated ones. According to the results TPC and AC values of untreated peel sample (control) were lower than treated ones. This finding is in agreement with the study of Coelho et al. (2019) who reported that OH resulted in a 58% higher recovery rate of polyphenols from tomato by-products than control samples. The difference may be explained by that OH treatment can cause changing permeability of cell membranes and structural damages, enhancing the release of phenolic compounds (El Darra et al., 2013). As seen in Table 1, the highest TPC and AC were obtained with electric field of 20 V/cm. Sarkis et al. (2013) reported that degradation of anthocyanin, a type of flavonoids, during OH at low voltages was lower, which is in agreement with the result of this study.
The ANOVA analysis for TPC and AC of peel extracts indicated that the model was highly significant (p < 0.05) with high values of determination coefficients (R2) (Table 2). This result represents a good correlation between the experimental and predicted data for TPC and AC of peel extracts (R2>0.99). For the model responses, lack of fit did not result in a significant p values, indicating that the model is sufficiently accurate for predicting the responses.
B. Effect of pretreatment parameters on TPC and AC of chestnut peel extracts
As seen in Table 2, linear and square terms of electric field had significant effect on TPC (p < 0.05), while linear term of salt solution and square term of extraction time were significant (p < 0.05). The most significant linear variable is electric field with F value of 347.31 while time was the most significant square one with F value of 6.94. The interaction between electric field and time was also significant (p < 0.05). The predicted model for TPC was given below:
(4)
With respect to AC, linear term of electric field was significant (p < 0.05) while square terms of salt solution and extraction time were significant (p < 0.05). None of the interaction terms had significant effect on AC (p > 0.05). The predicted model for AC was:
The positive values in the models indicate that an
increase of factors tends to increase the response values; on the other hand,
negative values indicate that an increase tends to decrease the responses. As
seen in Eq. 4 and 5, TPC and AC decreased with increasing electric field due to
term but
increased with increasing salt concentration and time due to the 
and 
terms. Square
and interaction terms of the independent variables had also positive and
negative values.
The response surface and contour plots were used to
demonstrate the effects of electric field, salt concentration and the time on
TPC (Fig. 1). A significant effect was produced on maximization of TPC by the
two-factor interaction between electric field and time (p < 0.05). As seen
in Figs. 1a and 1b the lowest electric field (code=-1) and the highest OH time
(code = +1) yielded highest TPC. This result could be explained by degradation
of polyphenols at higher voltages due to their being thermally sensitive and
unstable (Coelho et al., 2019). Because, increasing voltage gradient
(electric field strength) intensify electric current resulting in the ions and
charged molecules to move more and faster, so the higher the temperature
(Salari and Jafari, 2020). As agreement with the result of this study, Teh et
al. (2014) found that increasing of voltage decreased the content of
phenolics in the extract. Additionally, Demirdöven and Baysal (2014) reported
that contents of ascorbic acid of orange juices treated by ohmic heating at 69
°C (42 V/cm) and 70 °C (44 V/cm) were found 45.2 and 43.1 mg/100 mL,
respectively. Pereira et al. (2016) found that the yield of
anthocyanins, one of the most important groups of phenolic
