Supercritical carbon dioxide and pressurized liquid extraction of valuable ingredients from Viburnum opulus pomace and berries and evaluation of product characteristics

Supercritical carbon dioxide and pressurized liquid extraction of valuable ingredients from Viburnum opulus pomace and berries and evaluation of product characteristics

J. of Supercritical Fluids 122 (2017) 99–108 Contents lists available at ScienceDirect The Journal of Supercritical Fluids journal homepage: www.els...

2MB Sizes 0 Downloads 48 Views

J. of Supercritical Fluids 122 (2017) 99–108

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids journal homepage: www.elsevier.com/locate/supflu

Supercritical carbon dioxide and pressurized liquid extraction of valuable ingredients from Viburnum opulus pomace and berries and evaluation of product characteristics ˙ Rita Kazernaviˇciut ¯ e, ˙ Petras Rimantas Venskutonis ∗ Paulius Kraujalis, Vaida Kraujaliene, Department of Food Science and Technology, Kaunas University of Technology, Radvilen ˙ u˛ pl. 19, LT-50254, Lithuania

a r t i c l e

i n f o

Article history: Received 11 August 2016 Received in revised form 15 December 2016 Accepted 16 December 2016 Available online 21 December 2016 Keywords: V. opulus berries Pomace Supercritical carbon dioxide extraction Pressurized liquid extraction Response surface methodology Antioxidant capacity

a b s t r a c t Supercritical carbon dioxide extraction (SFE-CO2 ) of Viburnum opulus L. fruits and pomace was optimized using 2 level factorial and central composite design (CCD). The effects and interactions of temperature (T), pressure (P), extraction time (t) and CO2 flow (v) were estimated for the washed V. opulus berry pomace, while 2 most significant factors, P and t were used further to estimate the coefficients of quadratic model and to find optimal extraction parameters by response surface methodology (RSM) for the unwashed pomace. The highest extract yields at optimal parameters (P = 55–57 MPa, t = 120–131 min, T = 50 ◦ C and v = 2.5 L min−1 ) from washed, unwashed berry pomace and dried whole berries were 19.1, 14.6 and 6.6%, respectively. The oil in lipophilic fractions was composed mainly of oleic (42–51%) and linoleic (42–46%) fatty acids; it contained 963–1157 mg kg−1 tocopherols. SFE-CO2 residue was consecutively reextracted by pressurized acetone, water and ethanol yielding 11.59 ± 0.95, 27.58 ± 0.79 and 30.71 ± 0.59% of extracts, respectively, which demonstrated strong antioxidant capacity in DPPH and ABTS radical scavenging, oxygen radical absorbance capacity (ORAC) and total phenolics assays. In total, biorefining of unwashed pomace yielded 62.11% of extracts from the initial plant material. The fractions obtained may be considered as valuable functional ingredients for foods, nutraceuticals and other applications. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The growing interest in natural food ingredients has been an important factor in expanding the studies of less known horticultural plants in recent years. Nevertheless, some berry species remain poorly studied, including the genus of Viburnum (>230 species), which is grown for ornamental purposes and edible fruits (common names of V. opulus: guelder-rose, water elder, cramp bark, snowball tree and European cranberrybush). Evaluation of antioxidant potential and phytochemicals of several Viburnum genotypes demonstrated that berry juices possess strong radical scavenging capacity [1,2], due to high concentrations of phenolic constituents such as chlorogenic and quinic acids [3,4], and antimicrobial activity [1,5]. V. opulus fruits have been used as ingredients in sauces, jellies, marmalades and drinks [6,7]. However, the main problem in direct use of V. opulus berries for foods is associated with undesirable flavor notes, due to the presence of some unpleasantly smelling constituents [8]. Therefore, there is an interest in develop-

∗ Corresponding author. E-mail address: [email protected] (P.R. Venskutonis). http://dx.doi.org/10.1016/j.supflu.2016.12.008 0896-8446/© 2016 Elsevier B.V. All rights reserved.

ing extraction methods for the isolation of valuable fractions from the whole V. opulus berries and their pomace. Biorefining of berry pomace, which are often discarded as a waste or used inefficiently, is of particular importance because large amounts of valuable constituents are lost. Antioxidant properties of V. opulus were evaluated by DPPH• and • ABTS + scavenging, total phenolics (TPC), ferric reducing antioxidant power (FRAP) and oxygen radical absorbance capacity (ORAC) assays [3]. Coumaroyl-quinic acid, chlorogenic acid, procyanidin B2, and procyanidin trimer were the strongest antioxidants in juice, while the composition of antioxidants in the extracts of seeds obtained with organic solvents was different [4]. Until now research has been focused mainly on the waste of the major horticultural products; e.g. supercritical fluid extraction with carbon dioxide (SFE-CO2 ) was applied for citrus fruits [9] hazelnut, coffee and grape wastes [10,11], olive waste [12], raspberry pomace [13], apple peels [14] and sour cherry [15], while numerous other species remain under explored. It should also be emphasized that green technologies in biorefining agromaterials are highly preferable, particularly when the products are intended for nutrition. Green alternative methods for the isolation of antioxidant bioactive compounds from winery wastes and by-products, which also

100

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

include SFE-CO2 , were recently reviewed [16]. SFE-CO2 possesses many advantages (nontoxic, nonflammable, inexpensive and yields high purity extracts) and therefore can be successfully explored in food industry. Pure CO2 is particularly effective in extraction of lipophilic constituents; for instance, oil from sea buckthorn [15,17] and pomegranate [16,18]. However for more effective isolation of polyphenolic fractions polar co-solvents are required as it was demonstrated for guava seeds [17,19] and black chokeberries [20]. The solubility of various substances in CO2 highly depends on their properties and process parameters; therefore, SFE-CO2 should be optimized for every individual raw material. All previously performed studies used conventional extraction techniques for V. opulus and only one study applied SFE-CO2 for V. opulus seeds [21]; however the yield of oil was not indicated and extraction parameters were not optimized. A statistical tool, called Response Surface Methodology (RSM) is used for SFE-CO2 optimization in order to evaluate the effect of multiple factors and their interactions on dependent responses; whereas Central Composite Design (CCD) is the most popular form of RSM. The main aim of this study was comprehensive evaluation of SFE-CO2 process for obtaining lipophilic fractions from V. opulus berries and their pomaces. The objectives were to determine the optimal conditions of SFE-CO2 and to evaluate the extracts obtained as possible functional ingredients for food and other applications. In addition, SFE-CO2 residues were consecutively re-extracted using pressurized liquid extraction (PLE) with increasing polarity solvents, namely acetone, ethanol and water for obtaining higher polarity polyphenolic fractions.

2. Materials and methods 2.1. Materials The following V. opulus samples obtained from the SIA BestBerry (Latvia) were used in this study: (1) whole dried berries (moisture content 8.4%), (2) dried berry pomace (obtained after pressing juice, moisture content 9.3%), and (3) washed berry pomace (obtained after washing berry pomace with water to remove soluble juice residues and drying at 45–50 ◦ C for 6 h; moisture content 9.5%). Fatty acid methyl esters (FAME) and tocopherols, DL␣-T (99.9%), rac-␤-T (90 +%), ␥-T (99%) and ␦-T (95.5%), were from Supelco Analytical (Bellefonte, PA, USA); 2,2 azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), 6hydroxy-2,5,7,8-tetra-methyl-chroman-2-carboxylic acid (Trolox), microcrystalline cellulose (20 ␮m), fluorescein (FL) sodium salt, [2,2-azobis(2-methylpropionamidine) dihydro-chloride] (AAPH), Folin–Ciocalteau’s reagent and HPLC grade solvents were from Sigma-Aldrich Chemie (Steinheim, Germany). Ascorbic acid was from Across Organics (Geel, Belgium), KOH from Lachema (Neratovice, Czech Republic), n-hexane from Reachem Slovakia (Bratislava, Slovakia). Randomly methylated ␤-cyclodextrin (RMCD, Trappsol) was from CTD Holdings (High Springs, FL, USA), CO2 (99.9%) and helium from AGA (Vilnius, Lithuania).

2.2. Milling and particle size determination The samples were ground in an ultra-centrifugal mill ZM 200 (Retsch, Haan, Germany) using 500 ␮m hole size sieve. Dried berries before grinding were frozen in liquid nitrogen. Particle size distribution of ground berries was measured on a Mastersizer, Hydro 2000S (A) analyzer (UK) operating on a laser diffraction method [22]. Water was used as a dispersant for wet analysis, dispersant refractive index was 1.33, and particle refractive index was 1.53. Volumetric mass of ground materials was

as follows (in g cm−3 ): berries 0.479 ± 0.003, unwashed pomace 0.442 ± 0.003, washed pomace 0.431 ± 0.004. 2.3. Supercritical carbon dioxide extraction (SFE-CO2 ) SFE-CO2 was performed in a Helix extractor (Applied Separation, Allentown, PA) from 20 g of ground sample placed in a 50 mL cylindrical vessel (320 × 14 mm) between two layers of cotton wool (Fig. 1). The temperature of the extraction vessel was controlled by a surrounding heating jacket. The flow rate of CO2 in the system (v) was controlled manually by the micro-metering valve (backpressure regulator). The volume of CO2 consumed was measured by a ball float rotameter and a digital mass flow meter in liters per min (L min−1 ) at standard state: pressure (P) = 100 kPa, temperature (T) = 20 ◦ C, density () = 0.0018 g mL−1 . The conditions for extraction were set as shown in Table 1. Optimized T and P conditions were further used for the up-scaling SFE-CO2 using 10 L extraction vessel in a pilot extractor (Applied Separation, Allentown, PA). 2.4. Pressurized liquid extraction (PLE) PLE was performed on an accelerated solvent extraction apparatus ASE 350 (Dionex, Sunnyvale, CA, USA) in 34 mL extraction cells applying two cellulose and metal filters at the top and at the bottom. Unwashed V. opulus berry pomace residues (15 g) obtained after SFE-CO2 at optimal parameters was consecutively re-extracted by PLE with acetone, ethanol and water at the following parameters: 70 ◦ C, 10.3 MPa, 5 min pre-heating, 15 min static extraction (3 cycles × 5 min), 100% of cell flush volume and 100 s purge time with N2 for acetone and ethanol, and 120 ◦ C with 7 min pre-heating for water extraction. Dried material before extraction with water was mixed with diatomaceous earth (1:1). These extraction parameters were selected from our previous extraction experiments with raspberry pomace [13]. The extracts obtained by PLE were collected in the separate vials, acetone and ethanol were evaporated in a Büchi V–850 Rotavapor R–210 (Büchi Labortechnik AG, Flawil, Switzerland), while water residues were freeze-dried. Dry extracts were stored at −18 ◦ C. Soxhlet extraction with hexane was used as a reference method [23] using 20 g of material. The solvent was removed in a rotary vacuum evaporator at 40 ◦ C. 2.5. Determination of tocopherols Oily extracts were saponified by placing 0.1 g extract and 0.05 g antioxidant ascorbic acid in a screw-capped tube with 5 mL ethanol (95%) and 0.5 mL KOH (60%). After vortexing 60 s the tubes were flushed with nitrogen 60 s, capped and mixed for 2 h at room temperature. Afterwards 3 mL deionized water and 5 mL hexane were added and again vortexed 60 s. The suspension was extracted four times with 5 mL hexane. The organic layer was collected and evaporated to dryness with nitrogen; the residue was diluted with HPLC mobile phase to a final concentration of 1.6%. Tocopherols were determined by HPLC [24] using Shimadzu HPLC system with solvent delivery unit (LC-20A), system controller (CBM-20A), auto-sampler (SIL-20A), column oven (CTO20A) and fluorescence detector. A reverse-phase C30 column (particle size 5 ␮m, 250 × 4.6 mm) was termostated at 30 ◦ C applying isocratic elution with acetonitrile:methanol:dichlormethane (72/22/6, v/v/v). Injection volume was 20 ␮L, flow rate 1 mL min−1 . Tocopherols were detected using fluorescence detector at 290 nm excitation and 330 nm emission and identified by comparing their retention times with reference solutions prepared at different concentrations (0–10 ␮g mL−1 ) using mobile phase. The calibration

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

101

Fig. 1. Schematic diagram of 10 L pilot scale SC CO2 extractor. Table 1 Levels of independent variables in optimization of SFE-CO2 of V. opulus berry pomace. Analytical factors

Coded symbols

Levels −1.4

−1

0a

+1

+1.4

1st series (washed berry pomace) Extraction pressure, P (MPa) Extraction time, t (min) Extraction temperature T (◦ C) CO2 flow rate, v (L min−1 )

X1 X2 X3 X4

– – – –

35 60 30 1.5

45 90 40 2

55 120 50 2.5

– – – –

2nd series (unwashed berry pomace) Extraction pressure, P (MPa) Extraction time, t (min)

X1 X2

41 78

45 90

55 120

65 150

69 162

a

center points of experimental design.

curves were used to determine the concentration of tocopherols in the samples. Analyses were performed in triplicate. 2.6. Gas chromatographic (GC) determination of fatty acid profile Fatty acid methyl esters (FAMEs) were prepared by using BF3 catalyst [23] and analyzed on a HRGC 5300 with a flame ionization detector and 100 m length 0.25 mm (id), 0.20 ␮m film thickness fused silica capillary column SPTM -2560 (Supelco, Bellafonte, PA, USA). Analysis parameters were as follows: injection temperature 220 ◦ C; detector’s temperature 240 ◦ C; split ratio 100:1; oven temperature was programmed from 80 to 135 ◦ C at 4 ◦ C min−1 , from 135 to 185 ◦ C at 4 ◦ C min−1 , and from 185 to 240 ◦ C at 4 ◦ C min−1 . Carrier gas was helium at a flow rate of 20 cm3 s−1 . The compounds were identified by comparing their retention times with those of a commercial FAME mixture. Relative standard deviations (RSD) for fatty acids from 3 replicate runs were in the range of 1–3%. 2.7. Assessment of extract antioxidants properties •

The ABTS + scavenging capacity assay of PLE extracts, which is based on the inhibition of radical by the antioxidants was performed by the method of Re et al. [25]. Briefly, radical solution was • produced by mixing ABTS + with PBS at a ratio of 2:1 and kept in the dark during 12–16 h (stable for 2 days). For the assessment of • radical scavenging capacity ABTS + was diluted with PBS, to keep absorption 0.8 ± 0.02 at 734 nm 300 ␮L of extract solution in 50% ethanol (1% w/v) were pipetted into cuvettes and the final absorption was measured after 20 min. PBS was used as a blank. Trolox

solutions were used for calibration in the concentration range of 150–1000 ␮M. Radical scavenging capacity was calculated by the formulae: % inhibition = [(AB –AE )/AB ] × 100, where AB − absorption of blank samples; AE −absorption of the sample with pomace • extract. The ABTS + scavenging capacity was expressed in trolox equivalent antioxidant capacity (TEAC) in ␮mol TE g−1 . DPPH• scavenging capacity of extracts was determined by a slightly modified spectrophotometric method [26] using a 96 well microplate reader. For each well an aliquot of 7.5 ␮L of extract was mixed with 300 ␮L of DPPH• (6 × 10−5 mol L−1 ) and the decrease of absorbance was measured during 40 min at 515 nm by comparing with a blank sample containing the same amount of methanol and DPPH• . The final RSC values were calculated by using a regression equation, based on the calibration curve prepared by using 0.008-0.26 mg mL−1 Trolox solutions. The antioxidant capacity of each sample is expressed in mg TE g−1 . Total phenolic content (TPC) was measured with Folin-Ciocalteu reagent [27]. Briefly, microplates were filled with 140 ␮L FolinCiocalteu reagent solution, 10 ␮L extract or gallic acid (GA) standard and 100 ␮L 7% Na2 CO3 solution; distilled water was used as a blank. The absorbance was read after 90 min at 765 nm. Calibration curve was produced using 0.35, 0.2, 0.1, 0.05, 0.025, 0.0125 mg mL−1 GA solutions. The results were expressed in mg GA equivalents in g of dry weight (mg GAE g−1 ). Oxygen radical absorbance capacity (ORAC) assay was performed according to Prior et al. [28]. The method uses fluorescein as a probe and reactive oxygen species (ROS) generated by the thermal decomposition of AAPH, which can quench the fluorescent signal of fluorescein. All absorption measurements were performed in a

102

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

FLUOstar Omega 96-well microplate reader (BMG Labtechnologies GmbH, Ortenberg, Germany). The L-ORAC assay has been adapted to measure lipophilic antioxidants using 7% RMCD solution in acetone/water (50/50 v/v) to dissolve the lipophilic antioxidants [29]. Ten mg of oil were dissolved in 1 mL of 7% RMCD solution to a final 0.0625% concentration. The 7% RMCD solution was used as a blank. The samples (25 ␮L) and fluorescein (150 ␮L, 14 ␮M) solutions were placed in 96-well black opaque microplates with transparent flat-bottom, which were sealed and incubated for 15 min at 37 ◦ C. Afterwards the AAPH solution as a peroxyl radical generator (26 ␮L; 240 mM) was added with a multichannel pipette. The microplate was placed in a FLUOstar Omega fluorescent reader and after shaking fluorescence was recorded every 66 s (in total 120 cycles) at 485 nm excitation and 510 nm emission wavelengths. At least three independent measurements were performed for each sample. Raw data were analyzed using software Mars (BMG Labtech GmbH, Offenburg, Germany). Antioxidant curves (fluorescence versus time) were normalized and the area under the fluorescence decay curve



i=120

(AUC) was calculated as AUC = 1 +

fi , where f0 f0

is the initial flu-

i=1

orescence reading at 0 min and fi is the fluorescence reading at time i. The final ORAC values were calculated by using regression equation between the reference antioxidant trolox concentration and the net area under the curve (AUC). A series of solutions (0–200 ␮M) of trolox was prepared for calibration. The antioxidant activity was expressed in ␮mol TE g−1 oil and raw material. 2.8. Assessment of antioxidant capacity by using QUENCHER procedure •

ABTS + scavenging, ORAC and TPC assays were also applied directly to the homogeneous ground particles of V. opulus as • described by Pastoriza et al. [30]. In ABTS + assay 10 mg of the sample were weighed in a test tube, diluted with 40 ␮L methanol • and 2 mL ABTS + added. The mixture was vortexed for 30 min to facilitate the surface reaction, centrifuged at 4500 rpm for 3 min, and 300 ␮L of optically clear supernatant was transferred to the microplate. In ORAC assay 10 mg of the sample was transferred to a test tube and 2 mL fluorescein added. The mixture was vortexed at 37 ◦ C for 15 min, 175 ␮L of solution transferred to the microplate and 25 ␮L of AAPH solution added. In TPC assay 10 mg of the sample were transferred to a test tube with 1.4 mL FolinCiocalteau’s reagent solution (1:13). The reagents were mixed, kept for 3 min, neutralized with 0.6 mL of 7% sodium carbonate, vortexed for 77 min and centrifuged at 4500 rpm for 3 min. The absorbance was measured at 765 nm. When the samples exerted too high antioxidant capacity, they were diluted with microcrystalline cellulose serving as an inert material. Microcrystalline cellulose-reagent mixtures were pre• pared as controls. Trolox solutions (ORAC and ABTS + ) and gallic acid (TPC) were used for calibration curves, using microcrystalline cellulose as well. The results are expressed in ␮mol TE and GAE per g dry weight (DW). 2.9. Experimental design The first series of experiments was performed with washed V. opulus berry pomace using screening two levels fractional factorial or half fraction design 24−1 with included 4 center points which will study the effects of 4 factors in order to adjust the most important independent variable’s effects. This design has resolution IV which will estimate main effects, but 2FI’s will be confounded together. The initial conditions of SFE-CO2 were chosen according to previous study with sea-buckthorn pomace [31]. The effects of pressure

Fig. 2. Particles size distribution of ground V. opulus berry products.

(P), time (t), temperature (T) and CO2 flow rate (v) on the extract yield were studied simultaneously (Table 1). All experiments were replicated three times. The order of the experiments has been fully randomized. In the second series of extraction experiments 2 independent variables, which had the most significant effect on extract yield in the first series, were further used in CCD. The optimal T and P were employed as central experimental design values (Table 1). CCD using RSM was used to optimize SFE-CO2 conditions for the unwashed V. opulus berry pomace. For data analysis and model establishing the software Design-Expert version 7.0.0 (Stat-Ease Inc., Minneapolis, MN) was used. Extraction P and t were chosen as independent variables with 5 levels for each of them, while T and v were constant during all experiments, 50 ◦ C and 2.5 L min−1 , respectively. The complete design consisted of 13 experimental runs with 4 factorial (22 ), 4 axial and 5 center points. A quadratic polynomial regression model was used to predict responses. Optimized extraction conditions were further employed in SFE-CO2 of the whole dried V. opulus berries. 3. Results and discussion 3.1. Particle size analysis Mass transfer during extraction is intensified by increasing the specific surface area of the material; therefore particle size and their distribution in ground material are important process characteristics. Particle size distribution was measured by laser diffraction method using distilled water as a dispersant because particles are not soluble in water, while their wetting by water is quite efficient (Fig. 2, Table 2). It may be observed that the dominant particle size of ground V. opulus pomace was in the range of 600–700 ␮m, whereas dried V. opulus berries contained the highest amount of 500 ␮m size particles. It should be noted that dried berries before grinding were frozen in liquid nitrogen to improve milling and this pretreatment may have the effect on better homogeneity of ground material. 3.2. Optimization of SFE-CO2 parameters Numerous factors such as T, P, v and t as well as particles size, the type and the origin of the plant material, its chemical composition and others influence the yield in SFE-CO2 and the composition of extracts obtained. Therefore, optimization of the process is an important issue for obtaining maximum yield with a desirable composition. The effects of the independent parameters on the extraction yield of washed V. opulus berry pomace were examined using half fraction factorial design: the yield varied from 8 to 19.1 g extract 100 g−1 berry material (Table 3). The highest yield was obtained at t = 120 min, T = 50 ◦ C, P = 55 MPa, v = 2.5 L min−1 ; the ratio of the total solvent mass and the mass of treated material was 27.

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

103

Table 2 Particle size parameters of V. opulus ground samples.a Samples

d (0.1)

d (0.5)

d (0.9)

D [4,3]

D [3,2]

Washed V. opulus pomace Unwashed V. opulus pomace Dried V. opulus berries

28.9 49.8 25.8

179.0 247.7 181.6

641.9 688.4 499.4

268.7 314.9 226.8

46.5 78.6 56.9

a The values are expressed in terms of percent (%) as a mean of six determinations. D [4,3] – volume weighted mean, D [3,2] – surface weighted mean, d(0.1) – 10% of the particles are smaller than this diameter, d(0.5) – half of the particles are smaller or larger than this diameter, d(0.9) – 90% of the particles are smaller than this diameter.

Table 3 Experimental design parameters and extract yields of washed V. opulus pomace. Runs

Pressure, MPa

Temperature, ◦ C

Time,min

CO2 flow, L min−1

Yield, g 100 g−1

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

45 45 35 35 35 55 55 55 35 55 45 45

40 40 30 50 30 50 30 50 50 30 40 40

90 90 60 60 120 60 120 120 120 60 90 90

2 2 1.5 2.5 2.5 1.5 1.5 2.5 1.5 2.5 2 2

18.7 18.0 8.0 10.6 18.2 13.2 18.1 19.1 14.8 15.3 17.9 16.8

Table 4 Analysis of variance table for factorial model. Source

SS

df

MS

F

p-value

Model Pressure (P), MPa Temperature (T), ◦ C Time (t), min CO2 flow (v), L min−1 Curvature Residual Lack of Fit Pure Error Total

102.13 24.82 0.46 66.30 10.56 26.99 7.11 5.18 1.93 136.22

4 1 1 1 1 1 6 3 3 11

25.53 24.82 0.46 66.30 10.56 26.99 1.18 1.73 0.64

21.55 20.94 0.38 55.95 8.91 22.78

0.0010* 0.0038* 0.5578** 0.0003* 0.0245* 0.0031*

2.69

0.2188**

SS, sum of square; df, degree of freedom; MS, mean square; F, Fisher value. * significant. ** not significant.

Model evaluation is presented in the analysis of variance (Table 4). The significance of each parameter was determined using the Student test (p-value). The analysis of the quadratic regression models for extract yield showed that the model was significant (p < 0.05) with an F-value of 21.55 and in this case the “lack of fit” was not significant relative to the pure error, with a p-value of 0.218. The model shows that the factor with the largest effect on extract yield was t (p < 0.05, F = 55.95), followed by P (p < 0.05; F = 20.94) and v (p < 0.05; F = 8.91), while the effect of T was not significant (p > 0.05, F = 0.38). Interaction between factors had no significant effect on the yield (p > 0.05). The adequacy of the model was evaluated by the total determination coefficient (R2 ) value of 0.93, indicating a reasonable fit of the model to the experimental data. The Pareto diagram shows the estimated effects in decreasing order of importance (Fig. 3). The length of each bar is proportional to the standardized effect, which is the estimated effect divided by its standard error. In this chart Bonferroni limit split the results, which are the most significant. Response surface plots showing the effect of t, T, P and v on extract yield are presented in Fig. 4a, b. The graphs were obtained by fixing two variables at central design values, while varying the remaining two. Fig. 4a illustrates the linear effects of T and t on extract yield at a fixed P = 45.0 MPa and v = 2 L min−1 . Extraction t had the main effect on oil yield comparing with the effect of T; P had bigger effect than v on yield at T = 40 ◦ C and t = 90 min

Fig. 3. Pareto chart displays the evaluation of independent variables effects in decreasing order of importance of washed V.opulus berries pomace.

104

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

Fig. 4. Response surface plots of washed V. opulus pomace using 32 full factorial design: A –the effect of temperature and time on extract yield and B – the effect of pressure and CO2 flow rate on extract yield, maintaining other two parameters constant at their center point. Table 5 Experimental design parameters and extract yields of unwashed V. opulus pomace. Runs

Time, min

Pressure, MPa

Yield, %

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

120 120 150 78 120 162 90 120 120 150 120 90 120

55 41 45 55 55 55 45 55 55 65 69 65 55

14.50 13.80 14.25 13.99 14.35 14.30 13.60 14.42 14.52 14.32 14.20 14.18 14.62

(Fig. 4b). Extract yield increased by increasing P and increasing v in the selected variables range. CCD was further applied for unwashed V. opulus pomace to optimize 2 most significant variables, P and t, as it was determined in the previous factorial design, whereas other parameters as T and v were maintained constant, 40 ◦ C and 2 L min−1 , respectively. The reduction of the number of variables in the 2nd series of optimization was supported by the results in Table 3. For instance, in the 1st series of experiments, when 4 variables were optimized, the yield at 45 MPa, 40 ◦ C, 90 min and 2 L min−1 flow rate was 18.7%, while at 55 MPa, 50 ◦ C, 120 min and 2.5 L min−1 flow rate it was almost similar, 19.1%. The yield in the 2nd sieres varied from 13.6 to 14.6 g oil 100 g−1 sample (Table 5); the highest yield of extract at suggested optimal conditions, 131 min and 57 MPa was 14.6 g 100 g−1 , the ratio of total solvent mass and mass of treated material was 24. It is interesting to note that similar high extract yield (14.61%) of unwashed raspberry pomace was previously obtained by SFE-CO2 at P = 45 MPa, T = 60 ◦ C and t = 120 min [13]; however previously reported yields obtained with pure CO2 from other berry pomace were remarkably lower, e.g. 7.6% from freeze-dried residue of blueberry [31] and only up to 3% from black chokeberry pomace [32]. As a general rule, a component with high vapor pressure has higher solubility in a supercritical medium. The solubility of the most components in the superctitical fluids increases with the increase of their density, which can be accomplished by increasing the extraction pressure. Two opposite effects occur by increasing the temperature, at constant pressure: it reduces the solvent power of CO2 by a decrease of the density, and it increases the vapor pressure of solutes which can be more easily transferred to the supercritical phase [34].

Fig. 5. Response surface plot of unwashed V. opulus pomace using CCD: the effect of pressure and time on extract yield.

Response surface quadratic model evaluation is presented in the analysis of variance (Table 6). The model shows that the factor with the largest effect on extract yield was t (p < 0.05; F = 22.43) followed by P (p < 0.05; F = 21.97), while the interaction between factors was also significant (p < 0.05). The second-order terms as t2 and P2 were significant (p < 0.05). The “lack of fit” was greater than 0.1 meaning that the variation in the model points does not differ significantly from the variation in the replicated points; in this case model fit the data well. The adequacy of the model was evaluated by the total determination coefficient (R2 ) value of 0.94, indicating a reasonable fit of the model to the experimental data. Adjusted coefficient of determination (R2 ) of 0.9 is in agreement with the predicted coefficient (R2 ) of 0.82. Response surface plots showing the effect of t and P on extract yield are presented in Fig. 5. P and t has quadratic effect on oil yield and time had the main influence on oil yield comparing with the effect of T. Oil yield increased by increasing P up to 57 MPa and t up to 130 min; further increasing parameter values did not have positive effect on extract yield. Optimized conditions were finally applied to the whole dried V. opulus berries; the highest extract yield was 6.6 g 100 g−1 DW at 131 min, 57 MPa, 2.5 L CO2 min−1 and 50 ◦ C; the ratio of the total

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

105

Table 6 Analysis of variance table for RSM experimental design. Source

SS

df

MS

F

p-value

Model Time (t), min Pressure (P), MPa tP t2 P2 Residual Lack of Fit Pure Error Total

0.96 0.19 0.18 0.065 10.56 0.39 0.059 0.017 0.042 1.02

5 1 1 1 1 1 7 3 4 12

0.19 0.19 0.18 0.065 10.56 0.39 0.00841 0.00559 0.011

22.75 22.43 21.97 7.73 8.91 46.56

0.0003* 0.0021* 0.0022* 0.0273* 0.0245* 0.0002*

0.53

0.6844**

SS, sum of square; df, degree of freedom; MS, mean square; F, Fisher value. * significant. ** not significant. Table 7 Concentration of tocopherols (T, ␮g−1 ) in V. opulus berry extracts obtained at optimal SFE-CO2 parameters. Samples

␣-T

␤-T

␥-T

␦-T

Total

Dried berries Unwashed berry pomace Washed berry pomace Dried berries (Soxhlet)

495.9 ± 74.6a 623.2 ± 67.3c 560.5 ± 30.8b 535.5 ± 36.1b

317.1 ± 28.3b 315.3 ± 18.2b 228.6 ± 8.7a 301.8 ± 9.1b

30.0 ± 0.7b 14.8 ± 1.6a 12.2 ± 0.7a 29.3 ± 0.9b

314.3 ± 35.4c 191.7 ± 7.5b 162.1 ± 5.1a 293.6 ± 29.1c

1157 1145 963 1160

Results are expressed as a mean ± standard deviation n = 3. Significant differences among means in the columns are indicated by the letters a, b, c and were determined by one-way ANOVA, using the statistical package GraphPad Prism 5. Tukey’s Least Significant difference (LSD) was used to determine significant difference among the treatments at p < 0.05.

Table 8 Fatty acid composition of V. opulus extracts. Fatty acids

Dried berries

Unwashed berry pomace

Washed berry pomace

Palmitic (C16:0) Stearic (C 18:0) Oleic (C18:1) Linoleic (C18:2) Eicosenoic (C20:1) ␣–Linolenic (C18:3n3)

9.12 2.24 42.14 42.43 0.54 1.54

4.31 1.23 50.69 41.95 0.29 1.48

3.96 1.02 47.56 45.62 0.30 1.53

Results are presented as a mean value of duplicate analysis and are given in%.

solvent mass and the mass of treated material was 29.45. The yield of lipophilic extract from dried V. opulus berries obtained in a Soxhlet apparatus with hexane was 7.8 g 100 g−1 . It may be observed that the highest yields were obtained from the washed berry pomace (19.1 g 100 g−1 ), followed by the unwashed pomace (14.6 g 100 g−1 ) and the whole dried berries (6.6 g 100 g−1 ). It may be explained by the presence of polar (hydrophilic) compounds (e.g. sugars) in the whole berries and unwashed berry pomace, which are not soluble in CO2 and also may interfere in the process of extraction; whereas washed berry pomace consists mainly of seeds, which usually are rich in oil for many berry species. Finally, the SFE-CO2 process was up-scalled using a 10 L pilot extractor at optimal P and T, whereas t was prolonged to 240 min and v of CO2 was 13.56 kg h−1 , obtaining the ratio of the total solvent mass and the mass of extracted material 28.54. In this case extract yield was 17.5 ± 0.8%, i.e. only slightly lower compared with the maximal yield of washed berry pomace achieved in laboratory scale equipment. 3.3. Extraction of SFE-CO2 residues with pressurized liquids As it has been mentioned, V. opulus berries are rich in polyphenolic compounds [1–3], which are poorly soluble in CO2 . Consequently SFE-CO2 of berry pomace with pure CO2 is not effective for polyphenolic antioxidants, which are valuable and abundant constituents in berries and their pomace. It was also recently demonstrated in the study with sour cherry pomace: 80% ethanol addition in SFE-

CO2 was required for achieving better extraction of plyphenolics [15]. Therefore, other methods should be applied for the recovery of polar phenolic antioxidants from the plant material. In general, two main approaches may be used: (1) the yield of polyphenolics in SFECO2 of berry pomace may be increased by using polar co-solvents, particularly such as food-friendly ethanol and water [32,33,35]; (2) SFE-CO2 residues should be re-extracted with higher polarity solvents [13,15,20,36]. So far as previously reported data demonstrated that addition of a polar co-solvent into the CO2 flow is not sufficiently efficient (e.g. high quantities of ethanol are required) for the isolation of polyphenolics [10,15,33], the second approach was selected in this study. PLE of freeze dried blueberry residues with ethanol, water and their mixture gave 4.2–8% of extracts possessing strong antioxidant capacity [37]. Fractionated high pressure extractions from elderberry pomace were performed using SFE-CO2 , followed by enhanced solvent extraction with diverse CO2 /ethanol/H2 O solvent mixtures in order to obtain anthocyaninrich fractions; the maximum extraction yield in this case was 24.2% [37]. However, in the study of Paes at al. [32] the solvents were used separately for the extraction of the whole berry residues, while in our study the SFE-CO2 residue of unwashed V. opulus berry pomace obtained at optimal parameters was consecutively re-extracted by PLE with the increasing polarity solvents: acetone, ethanol and water. The unwashed V. opulus berry pomace was selected for PLE because it is the main juice pressing by-product. The yields of the extracts isolated from the residues were 11.59 ± 0.95, 27.58 ± 0.79 and 30.71 ± 0.59%, respectively; when recalculated for the initial mass of pomace they were 9.90, 20.82 and 16.79%, respectively. Thus, the sum of extracts isolated from berry pomace during the all steps of biorefining (SFE+PLE) constituted 62.1% from the initial plant material. Generally selection of the approach depends on the berry pomace composition and the tasks of its processing. For instance, when the yield of lipophilic fraction is rather low [33] it should be important to estimate if the separation of this fraction would be feasible from the technological and economical points of view. However, the main advantage of consecutive re-extractions with different methods and different polarity solvents is that it enables to obtain several fractions with different product properties and composition. Thus, SFE-CO2 with pure CO2 gives lipophilic

106

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

Table 9 Antioxidant capacity (TEAC in ABTS assay, ORAC) and total phenols content (TPC) of V. opulus determined by QUENCHER procedure in the solid ground material. Samples

TPC,mg GAE/g DW

TEAC,␮mol TE/g DW

ORAC,␮mol TE/g DW

Dried berries before SFE-CO2 Dried berries after SFE-CO2 Unwashed berry pomace before SFE-CO2 Unwashed berry pomace after SFE-CO2 Washed berry pomace before SFE-CO2 Washed berry pomace after SFE-CO2 Dried berries after Soxhlet extraction

44.5 ± 5.9b 40.9 ± 5.7b 50.4 ± 6.5b 54.9 ± 6.7bc 42.1 ± 3.9b 40.1 ± 5.4b 36.5 ± 4.4ab

643 ± 87a 707 ± 84ab 1007 ± 85b 1414 ± 178c 863 ± 80ab 791 ± 81ab 687 ± 110ab

1277 ± 204b 1515 ± 167b 2576 ± 324c 2174 ± 220c 659 ± 80a 612 ± 75a 459 ± 50a

Values represented as mean ± standard deviation (n=3). Significant differences among means in the columns are indicated by the letters a, b, c and were determined by oneway ANOVA, using the statistical package GraphPad Prism 5. Tukey’s Least Significant difference (LSD) was used to determine significant difference among the treatments at p < 0.05.

Table 10 • The yields, antioxidant capacity (TEAC in ABTS + and DPPH• scavenging, ORAC) and total phenols content (TPC) of unwashed V. opulus berries pomace extracts obtained by PLE. •

PLE solvent

Extract yield, %

TPC,mg GAE/g DW

DPPH• ,mg TE/g DW

ABTS + ,mg TE/g DW

ORAC,mmol TE/g DW

Acetone Ethanol Water

11.59 ± 0.95 27.58 ± 0.79b 30.71 ± 0.59c

132.6 ± 0.60 88.6 ± 0.30a 174.9 ± 0.40c

121.8 ± 0.85 106.9 ± 0.48a 267.4 ± 1.17c

376.8 ± 2.85b 331.0 ± 1.22a 602.3 ± 6.62c

5.75 ± 0.99a 5.32 ± 0.37a 8.72 ± 2.28b

a

b

b

Values represented as mean ± standard deviation (n = 3). Significant differences among means in the columns are indicated by the letters a, b, c and were determined by one-way ANOVA, using the statistical package GraphPad Prism 5. Tukey’s Least Significant difference (LSD) was used to determine significant difference among the treatments at p < 0.05.

fractions, which in case of berry pomace are rich in polyunsaturated fatty acids and tocopherols; while PLE with other, higher polarity solvents isolates polyphenol-rich fractions usually possessing strong antioxidant capacity as it will be shown in further sections.

3.4. Concentration of tocopherols and fatty acid composition of extracts Tocopherols are the most important bioactive constituents in many plant seeds. They are strong lipophilic antioxidants and possess vitamin E activity. It may be observed that ␣-T was dominating in the all V. opulus extracts, followed by ␤-T, ␦-T and ␥-T (Table 7). In general, the differences between the extracts were not remarkable; slightly lower total concentration of tocopherols was determined in the extracts of washed berry pomace. Considering that extract yield from this type of pomace was considerably higher than from the other samples, tocopherols in the extract of washed pomace may be more diluted with other extracted lipophilic constituents. The total amount of tocopherols isolated from 1 g of initial berry, unwashed and washed pomace DW by SFE-CO2 was 76.3, 167.2 and 183.9 ␮g g−1 , showing that washed berry pomace contains the highest concentration of tocopherols. It was reported, that the oil isolated from V. opulus seeds originated from Siberia contained 1100, 400 and 600 ␮g g−1 of ␣-T, ␥-T and ␦-T, respectively [21]. Considering the differences with our data, it may be suggested that the genotype of the previously studied plant was different from the plant investigated in our study; also the yield of oil in the previous study was not indicated, which makes the comparison difficult. The dominant fatty acids in V. opulus extracts were: oleic (42–50%), linoleic (41–45%), palmitic (3–9%), stearic (1–2%) and ␣–linolenic 1.5% (Table 8). It is in agreement with the previously published data [21], which reported linoleic (50.1%) oleic (45.7%), and palmitic (1.6%) acids as the major fatty acids in V. opulus berry seed oil extracted at 35 MPa and 50 ◦ C. It should be noted, that the analysis of tocopherols and fatty acids was performed using homogeneous extract. During product storage in refrigerator at 4 ◦ C it was observed that the extract separates into 2 layers, heavier yellow colour fraction and lighter dark colour transparent oil fraction. Comprehensive analysis of the composition of these fractions was beyond the scope of this study,

however, it may be hypothesized that yellow colour fraction may be rich in carotenoids.

3.5. Antioxidant capacity of V. opulus raw materials, extracts and residues Determination of the in vitro antioxidant capacity of plant extracts may be considered as a first step in the evaluation of their bioactivities and are widely used for the preliminary assessment. The majority of the in vitro antioxidant activity assays are based on electron/hydrogen atom transfer reactions (SET/HAT). It was concluded that ORAC, TPC, and one of the SET/HAT assays should be recommended for the representative evaluation of antioxidant properties [38]. SET based methods include the TPC assay with Folin–Ciocalteu reagent and TEAC measurement by the ABTS•+ decolorisation assay. All these methods were applied for assessing antioxidant potential of V. opulus extracts in our study. The antioxidant capacity of V. opulus extracts obtained using SFE-CO2 was assessed by L-ORAC assay. The antioxidant capacity values of extracts isolated from V. opulus washed pomace, unwashed pomace, and dried berries at optimal SFE-CO2 conditions (the highest yield) were 65.3 ± 1.8, 74.3 ± 2.2 and 142.4 ± 3.6 ␮mol TE g−1 , respectively. The ORAC value of Soxhlet extract of dried berries (141.1 ± 2.8 ␮mol TE g−1 ) was similar to that of SFE-CO2 extract; however, CO2 possesses many advantages comparing with hexane, first of all it is a green solvent, which does not cause any hazardous residues in the final food grade ingredients. It may be observed that the extracts obtained from dried berries possessed 2-fold higher antioxidant capacity values comparing with the pomace extracts. Most likely, some antioxidatively active compounds, which are present in the whole V. opulus berries and are being pressed out during juice production, may be present in the extracts isolated by SFE-CO2 . On the other hand, when calculated for 1 g DW of the initial raw material, the amount of antioxidants from washed pomace, unwashed pomace, and dried berries at optimal SFE-CO2 conditions was 12.47, 10.86 and 9.40 ␮mol TE, respectively. Some antioxidants may be strongly bound to other components in plant material matrix and are not extracted by the solvents. According to Serpen, et al. [39], the values obtained by using QUENCHER method, which has been adapted for insoluble food

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

components, in most cases demonstrate a significant antioxidant activity. It was hypothesized that free functional groups on the surface of insoluble particles also quench with radicals. Therefore, it was interesting to determine antioxidant potential of berry materials before and after SFE-CO2 . For this purpose antioxidant properties of dried berries and pomace before and after SFE-CO2 were evaluated by Folin-Ciocalteu (TPC), ABTS•+ (TEAC) and ORAC methods (Table 9). In most cases, the antioxidant activity indicators were not significantly different for the berry material before and after SFE-CO2 . However, the differences between berry materials may be observed and therefore possible effects of SFE-CO2 should be briefly discussed. For instance, dried berries before extraction possessed slightly higher TPC, while other antioxidant activity values were higher in their residues after SFE-CO2 . In case of unwashed pomace, TPC and ABTS•+ scavenging after SFE-CO2 were found to be higher and ORAC lower (although not significantly) than before extraction, whereas the values for all residues of washed berry pomace slightly decreased after extraction. Most likely, these findings may be explained by the fact that lipophilic fraction containing lower concentrations of electron donating antioxidants was removed during SFE-CO2 and the concentration of remaining nonlipophilic antioxidants in the unwashed berry pomace residue became higher, thus resulting in stronger radical scavenging capacity of this fraction. Removal of lipophilic substances and treatment of pomace at high P and elevated T in the extractor may also change structural properties of the pomace particles and the availability of radical scavenging functional groups in the assay reaction. It is interesting noting, that the sum of TPC isolated from 1 g of unwashed pomace DW during the all PLE steps with acetone, ethanol and water was 61.1 mg GAE, i.e. only slightly higher than the value determined by the QUENCHER method (54.9 mg GAE/g DW) for this type of pomace (Table 9). The differences in TPC and ABTSTEAC values before and after SFECO2 of washed berry pomace were not significant; in this case, considering the above mentioned explanation, the main part of the hydrophilic antioxidants, which may be present in the unwashed pomace, most likely, is removed during washing. The ORAC values were not significantly different for berry material before and after SFE-CO2 . It should be noted that ORAC assay is in principal different from the single electron transfer based assays involving Folin • Ciocalteu’s reagent and ABTS + scavenging. The polar co-solvents are usually used for obtaining higher yields of polyphenolic compounds in SFE-CO2 . For instance, in the SFE-CO2 of grape marc the highest yield of phenolic compounds and the highest antioxidant activity was obtained by applying 15% of water [40]. Our study used different approach for the exhaustive isolation of polyphenolics; instead of adding a co-solvent in SFE-CO2 consecutive PLE with the increasing polarity solvents was applied to the extraction residues. The antioxidant activity of extracts isolated from the unwashed V. opulus berry pomace residues by PLE with different polarity solvents was evaluated by • TPC, DPPH• , ABTS + and ORAC assays. In general, antioxidant activity indicators were comparatively high for all extracts (Table 10); however water and ethanol extracts demonstrated the highest and the lowest antioxidant capacity values, respectively. For instance, the TPC value in water extract was 179.9 mg GAE/g; for comparison TPC in the juice of various V. opulus cultivars was in the range of 5.4-10.6 mg GAE/g [3]. These findings encourage more comprehensive studies of the composition and distribution of individual phytochemicals in the extracts, which were beyond the scope of this study. In general, the results support that PLE is a rapid and effective extraction method for the isolation of different polarity fractions from V. opulus berries pomace.

107

4. Conclusion SFE-CO2 parameters were optimized to obtain the highest extract yield from V. opulus berries and pomace using CCD-RSM. The models were well fitting experimental results. Washed berry pomace presented the highest quantity of lipophilic extract 18.6%, comparing with the unwashed pomace (14.6%) and whole dried berries (6.6%), which were extracted at 57 MPa pressure, 50 ◦ C temperature, 2.5 L min−1 flow rate in 131 min. Pressure and extraction time were the most significant factors for extract yields. The extracts obtained from dried V. opulus berries possessed the highest antioxidant capacity and contained higher amounts of tocopherols compared with the extracts from pomace; however, considering remarkably higher extract yields, the total amount of extractable antioxidants was higher in pomace than in the whole berries. Consecutive re-extraction of SFE-CO2 residues with the increasing polarity pressurized solvents (PLE with acetone, ethanol and water) enabled to recover polyphenol-rich fractions possessing high antioxidant capacity. In general, SFE-CO2 was proved as an effective method for the isolation of lipophilic fractions from V. opulus berry pomace, while PLE was demonstrated as an efficient method for recovering higher polarity and high antioxidant capacity substances; consequently a multistep biorefining scheme of berries and their pomace may substantially increase process effectiveness and reduce the amount of insoluble residues. Acknowledgement This study was funded by Research Council of Lithuania, grant no. SVE-01/2014. References ˇ ˇ ˙ R. Daubaras, P. Viˇskelis, A. Sarkinas, [1] L. Cesonien e, Determination of the total phenolic and anthocyanin contents and antimicrobial activity of Viburnum opulus fruit juice, Plant Foods Hum. Nutr. 67 (2012) 256–261. ˇ ˙ E. Stackeviˇciene, ˙ J. Labokas, L. Cesonien ˙ R. [2] P.R. Venskutonis, V. Kraujalyte, e, Daubaras, Phytochemical characterisation of highbush blueberry (Vaccinium covilleanum) and European cranberry bush (Viburnum opulus) accessions grown in Lithuania, Pharm. Biol. 50 (2012) 551. ˇ ˙ P.R. Venskutonis, A. Pukalskas, L. Cesonien ˙ R. Daubaras, [3] V. Kraujalyte, e, Antioxidant properties and polyphenolics composition of fruits from different European cranberry bush (Viburnum opulus L.) genotypes, Food Chem. 141 (2013) 3695–3702. [4] A.A. Karac¸elik, M. Küc¸ük, Z. I˙ skefiyeli, S. Aydemir, S. De Smet, B. Miserez, P. Sandra, Antioxidant components of Viburnum opulus L. determined by on-line HPLC-UV-ABTS radical scavenging and LC-UV-ESI–MS methods, Food Chem. 175 (2015) 106–114. ˇ ˇ ˙ R. Daubaras, V. Kraujalyte, ˙ P.R. Venskutonis, A. Sarkinas, [5] L. Cesonien e, Antimicrobial activity of Viburnum opulus fruit juices and extracts, J. Verbrauch. Lebensm. 9 (2014) 129–132. [6] O. Rop, V. Reznicek, M. Valsikova, T. Jurikova, J. Mlcek, D. Kramarova, Antioxidant properties of European cranberrybush fruit (Viburnum opulus var. edule), Molecules 15 (2010) 4467–4477. [7] Y.S. Velioglu, L. Ekici, E.S. Poyrazoglu, Phenolic composition of European cranberry bush (Viburnum opulus L.) berries and astringency removal of its commercial juice, Int. J. Food Sci. Technol. 41 (2006) 1011–1015. ˙ E. Leitner, P.R. Venskutonis, Chemical and sensory [8] V. Kraujalyte, characterization of aroma of Viburnum opulus fruits by solid phase microextraction-gas chromatography-olfactometry, Food Chem. 132 (2012) 717–723. [9] P. Benelli, C.A.S. Riehl, A. Smania, E.F.A. Smania, S.R.S. Ferreira, Bioactive extracts of orange (Citrus sinensis L. Osbeck) pomace obtained by SFE and low pressure techniques: mathematical modeling and extract composition, J. Supercrit. Fluids 55 (2010) 132–141. [10] L. Manna, C.A. Bugnone, M. Banchero, Valorization of hazelnut, coffee and grape wastes through supercritical fluid extraction of triglycerides and polyphenols, J. Supercrit. Fluids 104 (2015) 204–211. [11] C. Da Porto, A.N. Lino, D. Decorti, The combined extraction of polyphenols from grape marc: ultrasound assisted extraction followed by supercritical CO2 extraction of ultrasound-raffinate, LWT-Food Sci. Technol. 61 (2015) 98–104. [12] F. Akay, A. Kazan, M.S. Celiktas, O. Yesil-Celiktas, A holistic engineering approach for utilization of olive pomace, J. Supercrit. Fluids 99 (2015) 1–7. ¯ e, ˙ P. Kraujalis, P.R. Venskutonis, Optimization of high pressure [13] N. Kryˇzeviˇciut extraction processes for the separation of raspberry pomace into lipophilic and hydrophilic fractions, J. Supercrit. Fluids 108 (2016) 61–68.

108

P. Kraujalis et al. / J. of Supercritical Fluids 122 (2017) 99–108

[14] A. Massias, S. Boisard, M. Baccaunaud, F.L. Calderon, P. Subra-Paternault, Recovery of phenolics from apple peels using CO2 + ethanol extraction: kinetics and antioxidant activity of extracts, J. Supercrit. Fluids 98 (2015) 172–182. ´ ˛ [15] L. Wozniak, K. Marszałek, S. Skapska, Extraction of phenolic compounds from sour cherry pomace with supercritical carbon dioxide: impact of process parameters on the composition and antioxidant properties of extracts, Separ. Sci. Technol. 51 (2016) 1472–1479. [16] F.J. Barba, Z.Z. Zhu, M. Koubaa, A.S. Sant’Ana, V. Orlien, Green alternative methods for the extraction of antioxidant bioactive compounds from winery wastes and by-products: a review, Trends Food Sci. Technol. 49 (2016) 96–109. [17] X. Xu, Y.X. Gao, G.M. Liu, Q. Wang, H. Zhao, Optimization of supercritical carbon dioxide extraction of sea buckthorn (Hippophae rhamnoides L.) oil using response surface methodology, LWT-Food Sci. Technol. 41 (2008) 1223–1231. [18] G.M. Liu, X. Xu, Q.F. Hao, Y.X. Gao, Supercritical CO2 extraction optimization of pomegranate (Punica granatum L.) seed oil using response surface methodology, LWT-Food Sci. Technol. 42 (2009) 1491–1495. [19] H.I. Castro-Vargas, L.I. Rodríguez-Varela, S.R.S. Ferreira, F. Parada-Alfonso, Extraction of phenolic fraction from guava seeds (Psidium guajava L.) using supercritical carbon dioxide and co-solvents, J. Supercrit. Fluids 51 (2010) 319–324. ˜ [20] T. Brazdauskas, L. Montero, P.R. Venskutonis, E. Ibanez, M. Herrero, Downstream valorization and comprehensive two-dimensional liquid chromatography-based chemical characterization of bioactives from black chokeberries (Aronia melanocarpa) pomace, J. Chromatogr. A 1468 (2016) 126–135. [21] B. Yang, M. Ahotupa, P. Määttä, K. Kallio, Composition and antioxidative activities of supercritical CO2 -extracted oils from seeds and soft parts of northern berries, Food Res. Int. 44 (2011) 2009–2017. [22] Authority ISO Standards, ISO, 3320-1, Particle Size Analysis-laser Diffraction Methods, Part 1: General Principles, Authority ISO Standards, 1999. [23] AOAC, Official Methods of Analysis of Association of Official Analytical Chemists, 15th ed., AOAC, Washington, DC, 1995. [24] J. Gruszka, J. Kruk, RP-LC for determination of plastochromanol, tocotrienols and tocopherols in plant oils, Chromatographia 66 (2007) 909–913. [25] R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M. Yang, C. Rice–Evans, Antioxidant activity applying an improved ABTS radical cation decolorization assay, Free Radic. Biol. Med. 26 (1999) 1231–1237. [26] W. Brand-Williams, M.E. Cuvelier, C. Berset, Use of a free radical method to evaluate antioxidant activity, LWT − Food Sci. Technol. 28 (1995) 25–30. [27] V.L. Singleton, J.A. Rossai Jr., Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents, Am. J. Enol. Viticult. 16 (1965) 144–158.

[28] R.L. Prior, H. Hoang, L. Gu, X. Wu, M. Bacchiocca, L. Howard, M. Hampsch-Woodill, D. Huang, B. Ou, R. Jacob, Assays for hydrophilic and lipophilic antioxidant capacity (oxygen radical absorbance capacity (ORAC-FL) of plasma and other biological and food samples, J. Agric. Food Chem. 51 (2003) 3273–3279. [29] D. Huang, B., Ou, M. Hampsch-Woodill, J.A., Flanagan, E.K. Deemer, Development and validation of oxygen radical absorbance capacity assay for lipophilic antioxidants using randomly methylated ␤-cyclodextrin as the solubility enhancer, J. Agric. Food Chem. 50 (2002) 1815–1821. [30] S. Pastoriza, C. Delgado-Andrade, A. Haro, J.A. Rufián-Henares, A physiologic approach to test the global antioxidant response of foods. The GAR method, Food Chem. 129 (2011) 1926–1932. [31] D. Cossuta, B. Simándi, J. Hohmann, F. Doleschall, T. Keve, Supercritical carbon dioxide extraction of sea buckthorn (Hippophae rhamnoides L.) pomace, J. Sci. Food Agric. 87 (2007) 2472–2481. [32] J. Paes, R. Dotta, G.F. Barbero, J. Martínez, Extraction of phenolic compounds and anthocyanins from blueberry (Vaccinium myrtillus L.) residues using supercritical CO2 and pressurized liquids, J. Supercrit. Fluids 95 (2014) 8–16. ˙ M. Pukalskiene, ˙ A. Pukalskas, P.R. Venskutonis, Fractionation of [33] L. Grunovaite, black chokeberry pomace into functional ingredients using high pressure extraction methods and evaluation of their antioxidant capacity and chemical composition, J. Funct. Foods 24 (2016) 85–96. ˜ [34] J.A. Mendiola, M. Herrero, M. Castro-Puyana, E. Ibánez, Supercritical fluid extraction, in: M.A. Rostagno, J.M. Prado (Eds.), Natural Product Extraction–Principles and Applications, RSC Publishing Cambridge, 2013, pp. 196–230. [35] A.M. Farias-Campomanes, M.A. Rostagno, M.A.A. Meireles, Production of polyphenol extracts from grape bagasse using supercritical fluids: yield, extract composition and economic evaluation, J. Supercrit. Fluids 77 (2013) 70–78. [36] M.J. Otero-Pareja, L. Casas, M.T. Fernández-Ponce, C. Mantell, E.J. Martínez de la Ossa, Green extraction of antioxidants from different varieties of red grape pomace, Molecules 20 (2015) 9686–9702. [37] I.J. Seabra, M.E.M. Braga, M.T. Batista, H.C. de Sousa, Effect of solvent (CO2 /ethanol/H2 O) on the fractionated enhanced solvent extraction of anthocyanins from elderberry pomace, J. Supercrit. Fluids 54 (2010) 145–152. [38] D. Huang, B. Ou, R.L. Prior, The chemistry behind antioxidant capacity assays, J. Agric. Food Chem. 53 (2005) 1841–1856. [39] A. Serpen, E. Capuano, V. Gökmen, V. Fogliano, A new procedure to measure the antioxidant activity of insoluble food components, J. Agric. Food Chem. 55 (2007) 7676–7681. [40] C. Da Porto, D. Decorti, A. Natolino, Water and ethanol as co-solvent in supercritical fluid extraction of proanthocyanidins from grape marc: a comparison and a proposal, J. Supercrit. Fluids 87 (2014) 1–8.