Application of response surface methodology for optimizing the recovery of phenolic compounds from hazelnut skin using different extraction methods

Industrial Crops and Products 91 (2016) 114–124

Contents lists available at ScienceDirect

Industrial Crops and Products journal homepage: www.elsevier.com/locate/indcrop

Application of response surface methodology for optimizing the recovery of phenolic compounds from hazelnut skin using different extraction methods Halil I˙ brahim Odabas¸ a,b,∗ , Ilkay Koca b a b

Gümüs¸hane University, Faculty of Engineering and Natural Sciences, Department of Food Engineering, Gümüs¸hane, Turkey Ondokuz Mayıs University, Faculty of Engineering, Department of Food Engineering, Samsun, Turkey

a r t i c l e

i n f o

Article history: Received 16 December 2015 Received in revised form 16 May 2016 Accepted 20 May 2016 Keywords: Hazelnut skin Phenolic Antioxidant activity Ultrasound-assisted extraction Microwave-assisted extraction Supercritical fluid extraction Response surface methodology

a b s t r a c t Hazelnut skin which is a rich source of phenolic compounds is a by-product of hazelnut roasting process. In present study, ultrasound-assisted extraction (UAE), microwave-assisted extraction (MAE) and supercritical carbon dioxide extraction (SCE) methods were applied for recovery of phenolic compounds from hazelnut skin. Response surface methodology was used to estimate optimum extraction conditions. Extraction studies were performed according to Box-Behnken Design. Temperature (UAE)/power (MAE), time and ethanol concentration were independent variables for UAE and MAE. The independent variables selected for SCE were temperature, time and pressure. The optimum extraction conditions were 45 min and 67.2–67.6% ethanol concentration, 600 W, 6 min and 55.03–56.23% ethanol concentration and 42.72–49.10 ◦ C, 59.83–60.00 min and 10.01–11.48 bar, for UAE, MAE and SCE respectively. Total phenolic content, FRAP and 1/EC50 values at the optimum conditions were 122.99–123.01 mg GAE/g, 612.20–613.25 mmol Fe(II)/g and 4.36 mL/mg for UAE, 111.53–111.55 mg GAE/g, 582.44–582.52 mmol Fe(II)/g and 2.48 mL/mg for MAE and 69.59–72.64 mg GAE/g, 426.25–465.52 mmol Fe(II)/g and 2.18–2.27 mL/mg for SCE, respectively. Maceration was performed for comparison with novel methods. UAE was found to be the best method of the extraction of phenolic compounds from hazelnut skin with the highest total phenolic content and antioxidant activity values. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Hazelnut (Corylus avellana L.) which belongs the Betulacae family is an important ingredient for processed foods, such as bakery, confectionery and dairy products. Only a small quantity of annual hazelnut production (∼10%) is consumed raw (Dervisoglu, 2006; Schmitzer et al., 2011). Hazelnut processing, which includes harvesting, cracking, shelling/hulling, and roasting processes, generates by-products such as hazelnut skin, hazelnut hard shell, hazelnut green leafy cover and hazelnut tree leaf (Shahidi et al., 2007). Hazelnut skin is a by-product of roasting process and represents about 2.5% of the total hazelnut kernel weight (Alasalvar et al., 2009). Currently, hazelnut skin has no commercial value. There-

∗ Corresponding author at: Department of Food Engineering, Faculty of Engineering and Natural Sciences, Gümüs¸hane University, 29100, Gümüs¸hane, Turkey. E-mail address: [email protected] (H.I˙ . Odabas¸). http://dx.doi.org/10.1016/j.indcrop.2016.05.033 0926-6690/© 2016 Elsevier B.V. All rights reserved.

fore, finding a feasible way to evaluate this waste product has great importance for the hazelnut industry. Previous studies have shown that hazelnut and hazelnut byproducts are rich sources of phenolic compounds (Alasalvar et al., 2006; Shahidi et al., 2007; Stevigny et al., 2007; Contini et al., 2008; Alasalvar et al., 2009; Del Rio et al., 2011; Altun et al., 2013). In addition, the majority of hazelnut phenolics are located in the hazelnut skin (Shahidi et al., 2007). The high antioxidant activity of hazelnut skin extracts were measured as 2,2-diphenyl-1-picrylhydrazyl (DPPH) scavenging activity (Shahidi et al., 2007; Contini et al., 2008; Contini et al., 2009; Alasalvar et al., 2009; Locatelli et al., 2010; Contini et al., 2012; Montella et al., 2013), hydrogen peroxide and superoxide radical scavenging activity (Shahidi et al., 2007), antiperoxyl radical efficiency (Contini et al., 2008), ferrous chelating capacity and ferric reducing ability (Contini et al., 2009) in different studies. Hazelnut skin phenolic extract has better DPPH radical scavenging activity than most common natural (␣tocopherol) and synthetic (BHA, BHT) antioxidants (Contini et al., 2008; Contini et al., 2012). Furthermore, in vivo studies indicate

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

that the phenolic extracts obtained from hazelnut skin are biologically active in rats (Contini et al., 2009). Hazelnut skin phenolics are formed mainly monomeric and oligomeric flavan-3-ols (95%). Flavonols, dihydrochalcones (3.5%) and phenolic acids (<1%) are other phenolic subclasses in hazelnut skin (Del Rio et al., 2011). Flavan-3-ols exhibit some beneficial effects to human health with their antioxidant, anticarcinogen, cardiopreventive, anti-microbial, anti-viral and neuro-protective properties (Aron and Kennedy, 2008). Thus, the phenolic extract obtained from hazelnut skin could potentially be used as a natural antioxidant, functional ingredient and dietary supplement in food and pharmaceutical industries (Shahidi et al., 2007; Alasalvar et al., 2009). The conventional methods for extracting phenolic compounds are commonly solvent extraction, heat reflux and Soxhlet extraction (Bai et al., 2010). Consumption of a large amount of organic solvent (generally ethanol, methanol, acetone, dimethylformamide) and the long extraction time are the main disadvantages of these methods (Li et al., 2005; Joana Gil-Chávez et al., 2013). Longer extraction times and using heat, increase the chance of oxidation, ionization and hydrolysis of phenolic compounds (Li et al., 2005; Naczk and Shahidi, 2004). In this context, various novel methods such as ultrasound-assisted extraction (Muniz-Marquez et al., 2013), microwave-assisted extraction (Wu et al., 2012), supercritical fluid extraction (Castro-Vargas et al., 2010), subcritical fluid extraction (Adil et al., 2007) and pressurized liquid extraction (Santos et al., 2012a) have been used for the extraction of phenolic compounds from plants. Extraction yield enhancement, reduction of solvent consumption, reduction of extraction cost, reduction of pollution to environment and reduction of extraction temperature are main advantages of UAE (Tao and Sun, 2015). MAE has advantages such as shortened extraction time and reduced solvent consumption (Bhuyan et al., 2015). SCE is the effective method with several advantages over conventional methods such as reduction the need for organic solvents, high selectivity, reduction of extraction time and easy separation of CO2 from the product (Lang and Wai, 2001). Many authors researched the extraction of phenolic compounds from hazelnut skin by conventional methods (Contini et al., 2008; Alasalvar et al., 2009; Locatelli et al., 2010). However there is limited research has been published about the use of novel extraction methods for the recovery of phenolic compounds from hazelnut skin. In addition to the best of our knowledge, no research has been done on the optimization and comparison of UAE, MAE and SCE process from hazelnut skin. Thus, the objectives of this study are to (1) evaluate the hazelnut skin by extracting phenolic compounds using novel methods (2) optimize UAE, MAE and SCE conditions for the hazelnut skin based on total phenolic content and antioxidant activity by response surface methodology; (3) compare the extracts obtained by novel methods with a conventional method (maceration) in terms of total phenolic content and antioxidant activity.

2. Materials and methods 2.1. Materials Hazelnut skins were obtained from the Fiskobirlik Integrated Hazelnut Processing Plant (Giresun, Turkey). The skins were filled in the polyethylene bags and delivered to the laboratory in a cooler. Hazelnut skins were ground using a blender (Waring Laboratory, Torrington, USA) and passed through a 1 mm sieve. Ground skins were defatted for 6 h using diethyl ether by Soxhlet apparatus. Defatted skins were dried in a vacuum oven (Nüve, Ankara, Turkey) to remove diethyl ether at 40 ◦ C for 1 h and stored in polyethylene bags at −18 ◦ C during the experiments. All chemicals and reagents

115

were analytical grade. Diethyl ether, Folin–Ciocalteau reagent, hydrochloric acid, glacial acetic acid, gallic acid, sodium carbonate, sodium acetate, 2,2-diphenyl-1-picrylhydrazyl (DPPH), Iron(II) sulfate heptahydrate and Iron(III) chloride hexahydrate (Sigma Aldrich Chemie GmbH, Steinheim, Germany); 2,4,6-tripyridyl-s-triazine (TPTZ) (Acros Organics, New Jersey, USA) and ethanol (Carlo Erba, Milan, Italy) were used in the experiments. 2.2. Methods 2.2.1. Ultrasound assisted extraction (UAE) An ultrasonic processor (VC 1500, Sonics and Materials Inc., Newtown, USA) with a 13 mm diameter probe was used for ultrasound assisted extraction. The probe was immersed to a depth of 25 mm in the sample. Samples were processed at a fixed power (1500 W), amplitude level (60%) and frequency (20 kHz). 1 g of defatted hazelnut skin sample was placed in a glass beaker and 50 mL of solvent was added. After extraction, the beaker was cooled to room temperature. The extract was filtered through Whatman paper No 1 and the solution was collected in a dark flask. 2.2.2. Microwave assisted extraction (MAE) Defatted hazelnut skin sample (3 g) was taken into a 500 mL round bottom flask and mixed with 150 mL solvent. A household microwave oven (MW71B, Samsung Electronics Ltd., Seoul, South Korea) with some modifications was used for microwave assisted extraction (Fig. 1). Frequency of microwave oven was constant and at 2450 MHz. After extraction, the flask was cooled to room temperature. The extract was filtered through Whatman paper No 1 and the solution was collected in a dark flask. 2.2.3. Supercritical carbon dioxide extraction (SCE) A supercritical fluid extractor (Spe-ed SFE-2, Applied Seperations Inc., Pennsylvania, USA) was used in this study. Extraction system has two pumps, main pump fitted with a chiller on the pump head for addition of liquid CO2 and a second pump (LabAlliance Model 1500, Scientific Systems Inc., Pennsylvania, USA) for the addition of co-solvent. 80% aqueous ethanol was used as cosolvent in this study. SC-CO2 and co-solvent was pumped at a fixed flow rate (2 L/min and 0.5 mL/min respectively) at the all extraction points. 2 g of defatted hazelnut skin sample was placed in the 10 mL of stainless steel extraction vessel. Extraction was started after reaching the desired pressure and temperature. After each extraction, the extract was collected in a glass vial. 2.2.4. Maceration (CSE) Defatted hazelnut skin (1 g) was extracted in a glass beaker with 50 mL of ethanol (50%, 70%, 90%) at a refrigerator (+4 ◦ C) for 24 h. The extract was filtered through Whatman paper No 1 and the solution was collected in a dark flask. 2.2.5. Experimental design and statistical analysis The three level Box-Behnken design with three factors was carried out to optimal levels of X1 (Temperature for UAE and SCE, Power for MAE), X2 (Extraction time for UAE, MAE and SCE) and X3 (Ethanol concentration for UAE and MAE, Pressure for SCE). Actual and coded values of the independent variables are shown in Table 1. Combinations of temperature (20, 40,60), extraction time (15, 30, 45 min), ethanol concentration (50, 70, 90%); power (300, 450, 600 W), extraction time (2, 4, 6 min), ethanol concentration (50, 70, 90%); temperature (40, 50, 60 ◦ C), pressure (10, 25, 40 bar), extraction time (20, 40, 60 min) were selected as independent variables for UAE, MAE and SCE, respectively. Total phenolic content, FRAP and DPPH radical scavenging activity (as 1/EC50 ) values of the extracts taken as the responses (Y ) for the design experiment. The 15 experimental points including three replicates at the central

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

116

Fig. 1. Schematic diagram of modified household microwave oven which is used in MAE experiments.

Table 1 Actual and coded values of independent variables. Coded values

Actual values UAE X1

X2

MAE X3

X1

X2

SCE X3

◦

−1 0 +1

X1

X2

X3

◦

20 C 15 min 50% 300 W 2 min 50% 40 C 20 min 10 bar 40 ◦ C 30 min 70% 450 W 4 min 70% 50 ◦ C 40 min 25 bar 60 ◦ C 45 min 90% 600 W 6 min 90% 60 ◦ C 60 min 40 bar

(Temperature: UAE and SCE, Power: MAE), X2 (Extraction time) and X3 (Ethanol concentration: UAE and MAE, Pressure: SCE).

point and total phenolic content, FRAP and 1/EC50 results are presented in Table 2. The response values in each trial were average of duplicates. The experimental data were fitted to the following quadratic polynomial model: Y = ˇ0 +

3  i=1

ˇi Xi +

3  i=1

ˇii Xi2 +

3 2  

ˇij Xi Xj

i=1 j=i+1

where Y is the predicted response; ˇ0 is the model intercept; ˇi , ˇii , ˇij are the linear, quadratic and interactive coefficients of the model respectively; Xi and Xj are the coded independent variables. The software Design Expert 9.0 (Trial version, Stat-Ease Inc., Mineapolis, USA) was used for the experimental plan, data analysis,

model generating, determination of optimum conditions and plot of response surfaces and contours. ANOVA was used to evaluate the statistical significance of independent variables and interactions amongst them. The adequacy of the models both obtained for UAE, MAE and SCE, were checked by evaluating coefficient of determination (R2 ), adjusted coefficient of determination (adj. R2 ), coefficient of variation (CV) and the Fisher’s test value (F-value). Models and regression coefficients were considered significant when p < 0.05. The relationship between independent variables and responses were analyzed by 3D response surface plots which represents the dependent variables in function of two independent variables when the third variable was kept constant at level 0 (in coded terms). Optimum conditions for UAE, MAE and SCE were calculated according to the desirability function. For this purpose, independent variables were kept in range while responses were maximized. 2.2.6. Analysis of the response variables 2.2.6.1. Measurement of the total phenolic content. The total phenolic content of the hazelnut skin extracts were determined with the Folin-Ciocalteu spectrophotometric method (Singleton and Rossi, 1965) with some modifications. 200 L diluted extract was mixed 1 mL of 0.2 N Folin-Ciocalteu reagent and the mixture was left to stand at dark for 8 min. Then, 2 mL of saturated sodium carbonate (75 g/L) was added, the mixture was shaken and kept in dark. After

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

117

Table 2 Coded Box-Behnken design with the analytical results and predicted values for UAE, MAE and SCE. Analytical results Coded

UAE

MAE

SCE

Run

X1

X2

X3

Y1

Y2

Y3

Y1

Y2

Y3

Y1

Y2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

−1 +1 −1 +1 −1 +1 −1 +1 0 0 0 0 0 0 0

−1 −1 +1 +1 0 0 0 0 −1 +1 −1 +1 0 0 0

0 0 0 0 −1 −1 +1 +1 −1 −1 +1 +1 0 0 0

96,63 96,63 127,88 121,31 88,19 71,63 49,13 48,19 65,69 68,50 40,38 59,13 116,94 110,06 118,81

438,63 478,27 712,79 569,10 517,90 451,84 256,96 166,13 311,46 481,57 212,37 412,21 475,04 443,58 466,71

4,00 4,01 4,66 4,39 4,05 4,11 3,78 3,82 3,90 3,78 3,75 4,00 4,05 3,96 3,95

66,13 83,31 78,63 98,94 52,22 109,25 42,06 67,22 84,72 95,81 55,03 68,47 91,91 92,84 93,16

285,87 442,57 361,43 541,66 267,30 555,90 237,57 334,18 456,81 516,27 274,73 345,95 502,64 503,88 506,36

1,52 1,88 1,84 2,22 1,43 2,51 1,12 1,70 1,91 2,07 1,14 1,79 1,96 1,87 2,00

0,12 0,26 65,41 41,42 45,02 39,7 14,08 0,29 15,48 66,12 0,18 9,07 19,38 18,98 19,62

0,98 1,49 403,38 204,17 256,81 245,45 165,07 2,44 138,6 426,09 1,51 157,18 170,19 163,6 167,71

Y3 0,00 0,01 1,51 1,43 1,44 1,28 0,17 0,00 0,79 2,18 0,00 0,81 0,87 0,83 0,84

Predicted values Coded

UAE

MAE

SCE

Run

X1

X2

X3

Y1

Y2

Y3

Y1

Y2

Y3

Y1

Y2

Y3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

−1 +1 −1 +1 −1 +1 −1 +1 0 0 0 0 0 0 0

−1 −1 +1 +1 0 0 0 0 −1 +1 −1 +1 0 0 0

0 0 0 0 −1 −1 +1 +1 −1 −1 +1 +1 0 0 0

102,92 102,92 122,29 122,29 73,50 73,50 49,21 49,21 63,82 83,19 39,52 58,89 112,61 112,61 112,61

420,15 420,15 603,88 603,88 440,69 440,69 261,92 261,92 348,83 532,56 170,05 353,78 512,02 512,02 512,02

4,07 4,07 4,37 4,37 3,99 3,99 3,99 3,99 3,67 3,96 3,67 3,96 4,05 4,05 4,05

60,80 90,72 73,97 103,89 57,82 103,68 46,45 60,44 83,67 96,83 56,36 69,53 91,85 91,85 91,85

279,45 459,98 355,78 536,32 285,95 562,48 230,99 315,53 435,76 512,09 284,79 361,13 504,29 504,29 504,29

1,41 2,01 1,78 2,38 1,68 2,28 1,14 1,74 1,80 2,16 1,25 1,62 1,90 1,90 1,90

8,30 0,00 49,79 39,05 46,88 36,14 11,21 0,47 10,33 72,70 0,00 16,15 23,68 23,68 23,68

32,60 39,29 394,52 201,49 268,34 250,80 158,79 0,00 95,59 423,56 0,00 172,46 166,98 166,98 166,98

0,09 0,09 1,37 1,37 1,32 1,32 0,14 0,14 0,71 2,28 0,00 0,81 0,90 0,90 0,90

X1 (Temperature: UAE and SCE, Power: MAE), X2 (Extraction time) and X3 (Ethanol concentration: UAE and MAE, Pressure: SCE). Y1 (Total phenolic compounds: mg GAE/g), Y2 (FRAP: mmol(Fe(II)/g), Y3 (1/EC50 : mL/mg).

2 h of reaction at room temperature, the absorbance of all samples was measured at 760 nm using a UV–vis spectrophotometer (Helios Gamma, Thermo Spectronic, Cambridge, UK). The results were expressed as mg gallic acid equivalents (GAE)/g defatted hazelnut skin.

2.2.6.2. Ferric-reducing antioxidant power (FRAP). The FRAP assay was carried out according to the procedure of Benzie and Szeto (1999) with some modifications. 50 L of diluted extracts were mixed with 0.95 mL of FRAP reagent (prepared by mixing 300 mM acetate buffer, pH 3.6, 10 mM TPTZ in 40 mM HCl, and 20 mM FeCl3 in proportions of 10:1:1). The absorbances of mixtures were measured at 593 nm after 5 min of incubation. FeSO4 was used as a standard and antioxidant power expressed as mmol Fe(II)/g defatted hazelnut skin.

2.2.6.3. DPPH radical scavenging activity. For the determination of DPPH free radical scavenging activity of extracts slightly modified method of Brand-Williams et al. (1995) was used. 50 ␮L of the diluted extract was added to 1 mL of the 100 ␮M DPPH solution prepared in absolute ethanol. The mixture was shaken and allowed to stand at room temperature in the dark for 30 min, and the absorbance was recorded at 515 nm. DPPH solution was used

as control. Reduction rate of DPPH was calculated with following equation. Reduction (%) =

Ac − As × 100 Ac

where Ac is the absorbance of control and As is the absorbance of extract. The sample amount (mg) necessary to decrease the initial DPPH concentration by 50% (EC50 ) was determined by an exponential curve. DPPH radical scavenging activity is defined as 1/EC50 (mL/mg). 3. Results and discussion 3.1. Ultrasound assisted extraction 3.1.1. Analysis of the model The responses (total phenolic content, FRAP and 1/EC50 values) and the coded values of independent variables (temperature, time and ethanol concentration) of each experiment are presented in Table 2. Total phenolic content of the extracts obtained from crude hazelnut skin with UAE ranged from 48.19 to 127.88 mg GAE/g. FRAP values ranged from 166.13 to 712.79 mmol Fe(II)/g and 1/EC50 values ranged from 3.75–4.66 mL/mg (Table 2). According to these experimental ranges, among the 15 extraction conditions (3 at center points), the maximum total phenolic compounds, FRAP and 1/EC50 values were obtained at run 3 (20 ◦ C, 45 min and 70% ethanol

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

118

Table 3 ANOVA results for the reduced quadratic models for UAE, MAE and SCE. Ultrasound assisted extraction Source

Model X1 X2 X3 X12 X13 X23 X11 X22 X33 Residual Lack of fit Total

Total phenolic compounds

FRAP

1/EC50

SS

F-value

p-value

SS

F-value

p-value

SS

F-value

p-value

11738.28 – 750.59 1180.49 – – – – – 9807.20 615.40 572.93 12353.68

69.94 – 13.42 21.10 – – – – – 175.30

<0.0001 – 0.0037 0.0008 – – – – – <0.0001

16.92 – 15.04 14.24 – – – – – 21.48

0.0002 – 0.0026 0.0031 – – – – – 0.0007

0.0097 – 0.0288 – – 0.0679 0.0190

0.2751

20.43

0.0475

0.51 – 0.17 – – – – 0.11 – 0.20 0.30 0.29 0.81

6.27 – 6.32 – – – – 4.10 – 7.55

3.00

227900.00 – 67517.10 63921.00 – – – – – 96426.00 49379.21 48847.84 277200.00

10.69

0.0884

Microwave assisted extraction Source

Model X1 X2 X3 X12 X13 X23 X11 X22 X33 Residual Lack of fit Total

Total phenolic compounds

FRAP

1/EC50

SS

F-value

p-value

SS

F-value

p-value

SS

F-value

p-value

5010.05 1790.41 346.63 1491.13 – 253.92 – 335.46 – 863.63 265.05 264.21 5275.10

25.20 54.04 10.46 45.01 – 7.66 – 10.13 – 26.07

<0.0001 <0.0001 0.0120 0.0002 – 0.0244 – 0.0130 – 0.0009

87.49 229.64 41.05 160.57 – 32.46 – 69.42 7.09 88.53

<0.0001 <0.0001 0.0004 <0.0001 – 0.0007 – <0.0001 0.0323 <0.0001

0.0002 0.0003 0.0086 0.0007 – – – – 0.0432 –

0.0095

110.37

0.0090

1.71 0.72 0.27 0.59 – – – – 0.13 – 0.25 0.25 1.97

16.94 28.58 10.63 23.22 – – – – 5.35 –

104.44

173900.00 65185.77 11653.30 45579.35 – 9215.04 – 19706.87 2013.85 25130.26 1987.07 1979.89 175800.00

6.82

0.1325

Supercritical carbon dioxide extraction Source

Model X1 X2 X3 X12 X13 X23 X11 X22 X33 Residual Lack of fit Total

Total phenolic compounds

FRAP

1/EC50

SS

F-value

p-value

SS

F-value

p-value

SS

F-value

p-value

6424.85 230.70 3443.67 2545.41 – – 435.77 – – – 546.58 546.37 7202.12

30.31 4.22 63.00 46.57 – – 7.97 – – –

<0.0001 0.0670 <0.0001 <0.0001 – – 0.0180 – – –

57.75 24.72 195.58 97.66 14.20 8.15 6.19 – – –

<0.0001 0.0011 <0.0001 <0.0001 0.0055 0.0213 0.0377 – – –

<0.0001 – <0.0001 <0.0001 – – 0.0312 0.0162 – –

0.0015

84.19

0.0118

6.26 – 3.29 2.77 – – 0.084 0.11 – – 0.13 0.13 6.39

116.74 – 245.46 206.91 – – 6.28 8.32 – –

653.34

243300.00 17362.23 137400.00 68588.82 9972.02 5720.65 4344.13 – – – 5618.33 5596.17 249000.00

38.41

0.0256

SS: Sum of squares.

concentration). While the lowest total phenolic compounds and 1/EC50 value were observed in the run 11 (40 ◦ C, 15 min and 90% ethanol concentration), minimum FRAP value was observed in the run 8 (60 ◦ C, 30 min and 90% ethanol concentration). Regression coefficients were determined with the least square method to develop second order quadratic polynomial models. Stepwise regression were used to identify the statistically significant (p < 0.05) terms and remove terms from the model equation. The following reduced second order models in terms of actual factors for total phenolic compounds, FRAP and 1/EC50 values of hazelnut skin as a function of temperature (X1 ), extraction ime (X2 ) and ethanol concentration (X3 ) were obtained:

Total phenolic compounds = −492.10 + 0.65X2 + 17.33X3 − 0.13X33

FRAP = −1327.59 + 6.12X2 + 51.78X3 − 0.40X33 1/EC50 = 3.86 + 9.75X2 + 1.23X11 − 52.75 The results of analysis of variance (ANOVA) were shown in Table 3. ANOVA, based on total phenolic compounds revealed that the model which had satisfactory levels of R2 (0.9502), adjusted R2 (0.9366) and CV (8.77%) were significant at p < 0.01. Because of the R2 and adjusted R2 values of the model which was developed for total phenolic compounds were higher than 0.80, there was a close agreement between experimental results and predicted values. When R2 is dramatically different from adjusted R2 , there can be non-significant terms in the model (Myers et al., 2009). The lesser CV value reveals a better reproducible model because CV is

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

119

the expression of standard deviation as a percentage of mean (Majd et al., 2014). Adequate precision is the measure of signal to noise ratio and greater ratios than 4 is desirable. Adequate precision ratio of 21.430 indicate that the model can be used to navigate the design space. Non-significant lack of fit indicate that the model was a good fit. The model which has been developed for FRAP is significant at p < 0.05 and has satisfactory R2 (0.8219), adjusted R2 (0.7733), CV (15.72%) and adequate precision (12.539) ratio. The results also showed that there was a significant lack of fit for the FRAP. Thus, appropriate transformation was performed, but was unsuccessful. This may occur when a very large experimental region is covered in the study (Capanzana and Buckle, 1997). Significant lack of fit discredits the model. However, if there were a large amount of data, a model with significant lack of fit could be used (Box and Draper, 2007). For the 1/EC50 , the generated model which had poor R2 value (0.6312) and adjusted R2 (0.5306) was significant (p < 0.01). Response surface methodology was not successful to develop an adequate model which describes DPPH radical scavenging activity (1/EC50 ) of the hazelnut skin extracts obtained by UAE. 3.1.2. Effect of temperature, time and ethanol concentration on responses As was presented in Table 3, the model showed that the most significant linear variable on total phenolic compounds was ethanol concentration with F value 21.10 (p < 0.01), followed by extraction time with F value 13.42 (p < 0.05). Meanwhile, quadratic term of ethanol concentration have significant effect on total phenolic compounds with F value 175.30 (p < 0.001). The most significant linear variable on FRAP value was extraction time with F value 15.04 (p < 0.01), followed by ethanol concentration with F value 14.24 (p < 0.01). The quadratic term of ethanol concentration have significant effect on FRAP value with F value 21.48 (p < 0.01). However, linear term of temperature and all of interaction terms had no significant effect on total phenolic compounds and FRAP value (p > 0.05). The linear term of extraction time and the quadratic term of ethanol concentration have significant effect on 1/EC50 value with F values 6.32 and 7.55, respectively (p < 0.05). The 3D response surface plots demonstrate the relationship between extraction parameters (extraction time and ethanol concentration) and responses (Fig. 2). From the figures, it can be seen that the total phenolic compounds, FRAP and 1/EC50 values increased with the increase of the ethanol concentration and about 65–70% reached a peak. After that value, ethanol concentration had a negative effect on total phenolic compounds and FRAP value. This effect may be attributed to the change of solvent polarity with the change in the ethanol concentration (Karacabey and Mazza, 2010). In addition, a certain concentration of ethanol can access plant cells, while high concentration of ethanol can cause protein denaturation, which prevents the dissolution of phenolic compounds and influences the extraction efficiency (Yang et al., 2009b). Adding a certain amount of water would increase the polarity of ethanolwater mixtures. Although 50% ethanol is the most polar solvent of all studied ethanol concentration levels, total phenolic compounds and FRAP value are lower than less polar ethanol concentration levels. This phenomenon might be attributed to the increasing viscosity with the increasing amount of water in the solvent (Sahin and Samli, 2013) and other factors such as weakening of solventmatrix interactions and swelling of plant material (Zuorro et al., 2014). Depending on the polarity of phenolic compounds, different proportions of ethanol-water mixtures from 30% (Wang et al., 2013) to 70% (Ghitescu et al., 2015) were found to be the optimal for UAE of phenolic compounds from different plants. As were seen in Fig. 2, when extraction was carried out for longer times, total phenolic compounds, FRAP and 1/EC50 values of the extracts obtained by UAE increased. Increasing extraction time provides longer contact of solids with the solvent and this may improve

Fig. 2. Response surface plots for the effect of independent variables of UAE on total phenolic compounds, FRAP and 1/EC50 values.

120

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

the diffusion of the phenolic compounds (Ghafoor et al., 2009). Elliptical contour plots indicate that there is an interaction between ethanol concentration and extraction time. The higher extraction time and medium ethanol concentration led to an increase in total phenolic compounds, FRAP and 1/EC50 values. Ethanol concentration was affected more on the both total phenolic compounds and antioxidant activity values (FRAP and 1/EC50 ) at higher extraction times. 3.1.3. Multi-response optimization of UAE process of hazelnut skin UAE process of phenolic compounds from hazelnut skin was optimized for determining maximum total phenolic compounds, FRAP and 1/EC50 values. Optimum extraction conditions were determined with desirability function. Optimum conditions of UAE of phenolic compounds from hazelnut skin by UAE were found as 45 min and 67.2–67.6% ethanol concentration. Extraction temperature has no significant effect on total phenolic compounds and antioxidant activity values (FRAP and 1/EC50 values). Under these optimum conditions predicted total phenolic compounds, FRAP and 1/EC50 values were 122.99–123.01 mg GAE/g, 612.20–613.25 mmol Fe(II)/g and 4.36 mL/mg, respectively. 3.2. Microwave assisted extraction 3.2.1. Analysis of the model The response data and the experimental design in terms of coded values are shown in Table 2. As shown in Table 2, total phenolic compounds of the extracts obtained from crude hazelnut skin with MAE ranged from 42.06 to 109.25 mg GAE/g. FRAP values ranged from 237.57 to 555.90 mmol Fe(II)/g and 1/EC50 values ranged from 1.12 to 2.22 mL/mg. According to these experimental ranges, among the 15 extraction conditions (3 at center points), the highest total phenolic compounds and FRAP and 1/EC50 value were obtained at run 6 (600 W, 4 min and 50% ethanol concentration), while the lowest values of all responses were obtained at run 7 (300 W, 4 min and 90% ethanol concentration). The following equations are reduced second order models in terms of actual factors which describe the relationship between responses and independent variables (microwave power (X1 ), time (X2 ) and ethanol concentration(X3 )). Total phenolic componds = −274.39 + 0.67X1 + 3.29 + 5.85X3 − 2.66x10−3 X13 − 4.22x10−3 X11 − 0.04X33

FRAP = −1844.17 + 4.64X1 + 65.79X2 + 32.30X3 − 0.02X13 − 3.25x10−3 X11 − 5.83X22 − 0.21X33

1/EC50 = −0.74 − 2.00x10−3 X1 + 0.09X2 + 0.05X3 − 4.75x10−4 X33

Fig. 3. Response surface plots for the effect of independent variables of MAE on total phenolic compounds, FRAP and 1/EC50 values.

ANOVA results for the quadratic polynomial models of MAE are presented in Table 3. The model for total phenolic compounds, FRAP and 1/EC50 were found to be significant according to the F values of 25.20 (p < 0.001), 87.49 (p < 0.001) and 16.86 (p < 0.001) respectively. The models have acceptable R2 (0.9498, 0.9887, 0.8709), adj.R2 (0.9121, 0.9774, 0.8192) and CV (7.32, 4.12, 8.87) values. Adequate precision ratios of 14.607, 26.942 and 13.527 indicate that the models can be used to navigate the design space. There was significant lack of fit values in case of total phenolic compounds and FRAP values, but the model which generated for 1/EC50 showed the insignificant lack of fit.

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

121

Table 4 Reference maceration results at different ethanol concentration. Ethanol concentration (%)

Total phenolic compounds (mg GAE/g)

FRAP (mmol Fe(II)/g)

1/EC50 (mL/mg)

50 70 90

58.82 ± 1.35a 44.55 ± 1.80b 36.99 ± 1.60c

394.55 ± 23.78a 364.27 ± 7.49b 331.24 ± 4.55c

1.59 ± 0.03c 1.52 ± 0.02b 1.03 ± 0.01a

Values are presented as means ± SD (n = 3). The different letters in the same column denote the significant differences (p < 0.05).

Table 5 Comparison of maximum values for maceration and optimum values for UAE, MAE and SCE. Extraction method

UAE MAE SCE Maceration

Comparison parameter Total phenolic compounds (mg GAE/g)

FRAP (mmol Fe(II)/g)

1/EC50 (mL/mg)

Extraction time (min)

122.91 111.55 72.80 58.82 ± 1.35

614.21 582.43 428.19 394.55 ± 23.78

4.39 2.48 2.26 1.59 ± 0.04

45.00 6.00 59.68 1440.00

3.2.2. Effect of power, time and ethanol concentration on responses All three independent variables (power, extraction time and ethanol concentration) were significantly effective on all responses (Table 3.). The most significant linear variable on total phenolic compounds, FRAP and 1/EC50 was power with F values 54.04 (p < 0.001), 229.64 (p < 0.001) and 28.34 (p < 0.01), respectively. Quadratic terms of power and ethanol concentration had significant effect on total phenolic compounds with F values 10.13 (p < 0.05) and 26.07 (p < 0.001), respectively. The interaction term of power*ethanol concentration had significant effect on total phenolic compounds and FRAP with F values 7.66 (p < 0.05) and 32.46 (p < 0.001), respectively. However, none of interaction terms had significant effect on 1/EC50 value. The quadratic terms of microwave power and ethanol concentration had significant effect on all responses. The 3D response surface plots for MAE were seen in Fig. 3. As was represented in Fig. 3 total phenolic compounds, FRAP and 1/EC50 value increased with the increase in the microwave power at a fixed ethanol concentration and temperature. Ballard et al. (2010) studied the microwave-assisted extraction of phenolic compounds from peanut skin. They showed that the quantity of phenolic compounds increased by increasing microwave power from 95 to 855 W. Microwave energy provides localized heating in the sample and it destroys the plant matrix so that target compound can diffuse out and dissolve in the solvent (Chan et al., 2011). With increase extraction times, total phenolic compounds, FRAP and 1/EC50 values were increased. The influence of power can be extended by the increasing extraction time (Proestos and Komaltis, 2008). Higher power with longer extraction time results in higher temperatures in the extraction vessel. Higher temperature and time could increase the solubility of phenolic compounds and decrease the viscosity of solvent (Dahmoune et al., 2015). Total phenolic compounds, FRAP and 1/EC50 values were increased with the increase in the ethanol concentration at fixed extraction time and microwave power. When the ethanol proportion of the solvent exceeded nearly 55%, total phenolic compounds, FRAP and 1/EC50 values reduced (Fig. 3A–C). Interaction between ethanol concentration and microwave power were seen in Fig. 3A–C. At high microwave power levels, increasing ethanol concentration results a decreasing total phenolic compounds, FRAP and 1/EC50 values. This effect can be attributed to the changes in polarity and dielectric properties of the solvent. Dielectric properties play an important role in microwave extraction because it depends heat distribution in sample (Dahmoune et al., 2015). Since

water has the higher dielectric constant than ethanol, the extraction mixture absorbs more microwave energy with increasing water proportions in solvent (Bouras et al., 2015). 3.2.3. Multi-response optimization of MAE process of hazelnut skin MAE process of phenolic compounds from hazelnut skin was optimized for determining maximum total phenolic compounds, FRAP and 1/EC50 values. Optimum extraction conditions were determined with desirability function. Optimum conditions of MAE of phenolic compounds from hazelnut skin were found as 600 W, 6 min and 55.03–56.23% ethanol concentration. Under these optimum conditions predicted total phenolic compounds, FRAP and 1/EC50 values were 111.53–111.55 mg GAE/g, 582.44–582.52 mmol Fe(II)/g and 2.48 mL/mg, respectively. 3.3. Supercritical fluid extraction 3.3.1. Analysis of the model Response surface methodology was employed to study the effects of pressure (100–400 bar), temperature (40–60 ◦ C), and extraction time (20–60 min) on total phenolic compounds and antioxidant activity (FRAP and 1/EC50 ) of hazelnut skin extracts obtained by SCE. The experimental data, ANOVA results and regression coefficients of the second order polynomial models for total phenolic compounds, FRAP and 1/EC50 values were showed in Tables 2, 3 and 4 respectively. As shown in Table 2; total phenolic compounds, FRAP and 1/EC50 values of the extracts obtained from hazelnut skin with SCE ranged from 0.12 to 66.12 mg GAE/g, 0.98–426 mmol Fe(II)/g and 0.00–2.18 mL/mg, respectively. Among these 15 extraction points the maximum total phenolic compounds, FRAP and 1/EC50 values were obtained at run 3 (50, 60 min and 10 bar), while the lowest values at run 4 (50 ◦ C, 60 min and 10 bar). The following reduced second order models as the functions of three independent variables with the linear, quadratic and interaction coefficients in terms of actual factors was derived by the least square method. Stepwise regression were used to identify the statistically significant (p < 0.05) terms and remove terms from the model equation. Total phenolic compounds = −22.88 + 1.91X2 + 0.20X3 − 0.03X23 FRAP = −632.12 + 11.63X1 + 21.78X2 + 10.83X3 − 0.25X12 − 0.25X13 − 0.11X23

122

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

Fig. 4. Response surface plots for the effect of independent variables of SCE on total phenolic compounds, FRAP and 1/EC50 values.

1/EC50 = 0.17 + 0.04X2 − 0.02X3 − 4.83x10−4 X23 − 5.77x10−5 X11 The results of ANOVA for SCE were shown in Table 3. The R2 (0.8921), Adj.R2 (0.8626) and the CV (35.51%) values indicate that the model for total phenolic compounds is adequate to describe the experimental results. The adequate precision value of 17.778

obtained from data indicated that the model can be used to navigate the design space. In addition, ANOVA results demonstrated that the model was significant at p < 0.001. ANOVA, based on FRAP value, revealed that the model which had satisfactory levels of R2 (0.9774), Adj.R2 (0.9605) and CV (15.87%) were significant at p < 0.001. Adequate precision ratio of 20.705 indicate that the model can be used to navigate the design space. The model which has been developed

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

123

for 1/EC50 is significant at p < 0.001 and has satisfactory R2 (0.9790), Adj.R2 (0.9706), CV (14.28%) and adequate precision (36.806) ratio. Lack of fit values of all models were significant at p < 0.05.

1/EC50 values was decreased with increasing ethanol concentration. Our findings are in agreement with Liyana-Pathirana and Shahidi (2005) and Yang et al. (2009a).

3.3.2. Effect of temperature, time and pressure on responses The analysis of the model showed that the most significant linear variable on total phenolic compounds was extraction time with F value 48.74 (p < 0.001), followed by pressure with F value 36.02 (p < 0.001) while temperature has no significant effect on total phenolic compounds (p > 0.05). Interaction term time*pressure had significant effect on total phenolic compounds with F value 6.17 (p < 0.05) while the interactions of other factors were no significant (p > 0.05). However, none of the quadratic terms had significant effect on total phenolic compounds (p > 0.05). The linear terms of temperature, extraction time and pressure were significantly effective (p < 0.001) on the FRAP values of extracts where time was the most significant one with F value 195.58. Interaction terms of temperature*time and temperature*pressure had significant effect on FRAP value with F values 8.15 and 6.19, respectively (p < 0.05). However, none of the quadratic terms had significant effect on FRAP value (p > 0.05). The most significant linear variable on 1/EC50 value was extraction time with F value 173.24 (p < 0.001), followed by ethanol concentration with F value 245.46 (p < 0.001). Interaction term of time*pressure and quadratic term of temerature had significant effect (p < 0.05) on 1/EC50 with F values 6.28 and 8.32, respectively. To illustrate the relationship between independent variables and responses 3D response surface plots was built. Fig. 4A, B, E show that total phenolic compounds, FRAP and 1/EC50 values of the hazelnut skin extracts decreased with increasing pressure. This result is in agreement with study of Gelmez et al. (2009), which supercritical CO2 was used for the extraction of phenolic compounds from roasted wheat germ. The solvation power of supercritical CO2 depends on its density, which increases with increasing pressure. The decreasing on total phenolic compounds with increasing pressure may be partially attributed to the low dispersion coefficient of the fluid. With increasing density and viscosity, the ability of solvent to penetrate the raw material reduced. Increasing extraction time has increased the total phenolic compounds, FRAP and 1/EC50 values specially at low pressure levels. Elliptical contour plots have shown that there were interactions between pressure and extraction time. Fig. 4C, D show that the effect of temperature*time and temperature*pressure interaction on FRAP values. It was clearly seen that increasing temperature at higher pressure levels and extraction times, decreased the FRAP values. Increasing temperature levels may lead to deterioration of antioxidative properties of extracts obtained by SCE (Ghafoor et al., 2012).

3.5. Comparison of UAE, MAE, SCE and maceration

3.3.3. Multi-response optimization of SCE process of hazelnut skin SCE process of phenolic compounds from hazelnut skin was optimized for determining maximum total phenolic compounds, FRAP and 1/EC50 values. Optimum conditions of SCE of phenolic compounds from hazelnut skin were found as 42.72–49.10 ◦ C, 59.83–60.00 min and 10.01–11.48 bar. Under these optimum conditions predicted total phenolic compounds, FRAP and 1/EC50 values were 69.59–72.64 mg GAE/g, 426.25–465.52 mmol Fe(II)/g and 2.18–2.27 mL/mg, respectively.

The results show that UAE, MAE and SCE methods (at optimum conditions) give better total phenolic compounds and antioxidant activity values than reference maceration (at maximum condition)(Table 5). The optimum UAE extract has approximately 9%, 68% and 108% higher total phenolic compounds comparing the optimum MAE, optimum SCE and maximum maceration extracts, respectively. Various researchers observed that total phenolic compounds and antioxidant activity of extracts obtained by UAE, MAE and SCE were more than conventional methods (Chemat et al., 2011; Dahmoune et al., 2015; Kazan et al., 2014; Khan et al., 2010; Majd et al., 2014; Proestos and Komaltis, 2008; Santos et al., 2012b). When comparing the novel extraction methods, UAE is more efficient than MAE and SCE, but MAE provides lesser extraction time than UAE and SCE. 4. Conclusions Response surface methodology was successful to develop an adequate model which describes total phenolic compounds and antioxidant activity values of hazelnut skin extracts obtained by UAE, MAE and SCE. Ethanol concentration, power and extraction time were demonstrated to be most significant parameters, affecting the total phenolic compounds and antioxidant activity of extracts obtained by UAE, MAE and SCE, respectively. Total phenolic compounds and antioxidant activity of the extract obtained by UAE at optimal conditions were higher than MAE and SCE at optimal conditions and reference maceration at maximum conditions. In addition, MAE reduced the extraction time significantly when compared to UAE, SCE and reference maceration. From the perspective of industrial applications; (1) Phenolic compounds from hazelnut skin, by-product of an industrial product, may evaluate in food and pharmaceutical industries as a good alternative to synthetic antioxidants such as Butylated hydroxyl anisole (BHA) and Butylated hydroxyl toluene (BHT), (2) UAE, MAE and SCE have been presented to be efficient methods for the extraction of phenolic compounds from hazelnut skin compared to the conventional method, (3) This study provides the knowledge about significance of parameters that may influence the effectivity of UAE, MAE and SCE processes of phenolic compounds from hazelnut skin, (4) It is expected that outcomes of this study can be used for the design of efficient UAE, MAE and SCE processes of phenolic compounds from hazelnut skin in commercial scale. Acknowledgements This work was supported by Ondokuz Mayıs University (Samsun, Turkey) through BAP Grant Number PYO.MUH.1904.11.012. We are grateful to the staff of Giresun Integrated Hazelnut Processing Plant for providing the hazelnut skin samples used in this study. References

3.4. Maceration The total phenolic compounds, FRAP and 1/EC50 results of reference maceration are showed in Table 4. Highest values of total phenolic compounds, FRAP and 1/EC50 were obtained at 50% ethanol concentration. Total phenolic compounds, FRAP and

Adil, I.H., C¸etın, H.I., Yener, M.E., Bayındırlı, A., 2007. Subcritical (carbon dioxide + ethanol) extraction of polyphenols from apple and peach pomaces, and determination of the antioxidant activities of the extracts. J. Supercrit. Fluids. 43, 55–63. Alasalvar, C., Karamac, M., Amarowicz, R., Shahidi, F., 2006. Antioxidant and antiradical activities in extracts of hazelnut kernel (Corylus Avellana L.) and hazelnut green leafy cover. J. Agric. Food Chem. 54, 4826–4832.

124

H.I˙ . Odabas¸, I. Koca / Industrial Crops and Products 91 (2016) 114–124

Alasalvar, C., Karamac, M., Kosinska, A., Rybarczyk, A., Shahidi, F., Amarowicz, R., 2009. Antioxidant activity of hazelnut skin phenolics. J. Agric. Food Chem. 57, 4645–4650. Altun, M., Celık, S.E., Guclu, K., Ozyurek, M., Ercag, E., Apak, R., 2013. Total antioxidant capacity and phenolic contents of turkish hazelnut (Corylus Avellana L.) kernels and oils. J. Food. Biochem. 37, 53–61. Aron, P.M., Kennedy, J.A., 2008. Flavan-3-ols nature, occurrence and biological activity. Mol. Nutr. Food Res. 52, 79–104. Bai, X.-L., Yue, T.-L., Yuan, Y.-H., Zhang, H.-W., 2010. Optimization of microwave-assisted extraction of polyphenols from apple pomace using response surface methodology and HPLC analysis. J. Sep. Sci. 33, 3751–3758. Ballard, T.S., Mallikarjunan, P., Zhou, K., O’Keefe, S., 2010. Microwave-assisted extraction of phenolic antioxidant compounds from peanut skins. Food Chem. 120, 1185–1192. Benzie, I.F.F., Szeto, Y.T., 1999. Total antioxidant capacity of teas by the ferric reducing/antioxidant power assay. J. Agric. Food Chem. 47, 633–636. Bhuyan, D.J., Quan Van, V., Chalmers, A.C., van Altena, I.A., Bowyer, M.C., Scarlett, C.J., 2015. Microwave-assisted extraction of Eucalyptus robusta leaf for the optimal yield of total phenolic compounds. Ind. Crops Prod. 69, 290–299. Bouras, M., Chadni, M., Barba, F.J., Grimi, N., Bals, O., Vorobiev, E., 2015. Optimization of microwave-assisted extraction of polyphenols from Quercus bark. Ind. Crops Prod. 77, 590–601. Box, G.E.P., Draper, N.R., 2007. Response Surfaces, Mixtures, and Ridge Analyses. Wiley-Interscience, Hoboken, N.J. Brand-Williams, W., Cuvelier, M.E., Berset, C., 1995. Use of a free radical method to evaluate antioxidant activity. Lwt-Food Sci. Technol. 28, 25–30. Capanzana, M.V., Buckle, K.A., 1997. Optimisation of germination conditions by response surface methodology of a high amylose rice (Oryza sativa) cultivar. Lwt-Food Sci. Technol. 30, 155–163. Castro-Vargas, H.I., Rodriguez-Varela, L.I., Ferreira, S.R.S., Parada-Alfonso, F., 2010. Extraction of phenolic fraction from guava seeds (Psidium guajava L.) using supercritical carbon dioxide and co-solvents. J. Supercrit. Fluids 51, 319–324. Chan, C.H., Yusoff, R., Ngoh, G.C., Kung, F.W.L., 2011. Microwave-assisted extractions of active ingredients from plants. J. Chromatogr. A 1218, 6213–6225. Chemat, F., Zill e, H., Khan, M.K., 2011. Applications of ultrasound in food technology: processing, preservation and extraction. Ultrason. Sonochem. 18, 813–835. Contini, M., Baccelloni, S., Frangipane, M.T., Merendino, N., Massantini, R., 2012. Increasing espresso coffee brew antioxidant capacity using phenolic extract recovered from hazelnut skin waste. J. Funct. Foods 4, 137–146. Contini, M., Baccelloni, S., Massantini, R., Anelli, G., 2008. Extraction of natural antioxidants from hazelnut (Corylus Avellana L.) shell and skin wastes by long maceration at room temperature. Food Chem. 110, 659–669. Contini, M., Baccelloni, S., Massantini, R., Anelli, G., Manzi, L., Merendıno, N., 2009. In vitro and in vivo antioxidant potential of phenolic extracts obtained from hazelnut skin by-products. In: Varvaro, L., Franco, S. (Eds.), VII International Congress On Hazelnut, 717–722. Dahmoune, F., Nayak, B., Moussi, K., Remini, H., Madani, K., 2015. Optimization of microwave-assisted extraction of polyphenols from Myrtus communis L. leaves. Food Chem. 166, 585–595. Del Rio, D., Calani, L., Dall’Asta, M., Brighenti, F., 2011. Polyphenolic composition of hazelnut skin. J. Agric. Food Chem. 59, 9935–9941. Dervisoglu, M., 2006. Influence of hazelnut flour and skin addition on the physical: chemical and sensory properties of vanilla ice cream. Int. J. Food Sci. Technol. 41, 657–661. Gelmez, N., Kıncal, N.S., Yener, M.E., 2009. Optimization of supercritical carbon dioxide extraction of antioxidants from roasted wheat germ based on yield, total phenolic and tocopherol contents, and antioxidant activities of the extracts. J. Supercrit. Fluids. 48, 217–224. Ghafoor, K., Al-Juhaimi, F.Y., Choi, Y.H., 2012. Supercritical fluid extraction of phenolic compounds and antioxidants from grape (Vitis labrusca B.) seeds. Plant Foods Hum. Nutr. 67, 407–414. Ghafoor, K., Choi, Y.H., Jeon, J.Y., Jo, I.H., 2009. Optimization of ultrasound-assisted extraction of phenolic compounds, antioxidants, and anthocyanins from grape (Vitis vinifera) seeds. J. Agric. Food Chem. 57, 4988–4994. Ghitescu, R.E., Volf, I., Carausu, C., Bühlmann, A.M., Gilca, I.A., Popa, V.I., 2015. Optimization of ultrasound-assisted extraction of polyphenols from spruce wood bark. Ultrason. Sonochem. 22, 535–541. Joana Gil-Chávez, G., Villa, J.A., Fernando Ayala-Zavala, J., Basilio Heredia, J., Sepulveda, D., Yahia, E.M., González-Aguilar, G.A., 2013. Technologies for extraction and production of bioactive compounds to be used as nutraceuticals and food ingredients: an overview. Compr. Rev. Food Sci. Food Saf. 12, 5–23. Karacabey, E., Mazza, G., 2010. Optimisation of antioxidant activity of grape cane extracts using response surface methodology. Food Chem. 119, 343–348. Kazan, A., Koyu, H., Turu, I.C., Yesil-Celiktas, O., 2014. Supercritical fluid extraction of Prunus persica leaves and utilization possibilities as a source of phenolic compounds. J. Supercrit. Fluids 92, 55–59.

Khan, M.K., Abert-Vian, M., Fabiano-Tixier, A.-S., Dangles, O., Chemat, F., 2010. Ultrasound-assisted extraction of polyphenols (flavanone glycosides) from orange (Citrus sinensis L.) peel. Food Chem. 119, 851–858. Lang, Q., Wai, C.M., 2001. Supercritical fluid extraction in herbal and natural product studies—a practical review. Talanta 53, 771–782. Li, H., Chen, B., Yao, S., 2005. Application of ultrasonic technique for extracting chlorogenic acid from Eucommia ulmodies Oliv. (E. ulmodies). Ultrason. Sonochem 12, 295–300. Liyana-Pathirana, C., Shahidi, F., 2005. Optimization of extraction of phenolic compounds from wheat using response surface methodology. Food Chem. 93, 47–56. Locatelli, M., Travaglia, F., Coısson, J.D., Martelli, A., Stevigny, C., Arlorio, M., 2010. Total antioxidant activity of hazelnut skin (Nocciola Piemonte Pgı): impact of different roasting conditions. Food Chem. 119, 1647–1655. Majd, M.H., Rajaei, A., Bashi, D.S., Mortazavi, S.A., Bolourian, S., 2014. Optimization of ultrasonic-assisted extraction of phenolic compounds from bovine pennyroyal (Phlomidoschema parviflorum) leaves using response surface methodology. Ind. Crops Prod. 57, 195–202. Montella, R., Coısson, J.D., Travaglıa, F., Locatellı, M., Malfa, P., Martellı, A., Arlorıo, M., 2013. Bioactive compounds from hazelnut skin (Corylus avellana L.): effects on Lactobacillus plantarum P17630 and Lactobacillus crispatus P17631. J. Funct. Foods 5, 306–315. Muniz-Marquez, D.B., Martinez-Avila, G.C., Wong-Paz, J.E., Belmares-Cerda, R., Rodriguez-Herrera, R., Aguilar, C.N., 2013. Ultrasound-assisted extraction of phenolic compounds from Laurus nobilis L. and their antioxidant activity. Ultrason. Sonochem. 20, 1149–1154. Myers, R.H., Montgomery, D.C., Anderson-Cook, C.M., 2009. Response Surface Methodology: Process and Product Optimization Using Designed Experiments. Wiley. Naczk, M., Shahidi, F., 2004. Extraction and analysis of phenolics in food. J. Chromatogr. A 1054, 95–111. Proestos, C., Komaltis, M., 2008. Application of microwave-assisted extraction to the fast extraction of plant phenolic compounds. Lwt-Food Sci. Technol. 41, 652–659. Sahin, S., Samli, R., 2013. Optimization of olive leaf extract obtained by ultrasound-assisted extraction with response surface methodology. Ultrason. Sonochem. 20, 595–602. Santos, D.T., Veggi, P.C., Meireles, M.A.A., 2012a. Optimization and economic evaluation of pressurized liquid extraction of phenolic compounds from jabuticaba skins. J. Food Eng. 108, 444–452. Santos, S.A.O., Villaverde, J.J., Silva, C.M., Neto, C.P., Silvestre, A.J.D., 2012b. Supercritical fluid extraction of phenolic compounds from Eucalyptus globulus Labill bark. J. Supercrit. Fluids 71, 71–79. Schmitzer, V., Slatnar, A., Veberic, R., Stampar, F., Solar, A., 2011. Roasting affects phenolic composition and antioxidative activity of hazelnuts (Corylus avellana L.). J. Food Sci. 76, S14–S19. Shahidi, F., Alasalvar, C., Liyana-Pathirana, C.M., 2007. Antioxidant phytochemicals in hazelnut kernel (Corylus avellana L.) and hazelnut byproducts. J. Agric. Food Chem 55, 1212–1220. Singleton, V.L., Rossi, J.A., 1965. Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. Am. J. Enol. Vitic. 16, 144–158. Stevigny, C., Rolle, L., Valentini, N., Zeppal, G., 2007. Optimization of extraction of phenolic content from hazelnut shell using response surface methodology. J. Sci. Food Agric. 87, 2817–2822. Tao, Y., Sun, D.W., 2015. Enhancement of food processes by ultrasound: a review. Crit. Rev. Food Sci. Nutr. 55, 570–594. Wang, J., Zhao, Y.-M., Tian, Y.-T., Yan, C.-L., Guo, C.-Y., 2013. Ultrasound-assisted extraction of total phenolic compounds from Inula helenium. Sci. World J. Wu, T., Yan, J., Liu, R., Marcone, M.F., Aisa, H.A., Tsao, R., 2012. Optimization of microwave-assisted extraction of phenolics from potato and its downstream waste using orthogonal array design. Food Chem. 133, 1292–1298. Yang, B., Liu, X., Gao, Y., 2009a. Extraction optimization of bioactive compounds (crocin, geniposide and total phenolic compounds) from Gardenia (Gardenia jasminoides Ellis) fruits with response surface methodology. Innov. Food Sci. Emerg. 10, 610–615. Yang, L., Jiang, J.-G., Li, W.-F., Chen, J., Wang, D.-Y., Zhu, L., 2009b. Optimum extraction process of polyphenols from the bark of Phyllanthus emblica L. based on the response surface methodology. J. Sep. Sci. 32, 1437–1444. Zuorro, A., Maffei, G., Lavecchia, R., 2014. Effect of solvent type and extraction conditions on the recovery of phenolic compounds from artichoke waste. Chem. Eng. Trans. 39, 463–468.