Microwave-assisted extraction of pectin from grape pomace

The utilization of microwave technique for the pectin extraction from grape pomace (Fetească Neagră and Rară Neagră), its influence on yield, galacturonic acid content, degree of esterification and molecular weight of pectin were analyzed. The optimal conditions of the extraction process were microwave power of 560 W, pH of 1.8 for 120 s. The pectin samples extracted by MAE in optimal conditions were analyzed by comparing with commercial apple and citrus pectin based on FT-IR analysis, thermal behavior, rheological characteristics and microstructure. The FT-IR analysis established the presence of different functional groups which are attributed to the finger print region of extracted pectin, while the rheological behavior presented a good viscoelasticity of pectin solutions. The obtained data assumes that grape pomace has a great potential to be a valuable source of pectin which can be extracted by simple and quick techniques, while maintaining analogous quality to conventional sources of pectin.

Rheological characterization of pectin solutions. In order to obtain 5% (w/w) solutions, pectin sample extracted by MAE in optimal conditions, commercial apple and citrus pectin were homogenized using Milli-Q water adjusted to pH 4 under constant stirring at 40 °C for 12 h. Then, the samples were cooled to room temperature (25 °C) and stocked under refrigeration at 4 °C for 16 h.
The dynamic viscosity of pectin solutions was determined with a Haake Mars 40 rheometer (Thermo Fisher Scientific, Waltham, MA, USA) using a cone (Ø 35 mm, 2°)-plate system at 20 °C. During measuring the dynamic viscosity (η, Pa·s) and shear stress (τ, Pa), the shear rate ( γ , s −1 ) was ranged between 0 and 100 s −1 . The stress sweeps of loss modulus (G′, Pa) and elastic modulus (G″, Pa) were determined at 1 Hz for measurement of the viscoelastic region. The frequency was presented a range from 0.1 to 100 Hz and the stress was selected within the linear viscoelastic region. (1)
The second significant factor which influences the extraction yield of pectin during MAE was the treatment time. It was established that pectin yield was increased when the extraction time was enhanced up to 500 s 38,39 . On the other hand, Thirugnanasambandham et al. 40 reported that the highest pectin yield extracted from dragon fruit peels was obtained at 400 W of microwave power, for 120 s at 45 °C and 24 g/mL of solid to liquid ratio. Also, for the extraction of pectin from sour orange peel, Hosseini et al. 41 obtained the following optimal conditions in order to obtain the maximum pectin yield, 700 W of microwave power, for 180 s at pH 1.5. These results correlated well with the experimental data obtained in our study for the highest yield of pectin from grape www.nature.com/scientificreports/   Table 3. In accord with the values in Table 2, the highest GalA content was realized at the correlation of the following parameter values, microwave power of 560 W, pH 2 for 120 s (81.24 g/100 g and 84.18 g/100 g for FN and RN pectin, respectively), while the lowest GalA content (50.92 g/100 g and 51.09 g/100 g for FN and RN pectin, respectively) at the correlation among microwave power of 280 W, pH 3 and 90 s. Figures 1D and 2D indicate that GalA content enhanced with the microwave power, when irradiation time increased. The similar results were obtained when a high microwave power was applied to extract pectin from sour orange peel 35 , lime peel 25 and lyophilized pumpkin flesh 42 . Therefore, it can be established that microwave heating may be an effective treatment in order to extract pectin from grape pomace without quality loss. Figures 1E and 2E present that a higher GalA content were obtained when the pH value was low (less than 2). Thus, the interactions, microwave power-pH and extraction time-pH demonstrated a statistically significant effect on the GalA content of pectin samples extracted by MAE. The same data was achieved by Hosseini et al. 41 , they obtained the following optimal conditions: microwave power of 700 W, irradiation time of 180 s and pH 1.5 for the extraction of pectin from sour orange peel. On the other hand, Lefsih et al. 43 reported that a range of pH from 1.5 to 3 didn't affect the GalA content of pectin from Opuntia ficus indica. Moreover, the decrease of pH values, the GalA content and yield of pectin were found to decrease 43 . In addition, increasing the irradiation time (Figs. 1F, 2F) enables an enhancement of the GalA content from grape pomace pectin. The enhanced GalA content can be explained by the improved penetration in the plant matrix of microwaves during the extraction 44 . Similar results were reported for the sweet lemon peel pectin extracted by microwave under the optimal conditions (microwave power of 700 W, pH 1.5 and irradiation time of 180 s) 44 .
Effect of extraction parameters on degree of esterification. The degree of esterification (DE) is a significant characteristic for the determination of the pectin applications in the food industry, which is related to its texturizing, emulsifying and gelling properties 29,45 . The experimental and predicted values of DE are shown in Table 2, while the ANOVA results for DE are presented in Table 3. The correlations among microwave power, irradiation time and pH which establish the evolution of DE of the grape pomace pectin are illustrated in Figs. 1G-I, 2G-I. According to the results showed in Table 2, all pectin samples had a DE higher than 50%, ranging from 62.28% (microwave power of 280 W, pH 3 for 90 s) to 82.29% (microwave power of 560 W, pH 2 for 120 s) for FN pectin and a range of 62.14-83.11% for RN pectin under similar extraction conditions. The obtained data for DE of the samples presented same tendency in the evolution of the yield and GalA content of pectin.
As can be seen from Figs. 1G-I and 2G-I, the DE of pectin enhanced gradually with the increase of microwave power. At higher microwave power, temperature of the pectin solutions enhanced to improve the diffusion of different compounds. Therefore, the combined influence of irradiation time and microwave power (Figs. 1G, 2G) was more significant than other combined variables presented in Figs. 1I and 2I, as such a higher microwave power needs a shorter extraction duration in order to obtain a high value of DE and vice versa 46 . Some researchers suggest to use the approach of low power with longer irradiation time for extraction as high microwave power presented might reduce purity of pectin 37,46,47 . On the other hand, researchers prefer to use the treatment of high power with short irradiation time for pectin extraction 13,35,44,48 .
Effect of extraction parameters on molecular weight. The molecular weight (M w ) of pectin is correlated with its gel-forming, thickening and stabilizing properties, which influence the utilization of pectin in the food industry 49 . Inadequate extraction conditions can affect the structure of pectin and thereby decrease its molecular weight 49,50 . For molecular weight of pectin from FN and RN pomaces, the 3D graphics are shown in Figs. 1J-L and 2J-L, respectively. The M w of pectin samples ranged from 4.15 × 10 4 g/mol (microwave power of 280 W, pH 3 for 90 s) to 4.57 × 10 4 g/mol (microwave power of 560 W, pH 2 for 120 s) and 4.21 × 10 4 g/mol (microwave power of 420 W, pH 3 for 120 s) to 4.60 × 10 4 g/mol (microwave power of 560 W, pH 2 for 120 s) for FN and RN pectin, respectively ( Table 2).
The microwave power and pH were significantly influenced the M w values of pectin solutions (Figs. 1K, 2K); this can be explained by that pectin underwent more hydrolysis The same data was obtained by Li et al. 51 , they reported the following optimal conditions, 152.63 W of microwave power, pH 1.57 for 3.53 min and 18.92 solid to liquid ratio for sugar beet pulp pectin. Also, they stated that pH of solution had the highest influence on average M w among the studied variables 51 . Moreover, Yoo et al. 42 noted that pectin extracted by MAE at pH 2 had more than 5 times higher M w than pectin extracted at pH 1. Microwave power implicates the conversion of high-frequency energy into heat energy which ensues in a more efficient pectin extraction 52 . On the other hand, Bagherian et al. 33 reported that M w decreased with the increasing of microwave power and heating time; they explained that continued heating may lead to pectin degradation (disaggregation of pectin matrix).
Optimization and validation of extraction conditions. The desirability function approach was utilized to optimize pectin characteristics (extraction yield, GalA content, DE and M w ) concurrently. The first characteristic ( y i ) was converted into desirability function ( d i ), which varied over the range presented in Eq. (12): Color. The pectin color is a significant factor affecting the appearance of gel produced and then the characteristic of the food product in which was added 53 . As can be seen in  (81.18). Also, the MAE treatment influenced significantly the chroma (C* ab ) and hue (h* ab ) of pectin samples. This can be explained by the fact that pectin extracted under higher power and temperature for prolonged time has a lower value of lightness (L*). Moreover, the color values of grape pomace pectin were predominantly due to tannins and anthocyanins which are the main polyphenolic compounds responsible for color in red grapes 54 . The similar tendency was reported for pectin extracted from lime peel 25 and apple pomace 55 .
FT-IR analysis. The different structural particularities of pectin extracted from grape pomace pectin (FN and RN) by applying microwave technique and correlate them to two commercial pectin samples (apple and citrus), FT-IR analysis was utilized. The FT-IR spectra of commercial samples (apple and citrus pectin) and pectin obtained by MAE in the optimal conditions are presented in Fig. 3. By comparing the spectra, FN and RN pectin had a peak around 3310 cm −1 which was ascribed to intermolecular bonding of O(6)H···O(3) 56 , while citrus and apple pectin had a shift at 3392-3366 cm −1 which was attributed to -OH and carbonyl C=O stretching vibrations 57 . The absorption peaks at 2929-2974 cm −1 which were found in pectin samples (commercial and extracted by MAE), was related to -CH (CH, CH 2 and CH 3 ) vibrations 58,59 . The C-H stretching detected at the peak of 2348 cm −1 is characteristic for polysaccharides chains 60 , while the peak of 1868 cm −1 was corresponded to the symmetric and asymmetric C-O stretching 61 . The band positions at 1714 cm −1 (FN and RN pectin) and 1733 cm −1 (commercial apple and citrus pectin) were assigned to undissociated carboxylic acid (COOH) and -CO from the group methyl ester (COOCH 3 ) 62 . The band identified at 1606 cm −1 was due to the asymmetric stretching vibration of the carboxylate ion (COO-) and C = C ring stretching of phenolic compounds 63 , while the peak at 1559 cm −1 was ascribed to the valence vibration of C = O bond 64 .
The absorption bands at 1399 cm −1 and 1261 cm −1 indicated enhancement quantity of carboxylates 65 and -CH bending 66 , respectively. The FT-IR spectra in the wavenumber between 1300 and 800 cm −1 are referred to as the 'fingerprint region' of carbohydrates, which enables the identification of major chemical groups in different polysaccharides 29,67 . Therefore, the peaks at 1226 cm −1 , 1223 cm −1 and 1208 cm −1 were assigned to the stretching vibration of C single bond, C = O stretching and C-O-H bending, respectively 25 . The peaks at around 1143 cm −1 and 1131 cm −1 were attributed to the C-O, C-O-C and C-C rings, which are characteristic to the structure of commercial (apple and citrus pectin) samples and pectin (FN and RN) extracted by MAE, respectively 68 . The bands identified at 1072 cm −1 (commercial apple and citrus pectin) and 1065 cm −1 (FN and RN pectin) were ascribed to C-O and C-C stretching of xyloglucan 69 and galactoglucomannan 68 , respectively. Some peaks are more intense in isolated pectin, such as at 1012 cm −1 (C-O, C-C and C-O-H stretching) 69 , 842 cm −1 (CH 2 bending linked to α-arabinose pyranoid ring) 68 , 828 cm −1 and 830 cm −1 (α-D-mannopyranose) 70 . Thermal analysis. The DSC was employed to examine the thermal characteristics and describe changes during thermal denaturation of pectin samples (FN and RN pectin extracted by MAE, commercial apple and citrus pectin) as illustrated in Fig. 4 and Table 4. Generally, thermal properties of pectin depend on the inter- Table 4. Color characteristics, thermal properties, creep and recovery parameters of pectin samples. Mean values and standard deviation, in brackets. ns, p > 0.05, *p < 0.01, **p < 0.001, ***p < 0.0001, a-d different letters in the same column indicate significant differences among samples (p < 0.0001) according to the LSD test with α = 0.05. FNP, Fetească Neagră pectin; RNP, Rară Neagră pectin; CAP, commercial apple pectin; CCP, commercial citrus pectin; L*, lightness of the color; C* ab , chroma; h* ab , hue angle; ΔH d , degradation enthalpy; T d , degradation temperature; J e , equilibrium compliance; J r , recoverable compliance; γ , shear rate; η, viscosity. www.nature.com/scientificreports/ dependence of three factors, its state transition occurred during processing, chemical composition and stability properties 71,72 . In the thermogram of pectin samples, exothermic peaks (degradation temperature) were registered at temperatures between 217.88 and 276.73 °C, while endothermic peaks (melting temperature) were not noticed. The endothermic peaks appear from water evaporation, melting-recrystallization of the crystallites, hydrogen bonding of GalA units, and also a conformational change of the pectin pyranose rings [72][73][74][75][76] . In this study, the absence of endothermic peaks indicates that no bound water was removed from the pectin samples 18 .   Fig. 5; all curves show a non-Newtonian fluid behavior with an enhancement in the shear stress and a decrease of the dynamic viscosity. This behavior was assigned to the decrease of the pectin intermolecular forces during the stress application 25,77 . It was noticed that RN pectin extracted by MAE had a higher dynamic viscosity than other samples, which means that source of pectin and different extraction parameters affect the pectin flow behavior 25,29 . The viscosity of RN pectin extracted by MAE at a shear rate of 100 s −1 was 2.53 Pa·s which was higher than viscosity of other samples, commercial citrus pectin (1.61 Pa·s), commercial apple pectin (1.11 Pa·s) and FN pectin extracted by MAE (1.01 Pa·s). Moreover, this value was also higher than the dynamic viscosity obtained at the same shear rate (100 s −1 ) for different concentration (0.5%, 1%, 2% and 3%) of lime peel pectin solution (less than 1 Pa·s) 25 , 1.5% and 2% pectin solutions of sour orange peel (less than 0.01 Pa·s) 44 and 30 g/L pectin solution of finger citron pomace (0.5 Pa·s) 70 .
Viscoelastic properties of pectin solutions. Figure 6 shows the viscoelastic characteristics of the 5% pectin solutions; the elastic modulus (G′) and loss modulus (G″) were determined in the linear viscoelastic region. The elastic modulus (G′) serves as the elastic constituent of the stress, while the loss modulus (G″) determines the energy lost via viscous flow 82 . All pectin samples had a higher G″ (liquid behavior) than G′ (solid behavior) in the 0.1-100 Hz frequency domain applied (Fig. 6). The values of both moduli enhanced proportionally with the frequency. Therefore, the G′ enhances more sharply with frequency correlated to the behavior of G″, until the two curves intersect and elastic constituents override the viscous. For that reason, the ability of pectin network  www.nature.com/scientificreports/ to keep the temporarily enforced energy enhances, and it involves more like an elastic solid 81 . Furthermore, the intersection of moduli (G′ and G″) shows the good viscoelasticity of pectin solutions 83 ; the lower the value of intersection moment, the major the role of elasticity 84 . The similar behavior was also noticed for the 5% pectin solutions from cacao pod husks 81 , lime peel waste 25 , pulp of gabiroba 85 and apple pomace 79 . In addition, the extraction methods had a considerable impact on the rheological parameters of pectin and samples extracted by microwave treatment could be considered adequate for using in diverse food products.
'Creep and recovery̕ analysis. In the 'creep and recovery̕ analysis (Fig. 7), the pectin samples were subjected to a constant stress during 360 s in order to assess the material deformation; the 'creep̕ test is from 0 to 180 s, while the 'recovery̕ test is from 180 to 360 s. The 'recovery̕ analysis allows access to the rheological behavior of material and lower shear rates for various systems 86 . The 'creep and recovery̕ analysis, through the creep compliance ( J ) as time function, distinguishes between the viscous and elastic phases. In viscoelastic materials, 'recovery̕ phase of the applied stress is partial, controlled by viscous or elastic characteristic of the samples, established at a transitional point between liquid and solid 86 . All samples (FN and RN pectin extracted by MAE; commercial apple and citrus pectin) manifested a non-Newtonian behavior, with a decrease of strain response during the applied stress, evidencing their viscoelastic characters. Commercial citrus pectin solution had a better recovery than other pectin samples. The creep and recovery parameters are presented in Table 4; the equilibrium and recoverable compliance ( J e and J r ), respectively) values for commercial citrus pectin and RN pectin were lower than commercial apple pectin and FN pectin. The highest shear rate ( γ ) was observed for FN pectin (0.1252 1/s), while the lowest was obtained for CCP (0.0004 1/s); the similar tendency was noticed for d(log(γ)/d(log(t)).

Microstructure.
The structural morphology of the pectin extracted by MAE from grape pomace (FN and RN) and commercial pectin samples (apple and citrus) was analysed by scanning electron microscopy (SEM). As Fig. 8 shows, the FN and RN pectin samples obtained by MAE appeared to be very different from commercial apple and citrus pectin, having a coarse and slightly ruptured surface. The structure appeared to be influenced by the accelerated enhancement of temperature and the high internal pressure associated with MAE method 46,87 ; similar results in terms of morphology were obtained by Liew et al. 46 . Moreover, microwave irradiation causes a great disintegration in the structural morphology of the raw material, which generates an increase of pectin yield 55,87 . The structure of commercial citrus pectin and grape pomace pectin (FN and RN) was found to have a large number of irregular particles with a rough surface (Fig. 8). This may due to the fact that citrus and grape pomace pectin are rich in insoluble fibres (e.g., lignin, cellulose and hemicellulose); similar structural morphologies have been noted in dried pomace from different vegetables and fruits 25,88 . There are some precise differences in the structure of the samples; the commercial apple pectin showed a slight tendency to curl, while the citrus and grape pomace pectin samples seemed to be ruptured. It was also noticed a more homogenous distribution of particle sizes, an increasing number and size of cavities in the structure of the MAE samples (FN and RN pectin), which could be due to pressure rise and sharp intracellular temperature 55 . From the obtained results, it can be concluded that MAE technique influenced the surface morphology of pectin samples extracted from grape pomace.

Conclusions
Pectin was extracted from grape pomace pectin (FN and RN) by MAE using three independent variables, each at three levels, microwave power (280, 420 and 560 W), irradiation time (60, 90 and 120 s) and pH (1, 2 and 3). The microwave power applied for pectin extraction and pH of solution were found to have a high impact on all four responses (extraction yield, GalA content, DE and M w ), while irradiation time had a great influence on pectin yield, GalA content and DE. The optimal conditions for pectin extraction were 560 W, pH of 1.8 for 120 s. The pectin sample extracted by MAE under optimal conditions were compared to CAP and CCP by FT-IR, rheological analysis, DSC and SEM. The viscosity of RN pectin extracted by MAE had a higher viscosity than viscosity of other samples. The microstructure of the pectin samples appeared to be very different from commercial apple and citrus pectin. The grape pomace was found to be a relevant and unconventional source of pectin with a high GalA content, DE and M w . The physicochemical properties, morphological characteristics and rheological behavior of pectin extracted by MAE from grape pomace denoted a promising field of different applications of this fiber in food industry.