Ultrafiltration of orange press liquor: Optimization of operating conditions for the recovery of antioxidant compounds by response surface methodology

Separation and Purification Technology 98 (2012) 255–261

Contents lists available at SciVerse ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Ultrafiltration of orange press liquor: Optimization of operating conditions for the recovery of antioxidant compounds by response surface methodology René Ruby-Figueroa, Alfredo Cassano ⇑, Enrico Drioli Institute on Membrane Technology, ITM-CNR, c/o University of Calabria, via P. Bucci, 17/C, I-87030 Rende (CS), Italy

a r t i c l e

i n f o

Article history: Received 11 May 2012 Received in revised form 20 July 2012 Accepted 31 July 2012 Available online 8 August 2012 Keywords: Ultrafiltration (UF) Orange press liquor Response surface methodology (RSM) Polyphenols Total antioxidant activity

a b s t r a c t In the present work the ultrafiltration (UF) of orange press liquor was analyzed by means of the response surface methodology (RSM). The liquor was clarified by using polysulphone hollow fiber membranes with a molecular weight cut-off (MWCO) of 100 kDa. The influence of different operating conditions, such as transmembrane pressure (TMP), axial feed flow rate (Qf) and temperature (T) on the membrane rejection towards polyphenols and the recovery of antioxidant compounds in the permeate stream was investigated. The experimental operating conditions were selected within the following ranges: TMP 0.2–1.4 bar, temperature 15–35 °C, and feed flow-rate 85–245 L/h. A total of 30 ultrafiltration experiments were performed. Judged by the lack-of-fit criterion, the analyses of variance (ANOVA) showed the regression model to be adequate. From the regression analyses, the membrane rejection towards polyphenols and total antioxidant activity (TAA) were expressed with quadratic equations of TMP, Qf and T. The predicted rejection towards polyphenols and TAA in the clarified liquor from the regression model were presented in 3D surface plots. Quadratic terms of TMP, T and Qf showed significant (p > 0.05) influence on the polyphenols rejection. Results indicated a strong interaction effect of TMP with T and Qf. For the TAA the quadratic effect of T was the most significant. In this case a strong interaction effect of TMP with Qf was detected. The desirability function approach was applied to analyze the regression model equations in order to optimize the recovery of antioxidant compounds in the permeate stream. The optimized operating conditions were 0.2 bar, 19.85 °C and 244.64 L/h (maximum TAA, minimum polyphenols rejection). Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Several studies have highlighted the structure–antioxidant activity relationship of flavonoids, hydroxycinnamic acids and coumarins extracted from plant materials [1]. In recent years, antioxidants have gained a great interest because of their potential as prophylactic and therapeutic agents in many diseases [2–6]. Consequently, the global market of antioxidants is increasing rapidly, because of the increased health risk in a constantly polluting environment. These agents also have cosmetic applications, leading to the development of researches at industrial and academic level to explore these molecules and their analogues. In this way there is a great interest in the separation, purification and recovery of antioxidant compounds from natural sources. Some citrus flavonoids have been reported to possess a variety of biological activities and pharmacological properties, including antiallergic, antidiabetic antiinflammatory, antiviral, antiprolifera⇑ Corresponding author. Tel.: +39 0984 492067; fax: +39 0984 402103. E-mail address: [email protected] (A. Cassano). 1383-5866/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2012.07.022

tive, and anticarcinogenic activities, in addition to having effects on mammalian metabolism [7,8]. The main use of citrus in food industries includes fresh and concentrated juice or citrus-based drinks. Since the juice yield of citrus is less than half of the fruit weight, the juice production is accomplished through the formation of very large amounts of by-product wastes. Peel and seed residues are the primary waste fraction and can be considered as an interesting source of phenolic compounds [9]. Over the past 10 years, different studies have been proposed on the recovery of flavonoids from by-products of orange juice processing based on the use of resins [10], organic solvents [11], enzymes [12], c-irradiation [13] and heat treatment [14]. These natural antioxidants offer, in fact, interesting perspectives and opportunities in the production of dietary supplements and functional foods and for their possible utilization by pharmaceutical and cosmetic industries. However, the proposed methodologies have drawbacks to some degree. For example, the extraction with organic solvents is characterized by safety problems (some of them are believed to be toxic), low efficiency and time consumption; heat treatment results in pyrolysis; enzymes can be denatured in enzyme-assisted extraction; c-irradiation assisted extraction is

256

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

still unknown in terms of safety. In addition, phenolic compounds are sensitive to light, heat and oxygen exposure. As reported recently by Galanakis [15] the recovery of antioxidants from agricultural by-products is today considered to be conducted in five principal stages: (1) a macroscopic pre-treatment; (2) a macro- and micro-molecules separation; (3) an extraction step; (4) an isolation and purification step; (5) a product formation. Pressure-driven membrane processes, such as microfiltration (MF), ultrafiltration (UF) and nanofiltration (NF) can be used in different stages, i.e. MF in macroscopic pretreatment, UF in macro- and micro-molecules separation, NF in purification. All these processes allow to carry out concentration and separation without the use of heat; the equipments need small space, are flexible and easy to scale-up; operating costs and energy consumption are low; products and co-products are of high quality. All these factors are very important for by-products valorization [16]. On the other hand, a disadvantage of pressure-driven membrane processes is related to the requirement of excessive pretreatment due to their sensitivity to concentration polarization and membrane fouling due to chemical interaction with feed constituents. Orange peels and pulp are semi-solid wastes obtained in the orange juice production which can be pressed to obtain a solution named press liquor. This liquor is a complex mixture containing soluble sugars (sucrose, glucose and fructose), insoluble carbohydrates, fibers, organic acids, essential oils, flavonoids and carotenoids [17]. UF can be considered a valid approach for separating and recovering valuable compounds from finely divided solid materials present in citrus press liquor. Specifically UF is able to remove macromolecules, such as pectins and proteins, from the liquor ensuring the production of a clarified solution containing health benefit compounds [18]. It has recently been used to clarify phenol containing beverages from the larger fragments of flavonols and hydroxycinnamic derivatives without affecting importantly the antioxidant profile of the permeate [19]. In a previous paper Ruby Figueroa et al. [20] investigated the influence of different operating conditions on flux decline in the UF of orange press liquor by using the response surface methodology (RSM) approach. This approach is widely used to analyze the effects of multiple factors and their interaction making the overall optimization of the process feasible [21–23]. The present study aimed at evaluating the effect of operating variables such as transmembrane pressure (TMP), temperature (T) and feed flow-rate (Qf) on the performance of hollow fiber UF membranes in terms of recovery of antioxidant compounds in the permeate stream through the Response Surface Methodology (RSM). This approach allowed to select the optimal operating conditions to minimize the retention of phenolic compounds and maximize the antioxidant capacity of the clarified solution.

2.2. UF equipment and procedures Experimental runs were performed by using a laboratory bench-plant equipped with a UF hollow fiber polysulphone membrane module supplied by China Blue Star Membrane Technology Co., Ltd. (Beijing, China) having an effective membrane area of 1600 cm2 and a nominal molecular weight cut-off (MWCO) of 100 kDa. A heat exchanger, placed into the feed tank, was used to keep the feed temperature constant. In Fig. 1 a schematic diagram of the UF experimental setup is depicted. Experiments were performed according to the total recycle configuration in which both permeate and concentrate streams were continuously recycled to the feed tank to ensure a steady state in the volume and composition of the feed. Table 1 summarizes the operating conditions tested. Each run was stopped after 3 h of operation, when a quasi-stationary permeate flux was reached. Permeate flux was gravimetrically measured at different time intervals (each 10 min). After each run the UF membrane module was submitted to a cleaning procedure with an enzymatic solution (Ultrasil 50, Henkel Chemicals Ltd., 1 g/100 g) at a temperature of 40 °C for 60 min. After a rinsing with distilled water for 15–20 min, the membrane module was cleaned with a 0.05 M NaOH solution at 40 °C for 60 min. A final rinsing with distilled water for 20 min was performed. After the cleaning step the water permeability was checked in fixed conditions (temperature 20 °C, feed flow-rate 245 L/h) and compared with the initial value (191.22 L/m2 h bar). The recovery of the hydraulic permeability after the cleaning procedure was higher than 90%. 2.3. Determination of total phenols content Total phenols were estimated colorimetrically by using the Folin–Ciocalteau method [24,25]. The method is based on the reduction of tungstate and/or molybdate in the Folin–Ciocalteu reagent by phenols in alkaline medium resulting in a blue colored

2. Materials and methods 2.1. Feed solution Citrus press liquors, obtained from blond orange peels, were supplied by Gioia Succhi Srl (Rosarno, Reggio Calabria, Italy). Their initial content of suspended solids was 7.13 ± 1.41 g/100 g. Liquors were left overnight at room temperature to let the majority of the cloud particles settle out. Partially clear liquor was recovered by decanting of the cloud layer. Liquors were then depectinised by adding 7 g/kg of pectinase (4 h at room temperature) from Aspergillus aculeatus (Sigma–Aldrich, Milan, Italy) and filtered with a nylon cloth. The obtained liquors (feed solution) presented a pH value and a total soluble solids (TSS) content of 3.6 and 8.6 °Brix, respectively. They were stored at 17 °C and defrosted to room temperature before use.

Fig. 1. Scheme of the UF experimental set-up. 1 – feed tank; 2 – feed pump; 3,6 – manometers; 4 – membrane module; 5 – thermometer; 7 – cooling coil; 8 – regulation valve; 9 – digital balance.

Table 1 Experimental range and levels of the independent variables for Box–Behnken design. Feed factors

TMP (bar) Temperature (°C) Feed flow rate (L/h)

Code

X1 X2 X3

Variation levels 1

0

1

0.2 15 85

0.8 25 165

1.4 35 245

257

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

product. Gallic acid was used as a calibration standard and results were expressed as gallic acid equivalent (GAE) (mg GAE L1). The absorbance was measured by using a UV–visible spectrophotometer (Shimadzu UV-160A, Japan) at 765 nm. The separation capability of the UF membrane was expressed in terms of membrane rejection as reported in the following:

  CP  100 R¼ 1 Cf

ð1Þ

where R is the membrane rejection towards polyphenols, Cp and Cf the polyphenols concentration in the permeate stream and feed solution, respectively.

2.4. Determination of the total antioxidant activity (TAA) Total Antioxidant Activity (TAA) was determined by an improved version of the ABTS radical cation decolourisation assay in which the chromogenic ABTS (2,2-azino-bis-(3-ethylbenzothiazoline-6-sulfonic acid) radical cation is generated by oxidation of ABTS with potassium persulphate before the addition of the antioxidant [26,27]. The ABTS decolourisation was measured as inhibition percentage of the absorbance at 734 nm. The concentration of antioxidant giving the same inhibition percentage of absorbance of the radical cation as 1 mM 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) was calculated in terms of the Trolox Equivalent Antioxidant Capacity (TEAC).

2.5. Experimental design Operating conditions, such as transmembrane pressure (TMP), temperature (T) and feed flow rate (Qf) were studied to evaluate their effect on the membrane performance expressed as polyphenols rejection and total antioxidant activity (TAA) in the permeate stream. Based on the capability of the experimental set-up and the preliminary single-variable tests, the ranges of operating conditions were chosen as follows: TMP = 0.2–1.4 bar, T = 15–35 °C and Qf = 85–245 L/h. The levels, code and intervals of variation of the operating variables are given in Table 1. Experimental runs were performed according to the Box–Behnken design (Table 2) which is composed of 30 runs divided in two blocks, each one with three central points (runs 13, 14, 15 and 28, 29, 30). The correlation of the operating variables and the responses based on the Box–Behnken design is fitted to a quadratic polynomial equation [28] using the least-square method of the form:

Y k ¼ b0 þ

3 3 2 X 3 X X X bi X i þ bii X 2i þ bij X i X j þ e i¼1

i¼1

ð2Þ

i¼1 ji1

where Yk is the response variable (Y1 total polyphenols rejection, Y2 for total antioxidant activity in the permeate stream), b0 the constant, e the residual (error or noise) term, bi the linear coefficients, bii the quadratic coefficients, bij the interaction coefficients and Xi the dimensionless coded variables (X1 for TMP, X2 for T and X3 for Qf). The system of equations was solved by using a multiple regression technique called method of least squares (MLSs). The criterion for choosing the bi estimates is that they should minimize the sum of the squares of the residuals [29]. Analysis of variance (ANOVA) was used to determine the linear and quadratic effect of the operating factors studied.

Table 2 Experimental design and results of Box Behnken design. Run

Block

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

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

TMP (bar)

Temperature (°C)

Feed flow rate (L/h)

Polyphenols rejection (%)

TAA (mM trolox)

X1

X2

X3

Y1

Y2

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

10.66 39.79 47.68 34.77 46.64 47.13 40.08 54.42 46.69 50.62 37.34 43.49 47.48 44.07 51.51

12.48 28.10 25.12 23.92 24.52 24.45 41.84 23.50 19.11 29.41 21.84 33.10 26.68 31.73 35.60

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

9.09 49.18 58.30 57.65 55.60 42.73 42.51 49.41 49.63 51.64 57.47 49.98 47.12 48.83 54.28

13.75 27.64 23.67 23.49 25.15 24.45 41.91 23.46 19.31 29.41 23.16 32.89 27.68 29.00 34.15

3. Results and discussion 3.1. Effect of operating conditions on the rejection of phenolic compounds The model obtained could explain the 70.37% (expressed in R-squared statistic) of the variability in the rejection of polyphenols and the standard error was equal to 3.73. In addition, other tests were performed to determine the capacity of the model to describe the experimental data. The lack-of-fit test showed a value of p = 0.0726 (F-ratio 4.69) and the residuals did not show a serial autocorrelation with a value greater than 0.05 (p = 0.5027; F-ratio 2.62) in the Durbin–Watson statistic; therefore, the model appears to be adequate to describe the observed data at 95% confidence level. The linear, quadratic and interaction between the factors studied are shown in a Pareto chart (Fig. 2). The interaction factor between TMP and T (b12) shows the most significant effect on the polyphenols rejection, followed by linear effect of temperature (b2) and linear effect of TMP (b1). The factors (b2), (b1), (b13) and (b33) appear to produce an increase in the polyphenols rejection while the factors as (b12), (b11) and (b22) produced a decrease in the same item. On the other hand, the factors (b3) and (b23) did not produce a significant effect (p P 0.05) in the polyphenols rejection. The interaction factor (b23) presenting the lowest effect were not included in the regression model equation. The quadratic regression (Eq. (3)) describes the effect of operating conditions on the polyphenols rejection:

Y 1 ¼ 14:03 þ 61:02X 1 þ 4:33X 2  0:30X 3  16:02X 21  1:72X 12 þ 0:09X 13  0:05X 22 þ 0:0007X 23

ð3Þ

258

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

b12 b2:Temperature b1:TMP b11 b22 b13 b33 b3:Feed flow rate b23

This effect was kept until 1.1 bar, where some of the particles (polyphenols) in the solid layer cross through the membrane due to the applied pressure (Fig. 3a and c). On the other hand, any change in the temperature will force molecules as proteins to assume a new equilibrium structure. An increase of temperature produced a gradual destabilization of the main non-covalent interactions, producing exposure of non-polar groups [30]. This destabilization could stimulate the solid layer formation onto the membrane surface, increasing the polyphenols rejection as showed in Fig. 3a and b. Meanwhile the Qf showed a strong interaction with the TMP and T (Fig. 3b and c). The main effect is to reduce the polyphenols rejection due to the reduction of the adjacent boundary layer thickness [31]. This effect was observed up to 165 L/h; after this point an increase in the recirculation flow rate produced an increase in the polyphenols rejection due to more particles arriving at the membrane surface which produce a pore blocking reducing the MWCO [32]. All these modifications can be attributed to concentration polarization and membrane fouling phenomena. During membrane filtration rejected solutes are convectively driven to the membrane surface where they build up a concentration polarization boundary layer near the membrane surface. In UF macromolecular solutes and colloidal species usually have insignificant osmotic pressure; in this case the concentration at the membrane surface rises up to a point where a gel precipitation layer forms on the membrane surface. This layer offers a major resistance to

+ -

0

2

4

6

8

Standardized effect Fig. 2. Standardized Pareto chart for total polyphenols rejection.

where Y1 is the predictive polyphenols rejection for the ultrafiltration process. The positive coefficient in the Eq. (3) means that these factors produce an increasing in the polyphenols rejection, whereas the negative coefficient in the equation produces a decrease when these factors are increased. The 3D plot (Fig. 3), built using the Eq. (3), shows the main effect of the studied variables and their interaction on the polyphenols rejection. The response surface of the polyphenols rejection was plotted against two operating variables, while the third variable was kept constant (level 0 in Table 1). It is possible to appreciate the significant variation in the polyphenols rejection when the operating conditions were modified in the investigated range. In particular, the polyphenol rejection increased by increasing TMP due to the effect of a solid layer compact that, acting as an additional layer, reduces the MWCO of the UF membranes [16].

Polyphenols rejection (%)

(a) 60 50 40 30 20 10 0

0

0.3

0.6

0.9

1.2

TMP (bar)

1.5

15

19

23

27

31

35

Temperature (°C)

Polyphenols rejection (%)

(b) 58 54 50 46 42 38

260 200 15

19

140 23

27

31

Temperature (°C)

35

80

Feed flow rate (L/h)

Polyphenols rejection (%)

(c) 60 55 50 45 40 260

35 30

200 0

0.3

140 0.6

TMP (bar)

0.9

1.2

1.5

80

Feed flow rate (L/h)

Fig. 3. 3D response surface with contour plot for total polyphenols rejection.

259

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

Residual polyphenols rejection

23

b22 b2:Temperature b13 b3:Feed flow rate b12 b11 b1:TMP b33 b23

13 3 -7 -17 0

10

20

30

40

50

0

60

Fig. 4. Plot of residuals against predicted response of total polyphenols rejection in the UF process.

permeate flow. Concentration polarization is considered to be reversible and is controlled by means of velocity adjustment. It is however directly related to fouling which is considered to be irreversible and indicates permanent loss of membrane permeability. In general, membrane fouling is affected by three main factors: membrane material properties, feed characteristics and operating parameters [33]. Weis et al. [34] reported that phenolic compounds are the major foulants in the UF process. Therefore, the effect of operating conditions on the molecules rejected (proteins, pectin and polyphenols), responsible of the solid layer formed on the membrane surface, should be strictly considered. Considering the previous work [20], it is possible to observe that there is a strong relationship between the type of fouling and polyphenols rejection, because changes in the fouling type produce changes in the polyphenols rejection. Finally the maximum polyphenols rejection was obtained when the intermediate pore blocking was the governing fouling mechanism. The residual (errors) of the polyphenols rejection (Fig. 4) shows a random distribution over and below the centerline suggesting that the model for the polyphenols rejection is statistically significant. 3.2. Effect of operating conditions on TAA The TAA was evaluated in the permeate stream to determine the effect of operating conditions in the preservation of antioxidant compounds. The regression model built for TAA at different operating conditions could explain the 69.02 percent of the variability (expressed in R-squared) with a standard error equal to 3.98. The p-value for lack-of-fit was equal to 0.3935 (F-ratio 1.44). This means that the model as fitted is adequate to explain the variability in the TAA at 95% of confident level. Since the p-value of the Durbin–Watson statistic test is greater than 0.05 (p = 0.9986), there is no indication of serial autocorrelation in the residuals at the 5.0% significance level. Fig. 5 shows the linear, quadratic and interaction coefficients of each operating variable plotted in the form of Pareto chart. It is possible to appreciate that the quadratic effect of T was the most significant followed by (b2), (b13) and (b3), whereas the linear effect (b1), interaction effect (b12, b23) and quadratic effect (b11, b33) did not show a significant effect (p P 0.05) on the TAA. The Eq. (4) shows the regression model that describes the linear, quadratic and interaction effect of operating conditions on the TAA in the permeate stream.

8:20X 21

 0:64X 12  0:09X 13  0:06X 22 þ 0:00009X 23 þ 0:0001X 23

1

2

3

4

Standardized effect

Predicted polyphenols rejection

Y 2 ¼ 44:60 þ 43:70X 1 þ 3:64X 2 þ 0:07X 3 

+ -

ð4Þ

where Y2 is the predictive TAA in the permeate stream for the UF process.

Fig. 5. Standardized Pareto chart for TAA in the permeate stream.

In Fig. 6a–c the response surfaces on the TAA are plotted against two operating variables while the third variable was kept constant (0 level). It can be seen that for all Qf values investigated, an increase in the T (until 28 °C) produced an increase in the TAA value of the permeate stream; at temperatures higher than 28 °C an increase in the T produced a decrease of the TAA value (Fig. 6b). A similar effect was observed for the TMP in the range 0.2–0.4 bar; after this point an increase in the TMP produced a decrease in the TAA value (Fig. 6a). In all the range of TMP values an increase in Qf enhanced the TAA of the clarified liquor. At the maximum value of Qf (245 L/h) an increase in TMP produced a decrease in the TAA (Fig. 6c). At the minimum level of Qf (85 L/h) the increasing in the TAA value took place up to a critical point of TMP where the TAA started to decrease when TMP was increased. The differences in the TAA values at different operating conditions can be correlated with the changes in the polyphenols rejection. The TAA is strictly correlated to the content of polyphenols of the orange press liquor. Proteggente et al. [35] analyzed the phenolic composition and the TAA of fresh Sicilian orange juice from pigmented and non-pigmented varieties of orange (Citrus sinensis L. Osbeck). They found that concentrations of anthocyanins and hydroxycinnamic acids are highly correlated to the TAA values while ascorbic acid seems to play a minor role. Their observations were consistent with a previous report in which the antioxidant action of similar varieties was ascribed to the phenolic content [36]. Recent studies have also shown that total phenols determined by Folin–Ciocalteu method can be correlated to the antioxidant activity determined by different methods (ABTS and DPPH assays, for instance) [37]. For this reason, the method described by Singleton et al. [25] has been proposed recently as a standardized method to use in the routine quality control and measurement of antioxidant capacity of food products and dietary supplements [38]. It has been also reported that thermal treatments, such as conventional and microwave cooking, frying and warm-holding of blanched products produce a modification of polyphenols content and also of antioxidant activity [39,40]. The observed decreasing of TAA after 28 °C, could be attributed to the increasing in polyphenols rejection when the temperature is raised (see also Fig. 3b). Fig. 7 shows the residual plot of TAA in the permeate stream: the dispersion is random over and below the centerline suggesting that the model for the TAA is statistically significant. 3.3. Optimization of multiple responses The desirability function (DF) was used to determine a combination of variables to optimize multiple responses. In this optimization the objective was to minimize the polyphenols rejection and maximize the TAA in the permeate stream, simultaneously. The minimum and maximum values of polyphenols rejection used to obtain the desirability function were 9.09 (dk = 0) and 58.30 (dk = 1). On the other hand for TAA the limits were 12.48 and

260

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

TAA permeate (mM trolox)

(a) 32 28 24 20 16 12

0

0.3

0.6

0.9

1.2

1.5

TMP (bar)

19

15

23

27

31

35

Temperature (°C)

TAA permeate (mM trolox)

(b) 37 34 31 28 25 260

22 19

200 15

19

23

27

140 31

35

Temperature (°C)

80

Feed flow rate (L/h)

TAA permeate (mM trolox)

(c) 41 37 33 29 25 260

21 17

200 0

0.3

140

0.6

0.9

TMP (bar)

1.2

1.5

80

Feed flow rate (L/h)

Residual TAA in permeate

Fig. 6. 3D response surface with contour plot for TAA in the permeate stream.

Table 3 Optimization results for the UF of citrus press liquor.

8 5

Optimized coded level of variables

X1: TMP (bar) X2: Temperature (°C) X3: Feed flow-rate (L/h)

Predicted responses

Polyphenols rejection (%) TAA (mM trolox)

2 -1 -4

Overall desirability

-7 12

17

22

27

32

37

0.2 19.85 244.64 28.45 31.28 0.59

42

Predicted TAA in permeate Fig. 7. Plot of residuals against predicted response of TAA in the permeate stream.

41.91 (mM trolox) for the minimum (dk = 0) and maximum (dk = 1), respectively. Results of the optimization are summarized in Table 3. In the optimized operating conditions predicted responses for polyphenols rejection and TAA in the clarified liquor were 28.45% and 31.28 mM Trolox, respectively. The maximum overall desirability found was equal to 0.59. The obtained results indicated a minimum polyphenols rejection of the UF membrane under operating conditions of minimal concentration polarization and fouling (Qf = 244.64 L/h; TMP = 0.2 bar). A similar behavior has been also observed in other studies. For instance, Cassano et al. [41] observed that the rejection of total phenols in the UF of grape must increased by increasing TMP due to an increasing of fouling. In the clarification of apple juice by UF de Bruijn et al. [42] identified low TMP and high tangential velocity as optimum conditions to minimize

fouling. Finally, Todisco et al. [43] noticed that the rejection coefficient of a ceramic UF membrane towards polyphenols of black tea increased linearly with increasing TMP and only at the highest cross-flow velocity the retention of polyphenols was found not to vary with the operating pressure.

4. Conclusions The effect of operating conditions in the UF of orange press liquor was studied to determine the effect on the membrane performance in terms of polyphenols rejection and recovery of antioxidant compounds in the clarified liquor. The response surface methodology was used to analyze the interaction between the investigated parameters. Temperature, TMP and feed flow rate play an important role in membrane retention. In addition, a strong relationship between membrane fouling and polyphenols retention was observed.

R. Ruby-Figueroa et al. / Separation and Purification Technology 98 (2012) 255–261

Optimization of multiple responses permitted to establish the operating conditions giving maximum recovery of TAA in the permeate and minimum polyphenols rejection, simultaneously. For an overall desirability of 0.59, polyphenols rejection of 28.45% and TAA of 32.28 mM Trolox in the clarified liquor, were estimated, respectively, in optimized operating conditions of 0.2 bar, 19.85 °C and 244.64 L/h.

References [1] F. Natella, M. Nardini, M. Di Felice, C. Scaccini, Benzoic and cynnamic acids derivatives as antioxidants: structure-activity relation, J. Agr. Food Chem. 47 (1999) 1453–1459. [2] D.V. Ratnam, D.D. Ankola, V. Bharswaj, D.K. Sahana, M.N.V. Ravi Kumar, Role of antioxidants in prophylaxis and therapy: a pharmaceutical perspective, J. Control. Release 113 (2006) 189–207. [3] P.M. Kris-Etherton, K.D. Hecker, A. Bonanome, S.M. Coval, A.E. Binkoski, K.F. Hilpert, Bioactive compounds in foods: their role in the prevention of cardiovascular disease and cancer, Am. J. Med. 113 (2002) 71–88. [4] S. Liu, J.E. Manson, I.M. Lee, R.C. Stephen, C.H. Hennekens, W.C. Wilett, Fruit and vegetable intake and risk of cardiovascular disease: the women’s health study, Am. J. Clin. Nutr. 72 (2000) 922–928. [5] N. Salah, N.J. Miller, G. Paganga, L. Tijburg, G.P. Bolwell, C.A. Rice-Evans, Polyphenolic flavanols as scavengers of aqueous phase radicals and as chainbreaking antioxidants, Arch. Biochem. Biophys. 322 (1995) 339–346. [6] J.W. Lampe, Health effects of vegetables and fruits: assessing mechanisms of action in human experimental studies, Am. J. Clin. Nutr. 70 (Suppl.) (1999) 475s–490s. [7] R.S.R. Zand, D.J.A. Jenkins, E.P. Diamandis, Flavonoids and steroid hormonedependent cancers, J. Chromatogr. B 777 (2002) 219–232. [8] W. Ren, Z. Qian, H. Wang, L. Zhu, L. Zhang, Flavonoids: promising anticancer agents, Med. Res. Rev. 23 (2003) 519–534. [9] A. Bocco, M.E. Cuvelier, H. Richard, C. Berset, Antioxidant activity and phenolic composition of citrus peel and seed extracts, J. Agric. Food Chem. 46 (1998) 2123–2129. [10] A. Di Mauro, B. Fallico, A. Passerini, E. Maccarone, Waste water from citrus processing as a source of hesperidin by concentration of styrene– divinylbenzene resin, J. Agric. Food Chem. 48 (2000) 2291–2295. [11] B.B. Li, B. Smith, Md.M. Hossain, Extraction of phenolics from citrus peels. I. Solvent extraction method, Sep. Purif. Technol. 48 (2006) 182–188. [12] B.B. Li, B. Smith, Md.M. Hossain, Extraction of phenolics from citrus peels. II. Enzyme-assisted extraction method, Sep. Purif. Technol. 48 (2006) 189–196. [13] H. Oufedjikh, M. Mahrouz, M.J. Amiot, M. Lacroix, Effect of c-irradiation on phenolic compounds and phenylalanine ammonia-lyase activity during storage in relation to peel injury from peel of Citrus Clementina Hort. Ex. Tanaka, J. Agric. Food Chem. 48 (2000) 559–565. [14] G.H. Xu, X. Ye, J. Chen, D. Liu, Effect of heat treatment on the phenolic compounds and antioxidant capacity of citrus peel extract, J. Agric. Food Chem. 55 (2007) 330–335. [15] C.M. Galanakis, Recovery of high added-value components from food wastes: conventional, emerging technologies and commercialized applications, Trends Food Sci. Technol. (2012), http://dx.doi.org/10.1016/j.tifs.2012.03.003. [16] H. Strathmann, L. Giorno, E. Drioli, An introduction to membrane science and technology, Consiglio Nazionale delle Ricerche, Roma, 2006. [17] E.M. Garcia-Castello, J.R. McCutcheon, Dewatering press liquor derived from orange production by forward osmosis, J. Membr. Sci. 372 (2011) 97–101. [18] N. Pap, M. Mahosenaho, E. Pongrácz, H. Mikkonen, M. Jaakkola, V. Virtanen, L. Myllykoski, Z. Horváth-Hovorka, C. Hodur, G. Vatai, R.L. Keiski, Effect of ultrafiltration on anthocyanin and flavonol content of black currant juice (Ribes nigrum L.), Food Bioprocess. Technol. 5 (2012) 921–928, http:// dx.doi.org/10.1007/s11947-010-0371-z. [19] C.M. Galanakis, E. Tornberg, V. Gekas, Clarification of high-added value products from olive mill wastewater, J. Food Eng. 99 (2) (2010) 190–197.

261

[20] R.A. Ruby Figueroa, A. Cassano, E. Drioli, Ultrafiltration of orange press liquor: optimization for permeate flux and fouling index by response surface methodology, Sep. Purif. Technol. 80 (2011) 1–10. [21] M.F. Anjum, I. Tasadduq, K. Al-Sultan, Response surface methodology: a neutral network approach, Eur. J. Oper. Res. 101 (1997) 65–73. [22] R.H. Myers, D.C. Montgomery, Response Surface Methodology: Process and Product Optimization Using Designed Experiments, John Wiley Sons Inc., New York, 1995. [23] M.C. Martí-Calatayud, M.C. Vincent-Vela, S. Álvarez-Blanco, J. Lora-García, E. Bergantiños-Rodríguez, Analysis and optimization of the influence of operating conditions in the ultrafiltration of macromolecules using a response surface methodological approach, Chem. Eng. J. 156 (2010) 337–346. [24] E. Roura, C. Andrés-Lacueva, R. Estruch, M.R. Lamuela-Raventós, Total polyphenol intake estimated by modified Folin–Ciocalteu assay of urine, Clin. Chem. 52 (2006) 749–752. [25] V.L. Singleton, R. Orthofer, R.M. Lamuela-Raventos, Analysis of total phenols and other oxidation substrates and antioxidants by means of Folin–Ciocalteu reagent, Method. Enzymol. 299 (1999) 152. [26] C.A. Rice-Evans, N.J. Miller, Total antioxidant status in plasma and body fluids, Method. Enzymol. 243 (1994) 279–298. [27] R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M. Yang, C.A. Rice-Evans, Antioxidant activity applying and improved ABTS radical cation decolorization assay, Free Radical Bio. Med. 26 (1999) 1231–1237. [28] M.N. Hyder, R.Y.M. Huang, P. Chen, Pervaporation dehydration of alcohol– water mixtures: optimization for permeate flux and selectivity by central composite rotatable design, J. Membr. Sci. 326 (2009) 343–353. [29] D. Bas, I.H. Boyaci, Modeling and optimization I: usability of response surface methodology, J. Food Eng. 78 (2007) 836–845. [30] S. Domodaran, Aminoacids, Peptides and Proteins, in: O. Fennema (Ed.), Food Chemistry, Marcel Dekker Inc., New York, 1997, pp. 321–430. [31] M. Moresi, I. Sebastiani, Pectin recovery from model solutions using a laboratory-scale ceramic tubular UF membrane module, J. Membr. Sci. 322 (2008) 349–359. [32] K.J. Hwang, C.Y. Liao, K.L. Tung, Analysis of particle fouling during microfiltration by use of blocking models, J. Membr. Sci. 287 (2007) 287–293. [33] H. Susanto, Y. Feng, M. Ulbricht, Fouling behavior of aqueous solutions of polyphenolic compounds during ultrafiltration, J. Food Eng. 91 (2009) 333– 340. [34] A. Weis, M.R. Bir, M. Nyström, C. Wright, The influence of morphology, hydrophobicity and charge upon the long-term performance of ultrafiltration membranes fouled with spent sulphite liquor, Desalination 175 (2005) 73–85. [35] A.R. Proteggente, A. Saija, A. De Pasquale, C.A. Rice-Evans, The compositional characterisation and antioxidant activity of fresh juices from Sicilian sweet orange (Citrus sinensis L. Osbeck) varieties, Free Radical Res. 37 (2003) 681– 687. [36] P. Rapisarda, A. Tomaino, R. Lo Cascio, F. Bonina, A. De Pasquale, A. Saija, Antioxidant effectiveness as influenced by phenolic content of fresh orange juices, J. Agric. Food Chem. 47 (1999) 4718–4723. [37] V. Roginsky, A.E. Lissi, Review of methods to determine chainbreaking antioxidant activity in food, Food Chem. 92 (2005) 235–254. [38] R.L. Prior, X. Wu, K. Schaich, Standardized methods for the determination of antioxidant capacity and phenolics in foods and dietary supplements, J. Agric. Food Chem. 53 (2005) 4290–4302. [39] C. Ewald, S. Fjelkner-Modig, K. Johansson, I. Sjöholm, B. Akesson, Effect of processing on major flavonoids in processed onions, green beans, and peas, Food Chem. 64 (1999) 231–235. [40] D. Zhang, Y. Hamauzu, Phenolics, ascorbic acid, carotenoids and antioxidant activity of broccoli and their changes during conventional and microwave cooking, Food Chem. 88 (2004) 503–509. [41] A. Cassano, A. Mecchia, E. Drioli, Analyses of hydrodynamic resistances and operating parameters in the ultrafiltration of grape must, J. Food Eng. 89 (2008) 171–178. [42] J. de Bruijn, A. Venegas, R. Borquez, Influence of crossflow ultrafiltration on membrane fouling and apple juice quality, Desalination 148 (2002) 131–136. [43] S. Todisco, P. Tallarico, B.B. Gupta, Mass transfer and polyphenols retention in the clarification of black tea with ceramic membranes, Innov. Food Sci. Emerg. Technol. 3 (2002) 255–262.